Social support on social media among young adult cancer caregivers: a mixed method study

Young adult cancer caregivers' (YACC) ability to thrive is enhanced when they feel socially supported. Social media (SM) is an important source of support for young adults yet little is known about how YACCs use SM during cancer caregiving. The purpose of this convergent mixed methods study was to illuminate factors associated with YACCs' use of, and social support received from, SM during the first six months of caregiving. In Aim 1, we interviewed YACCs (n=34) to describe their use of SM. We applied grounded theory coding methods (κ=0.88), which yielded key insights about perceptions and patterns of use: 1) YACCs primarily used SM to provide updates about the cancer patient as this was less burdensome than other communication strategies; 2) YACCs perceived SM as enabling access to social support they would not otherwise have accessed; 3) YACCs experienced negative consequences from posting online including misinformation and uncomfortable responses; 4) While negative experiences led some YACCs to avoid posting about cancer, most thought potential benefits outweighed harms. In Aim 2, we examined responses to YACCs' SM posts (n=2093) for prevalence of five types of functional social support-emotional, information, validation, companionship, instrumental-and how prevalence changed over the first six months of caregiving. Coding for support type (κ=0.95) was quantified to calculate prevalence; iii changes in prevalence over time were compared using linear regression modeling. Emotional and information support were the most prevalent support types; information (adjusted odds ratio (aOR): 1.15, 95% Confidence Interval (CI) 1.09-1.21) and companionship (aOR: 1.12, 95%CI 1.02-1.24) significantly increased in prevalence over time while emotional decreased (aOR: 0.90, 95%CI 0.85-0.94). Aim 3 examined the effects of SM platform and presence/absence of cancer-related content on support using linear mixed effects regression modeling. Instagram posts had more emotional support responses than Facebook posts (β=0.25, standard error (SE)=0.09, p=0.007). Cancer-related content was associated with more validation support (β=0.20, SE=0.07, p=0.002), but less emotional (β=-0.17, SE=0.07, p=0.02) and instrumental (β=-0.06, SE=0.02, p=0.001) support compared to posts without cancer-related content. Study findings yield insight into the use of SM for social support among YACCs, including future directions for SM-based support interventions. iv

SOCIAL SUPPORT ON SOCIAL MEDIA AMONG YOUNG ADULT CANCER CAREGIVERS: A MIXED METHOD STUDY by Echo Lyn Warner A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Nursing The University of Utah December 2019 Copyright © Echo Lyn Warner 2019 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Echo Lyn Warner has been approved by the following supervisory committee members: and by Lee Ellington , Co-Chair 10/02/2019 Anne Kirchhoff , Co-Chair 10/02/2019 Kristin Gates Cloyes , Member Ye Sun , Member Andrew Ralph Wilson , Member and by David B. Kieda, Dean of The Graduate School. Date Approved Date Approved Date Approved 10/02/2019 Date Approved , Chair/Dean of Barbara L. Wilson the Department/College/School of Date Approved Nursing ii ABSTRACT Young adult cancer caregivers’ (YACC) ability to thrive is enhanced when they feel socially supported. Social media (SM) is an important source of support for young adults yet little is known about how YACCs use SM during cancer caregiving. The purpose of this convergent mixed methods study was to illuminate factors associated with YACCs’ use of, and social support received from, SM during the first six months of caregiving. In Aim 1, we interviewed YACCs (n=34) to describe their use of SM. We applied grounded theory coding methods (κ=0.88), which yielded key insights about perceptions and patterns of use: 1) YACCs primarily used SM to provide updates about the cancer patient as this was less burdensome than other communication strategies; 2) YACCs perceived SM as enabling access to social support they would not otherwise have accessed; 3) YACCs experienced negative consequences from posting online including misinformation and uncomfortable responses; 4) While negative experiences led some YACCs to avoid posting about cancer, most thought potential benefits outweighed harms. In Aim 2, we examined responses to YACCs’ SM posts (n=2093) for prevalence of five types of functional social support—emotional, information, validation, companionship, instrumental—and how prevalence changed over the first six months of caregiving. Coding for support type (κ=0.95) was quantified to calculate prevalence; iii changes in prevalence over time were compared using linear regression modeling. Emotional and information support were the most prevalent support types; information (adjusted odds ratio (aOR): 1.15, 95% Confidence Interval (CI) 1.09-1.21) and companionship (aOR: 1.12, 95%CI 1.02-1.24) significantly increased in prevalence over time while emotional decreased (aOR: 0.90, 95%CI 0.85-0.94). Aim 3 examined the effects of SM platform and presence/absence of cancerrelated content on support using linear mixed effects regression modeling. Instagram posts had more emotional support responses than Facebook posts (β=0.25, standard error (SE)=0.09, p=0.007). Cancer-related content was associated with more validation support (β=0.20, SE=0.07, p=0.002), but less emotional (β=-0.17, SE=0.07, p=0.02) and instrumental (β=-0.06, SE=0.02, p=0.001) support compared to posts without cancerrelated content. Study findings yield insight into the use of SM for social support among YACCs, including future directions for SM-based support interventions. iv To my parents, my first teachers, for instilling in me the value of hard work and a job well done. To the women who cultivated my yearning for knowledge in grade school. For the ones who believed in me in high school and pushed me to be more. To all the undergraduate professors whom I admired but feared. To my graduate instructors, philosophers, and mentors who challenged my assumptions and broadened my world views. To the ones who taught me the most, Cyrus and Ophelia. I learn from you both every day. To my husband, Zachary, for teaching me to believe in myself and being my constant partner in the pursuit of this knowledge. To all the young adult cancer caregivers. TABLE OF CONTENTS ABSTRACT........................................................................................................................iii LIST OF TABLES ............................................................................................................. ix LIST OF FIGURES ........................................................................................................... xi LIST OF ACRONYMS AND ABBREVIATIONS ........................................................ xiii ACKNOWLEDGEMENTS ............................................................................................. xiv Chapters 1: INTRODUCTION .......................................................................................................... 1 Problem Statement ...........................................................................................................1 Significance .....................................................................................................................2 Specific Aims, Hypotheses, and Research Questions......................................................5 Innovation and Contribution ............................................................................................8 2: YOUNG ADULT CANCER CAREGIVERS SOCIAL SUPPORT THROUGH SOCIAL MEDIA: INTEGRATED REVIEW AND CONCEPTUAL FRAMEWORK ...10 Introduction ....................................................................................................................10 Young Adult Cancer Caregivers ....................................................................................10 Developmental Context of Young Adulthood ...............................................................13 Emerging Adulthood......................................................................................................14 Young Adulthood ..........................................................................................................15 Social Support in Online Communities .........................................................................16 Theoretical Framework and Models ..............................................................................20 Integration and Limitations of Caregiver Model and Social Support Theory ...............24 Conclusion .....................................................................................................................26 3: RESEARCH DESIGN AND METHODS .................................................................... 30 Introduction ....................................................................................................................30 Pilot Study Summary .....................................................................................................30 Mixed Methods Study Design .......................................................................................33 Participant Eligibility and Consent ................................................................................34 Sampling ........................................................................................................................35 Targeted Enrollment ......................................................................................................36 Aim 1 Purpose ...............................................................................................................36 Aim 1 Data Collection ...................................................................................................37 Aim 1 Data Analysis ......................................................................................................37 Aim 1 Limitations, Alternative Strategies, and Benchmarks ........................................38 Aims 2-3: Purpose .........................................................................................................39 Aims 2-3: Data Collection .............................................................................................39 Aim 2: Data Analysis .....................................................................................................41 Aim 3: Data Analysis .....................................................................................................42 Aims 2-3: Limitations, Alternative Strategies, and Benchmarks ..................................42 Mixed Methods Design and Integration ........................................................................43 Proposed Manuscripts………………………………………………………………………………………………44 4: YOUNG ADULT CANCER CAREGIVERS' USE OF SOCIAL MEDIA FOR SOCIAL SUPPORT……….……………………………………………………………………………………………..51 Abstract ..........................................................................................................................51 Introduction ....................................................................................................................52 Methods .........................................................................................................................54 Results ............................................................................................................................57 Discussion ......................................................................................................................65 Conclusion .....................................................................................................................69 5: SOCIAL SUPPORT EXCHANGES ON SOCIAL MEDIA DURING THE FIRST SIX MONTHS OF YOUNG ADULT CANCER CAREGIVING …………………….....................77 Abstract ..........................................................................................................................77 Background ....................................................................................................................79 Methods .........................................................................................................................83 Results ............................................................................................................................88 Discussion ......................................................................................................................95 6: CANCER RELATED CONTENT AND SOCIAL MEDIA PLATFORM INFLUENCE YACC SOCIAL SUPPORT ON SOCIAL MEDIA ……………………………………109 Abstract …………………………………………………………………………………………………………………109 Background ………………………………………………………………………………………………..………….111 Methods …………………………………………………………………………………………..…………………….114 Results ………………………………………………………………………………….……………………………….117 Discussionv ………………………………………………………………..…………………………………………119 Conclusions ……………………………………………………………………………………………………………124 7: SUMMARY …………………………………………………………………………………………………………..135 vii Introduction .................................................................................................................135 Review of Terms………………………………………..………..………………………………………………..136 Social Media Are Important Tools Used by Young Adults to Ease the Burden of Cancer Caregiving .......................................................................................................137 Young Adult Cancer Caregivers Experience Challenges and Benefits of Using Social Media During a Cancer Experience .............................................................................139 Social Support on Social Media Changes Over the First Six Months of Caregiving.. 141 Influence of Social Support on Young Adult Cancer Caregiver Wellbeing................144 Limitations……………………………………………………………………………………………….…………….144 Implications for Research ............................................................................................145 Implications for Clinical Practice ................................................................................147 Conclusion ...................................................................................................................148 Appendices A SEMISTRUCTURED INTERVIEW AND SURVEY SCRIPT .............................149 B SOCIAL MEDIA SCREENING FORM .................................................................159 REFERENCES.…………………………………………………………………………………………………..……..161 viii LIST OF TABLES Tables 2.1 Definitions ……………………………………………………………………………………..………………………27 2.2 Specific social issues and milestones for YACC by age………………………………………….27 2.3 Functional social support definitions and social media examples ……………………………28 3.1 Inclusion and exclusion criteria ………………………………………………………………………………46 3.2 Targeted enrollment ……………………………………………………………………………………………….46 3.3 Aim 1 qualitative analysis, first and second cycle coding ……………………………………….46 3.4 Aim 3 YACC, cancer patient, and social media independent variables …………………47 4.1 Sociodemographic characteristics of YACC (N=34) …………………………………………….70 4.2 Characteristics of cancer patients (N=34) ……………………………………………………………..71 4.3 Thematic categories, subthemes, and exemplary quotes for functional social support types on social media among YACC …………………………………………………………………………..72 4.4 Exemplary quotes for functional social support types on social media among YACC …………………………………………………………………………………………………………………………75 5.1 Functional social support definitions and social media examples ………………………….102 5.2 Sociodemographic characteristics of YACC (N=34) ……………………………………………103 5.3 Characteristics of social media posts (N=2,090 posts from N=33 YACC) ……………105 5.4 Prevalence of five types of functional social support exchanged between YACC and their followers on social media during the first six months of caregiving (N=2,090 posts from N=33 YACC) ……………………………………………………………………………………………………106 ii 5.5 Generalized linear models evaluating the odds of posts containing each type of social support over the six-month period, adjusted for platform type (N=2,090 posts from N=33 YACC) ……………………………………………………………………………………………………………………..106 6.1 Sociodemographic and cancer characteristics of YACC (N=34) …………………………..127 6.2 Characteristics of responses from social media posts (N=1,527 responses from N=33 participants) ………………………………………………………………………………………………………………128 6.3 Prevalence of five types of functional social support acquired by young adult cancer caregivers during the first six months of caregiving (N=1,527, N=33 YACC) …………..128 6.4 Univariate linear mixed effects models comparing social support by platform type and cancer content (N=1,527 responses from N=33 participants) ………………………………129 x LIST OF FIGURES Figures 2.1 Examples and definitions of terms ………………………………………………………………………..28 2.2 Cancer Family Caregiving Experience Model. Used with © permission from Fletcher, 2012 ……………………………………………………………………………………………………………….…………..29 2.3 Stress and Coping Social Support Theory. Adapted from Cohen, 2000 …………………29 3.1 Mixed methods design ………………………………………………………………..………………………….49 3.2 Aim 2 Hypothesized changes in functional social support over the first six months of caregiving …………………………………………………………………………………………………..…………..…..49 3.3 Mixed methods (MM) data integration …………………………………………………………………..50 3.4 Aims 2 and 3 hypotheses, analysis, and future research domains …………….……………..50 4.1 Frequency of commonly used social media platforms …………………………………………….76 5.1 Total number of social media posts exchanged between YACC and followers during the six months following the cancer patient’s diagnosis …………………………………………..…107 5.2 Change in the total number of posts containing each type of social support over time since patient diagnosis (N=2,090 posts from N=33 YACC) ……………………………….……..107 5.3 Change in the total proportion of posts containing each type of social support over time since patient diagnosis (N=2,090 posts from N=33 YACC) ………………………………………108 6.1 Mixed methods exploratory sequential design. Adapted from Creswell & Creswell, 2018 ...……………………………………………………………………………………………………………………....126 6.2 Differences in mean percent of responses to YACC’s social media posts that contain emotional support by platform type ……………………………………………………….………………….129 ii 6.3 Differences in mean percent of responses to YACC’s social media posts that contain information support by platform type …………………………………………………………………………130 6.4 Differences in mean percent of responses to YACC’s social media posts that contain companionship support by platform type ……………………………………………………………….…..130 6.5 Differences in mean percent of responses to YACC’s social media posts that contain validation support by platform type ………………………………………………………..………………….131 6.6 Differences in mean percent of responses to YACC’s social media posts that contain instrumental support by platform type …………………………………………………………………………131 6.7 Differences in mean percent of responses to YACC’s social media posts that contain emotional support by cancer-related content…………………………………….………………………….132 6.8 Differences in mean percent of responses to YACC’s social media posts that contain information support by cancer-related content………………………………………………….………….132 6.9 Differences in mean percent of responses to YACC’s social media posts that contain companionship support by cancer-related content………………………………………………………..133 6.10 Differences in mean percent of responses to YACC’s social media posts that contain validation support by cancer-related content…………………………………………………….………….133 6.11 Differences in mean percent of responses to YACC’s social media posts that contain instrumental support by cancer-related content ……………………………………………………………134 xii LIST OF ACRONYMS AND ABBREVIATIONS Acronyms/Abbreviations aOR: Adjusted odds ratio CI: Confidence interval Coef.: Coefficient DF: Degrees of freedom HCI: Huntsman Cancer Institute HIAYA: Huntsman Intermountain Adolescent and Young Adult Oncology Program PASS: Power Analysis Statistical Software PDAs: Paralinguistic digital affordances REDCap: Research Electronic Data Capture SD: Standard deviation SE: Standard error U.S.: United States YACC: Young adult cancer caregiver ACKNOWLEDGEMENTS It feels impossible to put into words my gratitude to all who have supported me in the pursuit of this knowledge. First, I want to say how extremely thankful I am to each of my committee members for their collective support, mentorship, and patience with me in the conduct of this research. Anne, you were the first person to believe I had the potential for a research career. It’s hard to thank you for this, because the breadth of what you have done for me is difficult to put into words. You’ve molded and shaped me in ways that I believe few mentors do for their mentees. I will always remember the many drafts we edited for my first research manuscript, and the patience it must have required of you. Your attention to detail is unmatched, and whenever I think I have finished a final draft of anything I go back and double check all the numbers, and then send it to someone else to double check because I know I will have inevitably missed something. Thank you for teaching me to value my mental and emotional wellbeing (by not working in an office without windows and the door shut ). Thank you for teaching me professionalism, mentorship, and collegiality. You are the person who taught me science, fostered my focus on adolescent and young adult oncology, and supported me personally and financially in professional networking at my first and many subsequent research conferences. Just really, really, a very big humongous thank you for believing in me and guiding me from my first step into scientific research to now. ii Lee, many times I have shared the sage advice you gave me at the start of my PhD. You said, over dinner, “Life doesn’t stop just because you’re in a PhD program.” I was so worried about the challenge of balancing my personal and professional lives as my husband and I were ready to start a family and I knew we would likely be moving out of state for his medical training. Thank you for being committed to supporting me as a whole person, including being a scholar, mother, runner, and student. Furthermore, I have you to thank for fostering my interest and eventual focus on young cancer caregivers. Your content expertise, intuition, and wisdom have guided me at every step of this process. Kristin, you were the first professor to ever teach me philosophy. To say I struggled with the concepts, rhetoric, and writing would be quite an understatement. Philosophy of Science, my first course from you, was without a doubt, the most intellectually challenging class of my education. In retrospect, I believe this was because this class not only taught me new ways of thinking about science, but new ways of thinking about life. As you lectured, my world views morphed and developed, just like the emerging adults I describe in this dissertation. You once quoted a saying I will never forget, “There is no one correct way to draw a face.” Those words have lived on a sticky note at my desk for the past five and a half years, reminding me that experience is lived and that life is not black and white, right or wrong. I deeply appreciate your encouragement during my pilot study that led to the submission and receipt of my F31 award from the National Cancer Institute, and eventually to this finished dissertation. Thank you for your kindness, understanding, and humor. xv ii Andy, our failed Skype chats will always bring a smile to my face. Technology, unlike statistics, never seemed to be reliable when we needed it. I appreciate your statistical tenacity and very patient instruction in the conduct of Aim 2 and 3 analyses for this dissertation. Thank you for sharing your expertise in handling repeated measures data and for kindly answering all of my R questions. I will always appreciate how you situated statistical concepts in a language I could (try) to understand. Ye, your role in my professional growth has been critical to furthering my understanding of social media analysis, network analysis, and communication research. I so appreciate you making yourself available to me as a mentor, professor, and social media expert. I did not even know what social network analysis was before taking your graduate seminar on the topic. Thank you for supporting me at every step of this process and for sharing your important insights on my work. I would be remiss if I did not acknowledge the critical importance of my mentor Dr. Deanna Kepka and her encouraging me to apply for entrance to the PhD program at the University of Utah’s College of Nursing. Deanna, you have been a role model to me over the years, not only because you are an absolutely fantastic rock star researcher and academic, but also because you are a loving mother, wife, daughter, and friend. When you became a young adult caregiver I watched you carefully balance the many roles and responsibilities you were met with, always with courage and compassion. You continue to inspire me and so many others! Thank you for the countless opportunities you facilitated for me as an aspiring researcher, which have been critical in my success as a doctoral student. I am so grateful to my first professional mentor, Reilly Larsen. Reilly, you taught xvi ii me so much about life when we worked together at Old Navy. Your spunky attitude and people skills always impressed me. As I tried to emulate these characteristics I learned leadership, gumption, and grit from you. You taught me how to be efficient and focus on ‘the most important thing’. You gave me a community of support that built my confidence, which I leaned on heavily to complete this degree. Thank you for always being my friend. I am extremely grateful to Austin Waters for his commitment to our recruitment process, his thoughtful insights during qualitative coding and social media data transformation, and for his extremely careful attention to detail in helping me format this final product. I am also very thankful to Taylor Nelson who was integral in completing the social media data manipulation that made it possible to conduct analyses for Aims 2 and 3. Thank you to the Huntsman-Intermountain Adolescent and Young Adult Oncology Program (HIAYA) and the patient navigators, Sara Salmon, Tomoko Tsukamoto, and Nicole Rey for your support in my recruitment and study conduct. Thank you also to Samantha Pannier and Sara Shaw for recruitment assistance, and to the HIAYA Patient and Family Advisory Board for your suggestions for my recruitment approach, materials, and study relevance. Thank you to the nannies, babysitters, and daycares who provided loving care to Cyrus and Ophelia so I could have dedicated time to plan, conduct, and write this dissertation. My deepest heartfelt thanks to Linda Bruno, Akeina Charles, Kayla DeWittBrown, Shentelle Lundeen, Chantalle Brown, Christinana Michelli, Jocee Hospodarsky, Marina Xinogalos, Melanie Speckman, Bailey Weiss, and all the teachers at Baby Steps Daycare & Preschool, and Ina Kinder Care. xvii ii Thank you to the Zogozogos, Annie, Jono, Marc-Aurel, Marie, Monica, and Shelley, for being my partners, peers, and colleagues in this journey. Thank you to the hosts of #phdchat and #phdmom groups and @AcademicChatter on Twitter, for supporting me throughout this process. Thank you to the National Center for Faculty Development & Diversity for many sessions of the 14-Day Writing Challenge and the Dissertation Success Program, during which I experienced my most productive writing. I also acknowledge financial support from multiple scholarships, fellowships, research and travel grants I received during the conduct of this work including the Region 6 Geographic Management of Cancer Health Disparities Program, the University of Utah College of Nursing, Huntsman Cancer Institute, Huntsman Cancer Foundation, Huntsman-Intermountain Adolescent and Young Adult Oncology Program, Marriott Daughters Foundation, the Jonas Nurse Leaders Scholar Program, and the National Cancer Institute under Award Number F31CA221000. The content herein is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Thank you to Christine Baczek, Faelene Larson, and Kristen Bayles for your administrative support with my F31 award. Thank you to my brother Leslie Walker for our Wednesday morning phone calls over the past three years. It has always been my goal to be at least as successful in education and life as you are. Even though our professions are very different, our experiences as first-generation college students are shared. Thank you for listening to and guiding me through this process, for distracting me when I needed it, and helping me to refocus when I needed that. Thank you to your wonderful wife Misty for loving and caring for me when my health was poor, and for my babies when I needed childcare at a xviii ii conference, support during my proposal defense, and the countless times I visited home and needed to run an analysis, write an abstract, or respond to email. To my husband Zachary, I would not have completed this degree, and probably would not have even begun without your encouragement. Words cannot convey the depth of gratitude I have for how you have understood and fulfilled my needs over the past five and a half years of my doctoral study. Thank you, and I love you. Lastly, thank you to the many, many courageous young adults who put their loved one’s needs before their own as they take on caregiving roles for cancer patients. This work would not have been possible without your willingness to share your experiences and social media data with me. xix CHAPTER 1 INTRODUCTION Problem Statement There are an estimated 20.9 million young adult caregivers in the United States (U.S.); 1.46 million provide care to a loved one with cancer.1 In general, younger adults ages 18-25 are more likely to provide care to someone who is two generations older, a female, and more likely a grandmother than mother.2 Adults ages 26-64 commonly provide care to someone in their parents’ generation.2 Caregiving may be uniquely burdensome to young adult cancer caregivers (YACC) because they are often unfamiliar with severe illness, have multiple caregiving responsibilities, and simultaneously experience developmental transitions of young adulthood. These transitions include milestones like completing education, establishing a career, developing intimate relationships, and attaining financial stability.3 Given the highly formative developmental years of young adulthood, social support may contribute to wellbeing differently for YACC than older caregivers. For example, young adults actively use social media to mobilize social support through broad social networks.4 Upon becoming cancer caregivers, they may turn to online support during the first six months, an acute period of transition for the cancer patient and YACC, to help them balance their new caregiving 2 role with existing responsibilities. Social support has been shown to mediate caregiver burden among those caring for patients with functional impairment,5 and is negatively related to depression, loneliness, and burden among cancer caregivers.5-8 Likewise, social support has protective effects on health,9-11 and may shield YACC from becoming overburdened when they find themselves in an unexpected caregiving role.12 However, YACC may have more unmet social support needs than older caregivers because their support networks often lack experience dealing with severe illnesses.13 Young adult cancer organizations across the U.S. have acknowledged the potential for social media communities to provide social support to YACC, and emphasize the acute need for social media interventions that improve support for YACC.14, 15 Yet, there are no existing studies of YACC that have investigated the process of gaining social support through social media, the extent to which social support is present on social media, and the change in social support over time. The aging U.S. population reflects a growing cohort of YACC and underscores the urgency to study social support on social media among YACC. Significance Studying social support on social media among YACC is an immediate research priority because young adults are increasingly adopting caregiving roles for an aging population.16 Caring for a loved one with cancer is not a normative task for YACC, making them more vulnerable to caregiving stress than older caregivers. For example, similar to older caregivers an unforeseen caregiving role may disrupt YACC wellbeing,17 3 and prompt fear of abandonment and loss of their loved one,12, 18, 19 but YACC experience these challenges at a time of life when they may be less established, have fewer resources to draw from, and have less stable social support that typically accumulates with age. Compared to older cancer caregivers, YACC report greater stress and depression during the first six months of caregiving, and the quality of their relationship with the cancer patient influences how stress and depression changes over this initial transition period.18 Also, YACC are more likely than older caregivers to express high levels of unmet needs within the first six months of diagnosis and commonly report unmet information needs related to their caregiving role.20 Moreover, key developmental milestones (e.g., emotional independence and financial stability) may be delayed when YACC put their aspirations aside to care for a loved one with cancer.21, 22 While caregiving can create feelings of closeness and positive memories,21, 23 it may limit YACC social engagement and lead to feelings of isolation and depression.19, 21 On the other hand, longstanding evidence indicates that cancer caregivers’ depression, loneliness, and burden are mediated by social support.5-7 For YACC, becoming a caregiver requires integrating new duties with existing responsibilities (e.g., caring for young children, completing education). Having social support during the acute transition period of the first six months of caregiving may help to ease the burden of a new caregiving role,24 but is not always achieved or sustained over time. YACC who engage social media networks for social support may be shielded from becoming overburdened in their new caregiving role.3, 12 Widespread use of social media among young adults demands recognition of a changing social world in the 21st century and underscores the necessity of engaging 4 existing online networks to support YACC. In 2016, Kent et al. emphasized the importance of studying social media to advance cancer caregiving research.14 This call to action established a precedent for studying how YACC use social media for social support. In 2014, approximately 89% of young adults in the U.S. used social media.25 YACC may use social media to engage with social networks, access resources, and connect with other caregivers. By using social media, individuals gain social support, garner feelings of trust, and establish and maintain close relationships.26, 27 For YACC, online connections may be preferred to traditional services, such as formal support groups.17, 21 In response to the recent emergence of cancer related discourse on social media, this research harnesses the advent of user-generated online data to extend beyond traditional methods of studying social processes by using social media data to study social support for YACC.27, 28 Social support on social media during the first six months of cancer caregiving has not been previously studied among YACC. Social media is likely an important source of support for this growing and potentially isolated community of caregivers. The immediate need for this work is demonstrated through multiple nationwide appeals for research that advances cancer caregiving and adolescent and young adult cancer science.14, 15, 17 Yet, there has been very little progress in either domain to improve supportive services for YACC; most research focuses on patients and older caregivers. By studying social support through social media with a specific focus on YACC, this research addresses a critical gap in both cancer caregiving and young adult cancer research. This research is a crucial first step toward understanding how YACC and 5 caregivers of patients with other diseases use social media for social support, which is required to inform future innovative technology-based interventions. Specific Aims, Hypotheses, and Research Questions Building upon prior work in adolescent and young adult cancer patient social wellbeing, this research uses a theory-based approach to study social support through social media among YACC during the first six months of caregiving. A convergent mixed methods design is applied.29 A six-month timeframe after diagnosis was selected to develop focused knowledge of YACC supportive needs during the acute period of transition to a new caregiving role. Generally speaking, there is evidence that social support diminishes over the first six months of cancer caregiving.18 It is likely that social support on social media among YACC also diminishes over this period, potentially because YACC use social media to seek social support less often or because their social networks become less responsive over time. YACC may use social media less often because the process of writing about their caregiving experiences and sharing them on social media may be difficult and stressful,30 or because their supportive needs are met by offline resources. On the other hand, their social networks may become less responsive over time because young adults lack experience coping with severe illnesses, many online relationships consist of “weak” ties (i.e., individuals come and go as they please) which may not provide as robust social support, or because YACC have not met those they seek support from in person, and may never meet or interact in the future, and so the social responsibility to respond may be diminished over time.31 This study answers these questions through three aims. First, a stratified 6 purposive sample of N=34 YACC was recruited for qualitative interviewing to explore the process of using social media for social support (Aim 1). Wang et al. revealed that different techniques are used in online cancer communities to facilitate emotional (disclosing fear, sadness, anger) and information (asking questions) support,32 but this study did not describe other types of social support and was not focused on YACC. Therefore, Aim 2 expands prior research by creating and analyzing a large dataset of YACC social media posts for evidence of five functions of social support (emotional, information, instrumental, companionship, validation) in exchanges between YACC’s social media posts and responses to their posts. In addition, it may be the case that specific individuals, such as those who are female, younger age, or of a racial/ethnic minority, may be able to sustain social support more evenly over the first six months of caregiving because they are more likely to use social media than their counterparts.33 This is important because caregivers who are female, younger, and of race/ethnic minority tend to have poorer psychological outcomes.18, 34 Moreover, certain sociodemographic groups may seek different types of functional social support during the first six months of caregiving. For example, caregivers with higher socioeconomic status may seek to sustain emotional support at a high level during the first six months, whereas a caregiver with lower socioeconomic support may seek to sustain a high level of instrumental support to meet the financial burdens that may accompany the caregiving experience, but this has not been previously studied on social media. The purpose of Aim 3 is to contribute new knowledge of social media and YACC characteristics that are associated with social support during the first six months of caregiving. 7 The mixed methods design of this project is pivotal for understanding the results about YACC’s social support through social media. Specifically; Aim 1 provides crucial information about how YACC seek social support to contextualize the interpretation of social media data for Aims 2-3. In Aim 2, content analysis was used to quantitize textual social media posts and responses to identify the types and prevalence of five types of functional social support that is exchanged between YACC and their followers during the first six months of caregiving. A longitudinal dataset comprised of a random sample of 10% of YACC’s social media posts made during the first six months of caregiving was analyzed in Aim 3. The social support outcome variables from Aim 2 are transformed to depict the percent of responses to each social media post that contained each of the five types of social support by social media platform type (Facebook vs. Instagram) and cancer-related content (No vs. Yes) using linear mixed effects models (Aim 3). This mixed method approach comprehensively answers the overarching research question: How is social media used for social support by YACC, and how does social support on social media change during the first six months of caregiving? Aim 1: Explore and describe the process of social media use for social support among YACC. Aim 2: Identify the type, prevalence, and changes in functional social support exchanges between YACC and their followers on social media. H1: Emotional and information support will be more prevalent than other types of functional social support. H2: All types of functional support will diminish in prevalence over time. Aim 3: Identify possible effects on functional social support among YACC. 8 RQ: How does platform type influence the distribution of social support contained in responses to YACC’s posts? H1: Social media posts with cancer-related content will contain more social support compared to posts without cancer-related content. Innovation and Contribution In sum, studying social support through social media among YACC during the first six months of caregiving fills a critical knowledge gap and growing phenomenon in cancer caregiving and young adult cancer research. Despite the growing population of YACC and widespread use of social media among young adults, few studies have investigated the prevalence of social support on social media among cancer caregivers, and none have targeted the experience of young adults. This research is needed to improve supportive services for YACC. Furthermore, although social support declines over the first six months of caregiving,24, 35 to date no studies have assessed the change in social support on social media during this acute transitional period. This research contributes crucial information that is necessary for developing technology-based social media interventions to improve social support for YACC, and it directly aligns with the National Cancer Institute’s request for exploratory research on cancer caregiver wellbeing.36 Specifically, this work expands existing literature by addressing three critical objectives in YACC research: evaluating the feasibility of studying social support on social media, defining the role of social media in promoting functional social support, and informing future social media interventions to help YACC leverage social media for the functional social support they 9 need to fulfill their caregiving role. The impact of this work may extend to young adult caregivers of other disease states by contributing new information for supporting young adult caregivers through social media research and clinical supportive services. CHAPTER 2 YOUNG ADULT CANCER CAREGIVERS SOCIAL SUPPORT THROUGH SOCIAL MEDIA: INTEGRATED REVIEW AND CONCEPTUAL FRAMEWORK Introduction This chapter reviews the literature that characterizes YACC, their social support, and their use of social media. Definitions related to social support, social media, and YACC are provided in Table 2.1 and examples of these terms on social media are demonstrated in Figure 2.1. The developmental stages of young adulthood are described and defined from the perspective of YACC roles. Lastly, the integration of the Cancer Family Caregiver Experience Model with the Stress and Coping Social Support Theory is described. For this study, the Cancer Family Caregiver Experience Model situates this research within the context of the broader cancer caregiving experience, and the Stress and Coping Social Support Theory provides the definitions of five different functions of social support, which are the primary outcomes of Aims 2-3. Young Adult Cancer Caregivers YACC are individuals ages 18-39,2 who care for a spouse, partner, friend, sibling, parent or grandparent who has been diagnosed with cancer.16, 21 There are approximately 11 one and a half million YACC in the U.S.1 Distinct differences exist between young adult and older caregivers. For example, the average time spent caregiving among young adults is 5.8 years, significantly longer than that of older adult caregivers at 3.7 years.16 The longer-lasting caregiving relationship among younger caregivers suggests that providing care may be more burdensome for them than older adults, because it requires sustained attention over a longer time period, occurs during a transitional developmental stage, and limits social engagement,21 which is a cornerstone of young adulthood.22, 37, 38 YACC may establish social networks during the initial transition to a caregiving role, which set the stage for their future caregiving relationship, how they balance existing responsibilities, and how they use social media for social support. Caring for a loved one with cancer is not a normative task for YACC, and taking on this unexpected role makes them more vulnerable to caregiving stress than older caregivers. Similar to older caregivers, an unforeseen caregiving role may disrupt YACC wellbeing because it prompts fear of abandonment and loss of their loved one.12, 18, 19 Unique to YACC, becoming a caregiver requires integrating new duties with existing responsibilities of young adulthood such as completing education, starting new careers, and caring for young children. Furthermore, key developmental milestones (e.g., emotional independence and financial stability) may be delayed when YACC put their own aspirations aside to care for a loved one with cancer.21, 22 While providing informal care can create feelings of closeness and positive memories,21, 23 it may limit YACC social engagement and lead to feelings of isolation and depression.19, 21, 23 Nonetheless, longstanding evidence indicates that cancer caregivers’ depression, loneliness, and burden is mediated by social support.5-7, 35 12 Having social support during the acute transition period of the first six months of caregiving may help to ease the burden of a new caregiving role, but social support is not always achieved or sustained over time.24 The current literature is limited, but does point to the heightened vulnerability of YACC to experience caregiving stress and depression compared to older caregivers, especially during the first six months of caregiving.18 This susceptibility of YACC could be due, in part, to their high level of unmet supportive needs (e.g., information, emotional, social needs).20 Likewise, prior research suggests that individuals join online communities primarily for information, friendship, and social support,39 which includes the two most commonly cited unmet supportive needs among YACC. Therefore, when YACC find themselves lacking information and emotional support, they may reach out to online communities on social media to fulfill these needs.32, 39 The broad application of using social media for social support during stressful times underscores the relevance of studying how YACC obtain online social support as they adapt to a new role during the first six months of their caregiving experience. However, social media has been underutilized as a lens for studying social support among cancer caregivers.40 This research is the first step toward improving our understanding of how YACC use social media for social support. Overall, YACC are an understudied population with divergent social support needs compared to older caregivers. In the last two decades, the paucity of information on YACC supportive care in comparison to pediatric or adult cancer has been recognized, and has resulted in a national prioritization of young adult cancer research.14, 15, 17 These calls to action accentuate the need to study innovative approaches that advance cancer 13 caregiving research among YACC. The knowledge generated through this work is essential for developing technology-based interventions to assist YACC with sustaining social support as they take on a new caregiving role. Developmental Context of Young Adulthood Young adulthood encompasses the distinct developmental periods of emerging adulthood (approximately ages 18-25 years) and young adulthood (approximately ages 26-39 years),38 during which social support may vary considerably. The heterogeneity of social development during young adulthood necessitates research designs that consider dynamic social support needs among YACC. For example, emerging adults who value change and exploration may feel restricted by caregiving responsibilities, whereas young adults who see themselves as more established in their world-views, occupation, and family life may feel displaced by unexpected caregiving expenses.37, 38 Additionally, when YACC seek social support through social media, their social networks may be unresponsive or provide undesired types of support because they lack experience coping with severe illness, are unfamiliar with caregiving demands, and have yet to attain social stability that is typically achieved with age.13 Given the complexity of this period, YACC may find themselves in a situation where they are at risk of becoming isolated or overburdened and have to put aside goals and plans for the future to fulfill their caregiving role. There is an immediate need to support YACC across these developmental periods, but existing caregiving research has not addressed this need. The following sections provide a detailed description of how the developmental periods of 14 emerging adulthood and young adulthood inform YACC research studying social support on social media. Emerging Adulthood Emerging adults 18-25 years of age are in a period of exploration and discovery.38 They are no longer adolescents, but they have typically not yet achieved independence nor fully transitioned into adult roles in their employment, education, or family life. During emerging adulthood, individuals aim to establish themselves financially, socially, and emotionally separate from others. Emerging adults explore possible career paths and experiment with new relationships in an attempt to define their own identity.38 This period may be disrupted when an emerging adult becomes the primary caregiver to a cancer patient. Caring for a loved one with cancer may limit an emerging adult’s ability to socialize with others, develop romantic relationships, and form selfidentity. Fulfilling caregiving responsibilities may leave emerging adults feeling restricted and isolated. Caregiving may also complicate interpersonal relationships with the cancer patient.21 For example, an emerging adult who is providing care for a grandparent with cancer may experience social and emotional growth as they strengthen the bond with their loved one. In contrast, emerging adults who become the primary caregiver to a parent with cancer may experience resentment and guilt from “role reversal,” when they perform caregiving tasks for a parent who was recently their primary carer.21, 41 Whereas an emerging adult caregiver had previously been in the process of establishing themselves separately from their parents, taking on a caregiving role of this nature may create conflict.41 15 In the event that an emerging adult becomes a cancer caregiver, they may reach out to social networks through social media to express emotions, ask questions, and seek support. Identity formation is a key component of emerging adulthood, and social media is often used as a channel to explore different identities.42 When a caregiving role limits social engagement, social media may serve as a platform for YACC to establish their own world views and experiment with different social identities without compromising their ability to fulfill caregiving actions. Social issues and milestones that may be disrupted by an emerging adult taking on a cancer caregiving role are described in Table 2.2. Young Adulthood Establishing an adult role in the workplace, social circles, and family life are markers of young adulthood between the ages of 26-39.22, 37 During young adulthood, YACC are motivated to care for an older individual, particularly an older family member, either out of obligation or attachment.21 Yet, in a new caregiving role, YACC may fail to establish emotional independence and sustained employment21 which are key cornerstones of young adulthood.22 Moreover, while young adult caregivers experience positive outcomes of caregiving (e.g., closeness, positive memories, avoiding nursing home placement), caregiving responsibilities can limit social engagement with family (e.g., spouse, partner, children), friends, and social networks.21 Of particular importance during young adulthood, the responsibilities of cancer caregiving may affect employment in many ways, such as taking time off, reducing full-time employment to part-time, and being distracted or unable to work.43 16 YACC become caregivers during a time when they are often caring for others including young children or aging parents.44 Whether the cancer care recipient is a spouse, parent, grandparent or friend, the additional role of caring for a loved one with cancer may be burdensome. Establishing robust social support networks during the initial transition to a caregiving role, the first six months after the cancer diagnosis is received, may help YACC adjust to having multiple caregiving roles. While there are distinct developmental differences between emerging and young adults, they share a common behavior in their high prevalence of social media use.33 What has been referred to as the millennial and Z generations includes those born from 1981-1996 and 1997 onward, respectively.45 In 2016, these individuals represent the current cohort of emerging and young adults ages 18-35.46 Social media use is prevalent among millennials, with approximately 90% using at least one social media platform daily.33, 47 As high users of social media, young adults may rely more heavily on online interactions for social support regarding personal matters than individuals of other ages.33, 47 While limited, there is evidence that social media use may be more prevalent among racial/ethnic minorities,33 and thus this is a parameter of interest in the proposed research. Specific social issues and milestones that may be disrupted by a young adult taking on a cancer caregiving role are described in Table 2.2. Social Support in Online Communities The growing popularity of social media use among young adults provides the opportunity to support YACC in creative ways.26, 33 Social media includes popular social networking sites such as Facebook, Instagram, and Twitter. YACC who engage existing 17 online networks for social support may be shielded from becoming overburdened in their new caregiving role.3, 12 While the nature of social support has yet to be explored in YACC interactions on social media, previous research indicates that using social media may improve social support, help YACC garner feelings of trust, and help users establish and maintain close relationships.26, 27 Individuals also use social media to share personal health information, request guidance or feedback about disease, and receive information and emotional support.48, 49 Studying social support in online communities is important because online communication may differ from face-to-face interactions, especially because nonverbal cues (e.g., nodding in agreement, personal touch etc.) are absent in online interactions.31 Online support may benefit coping differently than face-to-face support.30 One reason for this may be that being physically separated during an interaction online provides an opportunity for deeper self-reflection by social media users.30 Meaning that, as individuals post about their cancer caregiving experiences they may become more attuned to caregiving demands and stress. Reduction in nonverbal cues on social media may be addressed through support matching, similar to language-style matching, in which support messages are aligned with the type of support requested.30, 50, 51 For example, individuals who post seeking social support for an uncontrollable situation (e.g., terminal cancer diagnosis) benefit more from emotional support than information support, whereas when an individual who posts seeking support for a controllable situation (e.g., scheduling an appointment) benefits more from information support.30 When online networks are responsive to the type of support and language that is used in a post, users feel more confident in the 18 network’s ability to provide emotional and information support.50 The first step in examining support matching among YACC on social media is to study the types, prevalence, and change in social support on social media during the first six months of caregiving. When it comes to online cancer communities, little research has occurred studying the types and prevalence of social support online and how social support is maintained over time. However, a growing body of literature supports the further investigation of social support on social media among cancer caregivers. In 2015, Badr interviewed cancer caregivers, finding that 84% expressed interest in using social media as a resource for dealing with cancer, with 54% desiring emotional support and 77% endorsing information support.40 Cancer caregivers also support the use of social media to help coordinate care and avoid conflicts with the cancer patient.40 While not focused explicitly on caregivers, Twitter posts related to young adult cancer included a range of emotional expressions such as hope, sadness, fear, humor, anger.27 Tweets with detailed experiences, religious references, and empathy were the most likely to be shared or liked, which are indications of social support, whereas those expressing sadness and anger were less likely to be shared.27 Moreover, users who seek emotional support tend to remain involved in online social communities longer than those who seek information support.32 In contrast, there may be unintended negative consequences of using social media for social support as a cancer caregiver. Optimism and positivity are cornerstones of online cancer communities.27 Yet, social media posts that are interpreted to be negative may not be offered support in the same way as more positive posts. Emerging adults who encounter negative social media posts may be uncertain about how to respond because 19 they are inexperienced in dealing with severe illness and are unfamiliar with caregiving demands.38 This uncertainty may lead them to respond in an unsupportive manner, provide an incongruent response (e.g., a mismatch of support, like offering emotional support when information support was sought), or not respond at all. Another potential negative consequence of online social interactions may be the sharing of clinically inaccurate information on social media. When a YACC seeks support through social media they may encounter health information that is of low quality or inaccurate information, estimated to be as high as 70% of online health information sources.52, 53 Similarly, social support that is unsolicited may be detrimental to coping,54 yet as many as two-thirds of social media posts may contain unsolicited advice about disease management.48 Unintended negative consequences of seeking social support through social media (e.g., negative support, incongruence, absence of response, and inaccurate information, and unsolicited advice) are examined to enhance and expand the analysis and contextualize the negative consequences of using social media for social support as a YACC. Furthermore, it is possible that YACCs with certain characteristics may be more likely to receive different types of social support as well as sustain or lose social support over time. Previously, factors such as type of cancer diagnosis, cancer treatment, hospital admission (e.g., public vs. private hospital), and diagnosis stage have been identified as mediators of caregivers’ unmet needs, and are also likely to influence their social support on social media.20 More female social media users may provide emotional support than male users.49 Factors that have shown direct negative associations with psychological adjustment included simultaneously caring for a child and being unemployed.55 20 Furthermore, sociodemographic characteristics that are specifically related to social support include age, attachment style, and Facebook use.4, 18 Given the paucity of research studying how sociodemographic characteristics influence social support on social media it is imperative that future researchers collect and analyze variables like gender, race/ethnicity, and income that are associated with heightened caregiver burden among general caregiver populations.34 While not all of these factors are considered in the current analysis, these are important areas for future exploration. Theoretical Framework and Models Two perspectives guide this research. First, the Cancer Family Caregiving Experience Model posits three elements of central importance to the informal caregiving experience: context, stress process, and cancer trajectory (Figure 2.2).35 There are multiple points during the stress process in which social support among YACC using social media may be affected. In the model, primary stressors initiate the stress process, including things like decreased social support, poor coping, and patient-related illness factors such as disease site, stage, prognosis and duration, symptoms, functional dependency.35 To address complex primary stressors, YACC are required to meet caregiving demands, which include managing complex medications, navigating the healthcare system, and reducing symptom burden.56, 57 Secondary stressors arise as caregiving demands begin to affect other aspects of life. Changes in relationships is a top secondary stressor and includes loss of social support and feelings of isolation.58-61 For YACC, secondary stressors may arise as caregiving demands begin to affect their social interactions, romantic relationships, parenting responsibilities, employment, and 21 education. However, the majority of research on secondary stressors has focused on older adult caregivers, and does not consider how caregiving demands might be different for YACC. As primary and secondary stressors begin to weigh on YACC they start to appraise their situation. Appraisal is an evaluative period during which caregivers decide what the stressors they are dealing with mean for them and they negotiate how to cope with each stressor.35 Appraisals can be both positive (e.g., confidence, preparedness) and negative (e.g., uncertainty, hopelessness).35, 62-64 While the appraisal of caregiving stressors occurs on an individual basis, this process may be heavily influenced by YACC social networks, but this has not been explored in existing literature. For example, YACC may try to compare their experience to others they perceive are similar to them using social media. As YACC seek support on social media, they may use this process as an opportunity to appraise their situation in terms of their caregiving role, family and work responsibilities, their social life, and their goals, compared to others they interact with on social media. This type of appraisal may have benefits, such as connecting individuals with other YACC caregivers who can validate their experiences, or it may have negative consequences if YACC appraise their situation to be less optimal in comparison to others in their social network. In response to appraisals, a cognitive-behavioral response occurs, which can be likened to the coping process, a key component of the Stress and Coping Social Support Theory, described hereafter. Examples of appraisal among cancer caregivers have been widely studied and include social comparison, support seeking, emotional expression, venting, and disengagement.65, 66 Similarly, using social media is a common way in 22 which YACC perform these same actions of social comparison, support seeking, and mood expressions. For example, a YACC might post the following on Facebook with an accompanying “frustrated” emotion tag, “Does anyone else ever feel like they don’t have enough time for themselves?” In this example, the YACC may be seeking validation support by asking a question, intending to elicit answers that will allow them to compare themselves to others who are like them, and expressing frustration with their current situation. The extent to which YACC use social media for support by engaging in these coping mechanisms is unknown. Finally, the stress process influences health and wellbeing. While positive outcomes such as increased life satisfaction and sense of meaning can occur during a caregiving experience, negative outcomes such as depression, emotional disturbance, and other poor mental health outcomes are common among cancer caregivers.35, 56 Previous literature suggests that YACC in particular experience a higher level of negative mental health outcomes (e.g., depression, anxiety) compared to older caregivers.18, 20, 59 In summary, the model provides a framework for comprehending that social support among YACC using social media may vary at different stages of the stress process and that social media is also an important but previously undescribed feature of the broader sociocultural context in which caregivers exist. It is likely that social support diminishes over time as support networks become more accustomed to the caregiving situation,35 and this can be stressful for YACC as it affects their health and wellbeing. To date, the majority of research has focused on older cancer caregivers and almost none has studied how social media use influences the stress process. While this model is foundational for the study of cancer caregiving, it does not fully encapsulate how social 23 support changes over the initial period of transition to a caregiving role among YACC. In other words, the heterogeneity of social development during young adulthood is not specifically explicated in the Cancer Family Caregiving Experience Model.37, 38 The Stress and Coping Social Support Theory provides further guidance and defines social support for this study.67, 68 Social support is a concept that has been widely researched over the past few decades and is a topic of interest among those studying the cancer caregiving experience. There are two different types of social support described in the literature: structural and functional support. Structural support refers to the people and relationships that create a connected social network. In contrast, functional support refers to the specific actions that are provided by members of a social network.9 Functional support is further divided into either perceived support or received/enacted support.9 While perceived support relies on a subjective judgement of support by the person seeking support, received/enacted support relies on the supportive actions that are offered by others to the person seeking support. This study focuses on the latter form of functional support, received/enacted support, because the retrospective nature of the study may compromise the validity of participants’ subjective interpretations regarding how they perceived the responses to their social media posts. This theory posits five different types of received/enacted actions that individuals perform to show support and help relieve stress; which are termed functional social support (Table 2.3).67, 69 Each type of functional social support addresses different social needs.67, 69 However, when functional social support originates from inappropriate sources or in undesired forms it may be detrimental to coping.67, 69 Similar to the cognitive-behavioral response process that occurs in the previous model, 24 this theory postulates that social support enhances coping, which mediates the relationship between stress and health (Figure 2.3).67, 69 Building upon this framework, having social support through social media may help YACC facilitate specific types of social support to address different needs (e.g., sympathy, child care, advice). As these social needs are met through social media interactions, YACC’s ability to balance a new caregiving role is improved. In the analyses for Aims 2-3, the functional social support definitions were used to code social media data. This theory further informs the overall premise for this study by postulating that social support enhances coping, thereby reducing caregiver stress and improving caregiver health. While this perspective explains how social support can be beneficial to caregiver health it does not account for changes in social support over time. The Social Support and Coping Theory explains how the social support on social media may enhance coping and improve health outcomes among YACC. Furthermore, future research that evaluates how social support from social media affects coping with a caregiving role among YACC is essential for designing future social media interventions, but is beyond the scope of this study. Integration and Limitations of Caregiver Model and Social Support Theory Studying social support on social media among YACC is best understood by blending the Cancer Family Caregiving Experience Model and the Stress and Coping Social Support Theory. Precisely, the model posits a caregiving stress process that provides the conceptual foundation for understanding the position of social support in the 25 overall context of the caregiving experience. The theory provides the theoretical underpinnings and the social support definitions used in this study. These perspectives inform the current project and pave the way for future research by explaining the points during the caregiving stress process that are most amenable to intervention: the appraisal process and the cognitive-behavior response.35 Acknowledging some limitations of these perspectives positions this research within the caregiving, social support, and social media research fields. While the Cancer Family Caregiving Experience Model provides a comprehensive conceptual guide for situating this research within the overall cancer caregiving experience, it may not satisfy or explain the varying needs of YACC because it does not describe how social contexts and the stress process differ for caregivers at various ages and developmental stages of life. Further, it does not acknowledge the robust online social interactions that exist among cancer caregivers on social media or postulate about how these interactions influence the caregiving experience. While the Stress and Coping Social Support Theory has been widely used to study social support on social media, there are no existing applications of this theory to a young adult cancer caregiving context specifically,70 meaning that, while this model provides sound theoretical definitions of the functions of social support and informs the interpretation of social support outcomes on caregiver wellbeing, it may not fully explain how social support on social media benefits YACC’s wellbeing.70 Additionally, functional social support is a form of “enacted” social support that guides the definition of important terms for this study including seeking and offering, but these definitions do not explain how social support actions are subjectively interpreted by the YACC on 26 social media nor do they necessarily capture the intention of the responder, and this is noted as a limitation. Conclusion In conclusion, this research addresses a critical gap and growing phenomenon of YACC research with a defined focus on social support and social media. To further the field of cancer caregiving research, and fully support YACC, there is an immediate need to understand the parameters of social support in social media for YACC. The results explain how social media is used for social support by YACC. 27 Table 2.1 Definitions Social media Social media are the tools used to engage with an online community. For the purposes of this study, the term social media encompasses social networking sites like Facebook and Instagram. Young adult An individual aged 18-39 years who has the main cancer caregiver responsibilities of informal patient care for a loved one with (YACC) cancer. A YACC may be a family member, friend, co-worker, or neighbor of a cancer patient. Functional social Functional support types include emotional, information, support instrumental, companionship, validation. Functional social support may be supportive, neutral, or adverse to the YACC’s ability to cope with a stressor. Seek When a YACC makes a post on social media in an attempt to elicit an action from their social network members Offer When a member of the YACC’s social network interacts with a post on social media by commenting, “liking”, sharing, mentioning, or responding Exchange An original post or response shared between a YACC and a member or members of their social network. Exchanges are the focus of Aim 2 analyses. Response A reply made on a YACC’s Facebook or Instagram post. Replies are the focus of Aim 3 analyses. Table 2.2 Specific social issues and milestones for YACC by age Stage Age Social Social issues/milestones disrupted by relationships a YACC role Emerging 18-25 Friends, • Role reversal between YACC and cancer adulthood years school and patient (e.g., parent/child) work peers, • Delays or gaps in higher education parents • Work/employment interruptions • Barriers to adequate health insurance coverage • Hindered achievement or maintenance of financial independence from parents Young 26-39 Partners, • Work/employment interruptions adulthood years family, • Difficulty developing and/or maintaining friends, intimate partner/spouse relationships school and • Difficulty fulfilling multiple caregiving work peers roles (e.g., young children) • Difficulty achieving/maintaining financial independence from parents • Geographic distance from support system 28 Table 2.3 Functional social support definitions and social media examples67 Emotional Sympathy, caring, Comments including: “You are acceptance amazing!” “Our hearts are with you” Instrumental Transportation, household Links to scholarships, fundraising. chores, child care, finance Posts seeking childcare, transportation. Information Knowledge, information, Links to articles/videos related to advice, alternative action caregiving or cancer questions. Companionship Availability of persons to Links to YACC support groups, spend time with chats, or events. Posts planning activities with others. Validation Feedback, social Comments such as “My mother comparison was diagnosed with breast cancer, too”. Figure 2.1: Examples of definitions of terms 29 Figure 2.2: Cancer Family Caregiving Experience Model. Used with © permission from Fletcher, 2012. Figure 2.3: Stress and Coping Social Support Theory. Adapted from Cohen, 2000 CHAPTER 3 RESEARCH DESIGN AND METHODS Introduction The following chapter provides a description of the approach, study design, methods, data collection, statistical analysis, and mixed methods integration performed in this study. The chapter begins with a summary of the pilot study that informed this proposal. Following is a description of the mixed-methods study design and participant eligibility, inclusion and exclusion criteria, the consent process, sampling, and targeted enrollment. Next, a description of the qualitative methods, data analysis, and limitations for Aim 1 are described. This is followed by a description of the quantitative methods, statistical analysis, and limitations for Aims 2-3. A summary of the mixed methods integration follows, including how the qualitative and quantitative aims were collectively performed, interpreted and disseminated. The chapter closes with a preliminary overview of the three results chapters. Pilot Study Summary The purpose of the pilot study was to explore the feasibility of analyzing social support through social media posts related to young adult cancer. In this cross-sectional 31 study, we content analyzed Instagram posts about young adult cancer. During the study, methods and a protocol were developed and refined for assigning latent and manifest codes to each social media data observation (i.e., an Instagram post). In addition, the feasibility of using content analysis software was assessed, and a customized coding dictionary was created in Yoshikoder to categorize text from Instagram posts according to specific definitions related to social support (e.g., positivity, negativity of terms). The methods, results and limitations of the pilot are reviewed next. First, a coding schema was iteratively developed and applied to categorize text from a random sample (N=50) of public Instagram posts that were tagged with the hashtag #youngadultcancer. The text was compiled into a single dataset and later separated according to treatment status (active treatment vs. survivorship) and type of user (individual vs. organization (e.g., hospital, business, non-profit organization)). Lexical characteristics (i.e., the proportion of positive, negative, and emotional text) and social support (i.e., the number of likes and comments) were defined based on extant literature, coded for each post, and analyzed in Yoshikoder.71 Lexical characteristics of posts were compared by survivorship status. Posts made by those in survivorship had a higher mean number of likes (54.5, SD=37.8) than posts made by those in active treatment (32.3, SD=16.2, p=0.03). There were more positive than negative terms used by those in survivorship (30 vs. 13 terms, p<0.01) and in active treatment (20 vs. 9 terms, p=0.04). In addition, posts made during survivorship also had more emotional text (79.6%) compared to those made during active treatment (34.9%, p<0.01). In terms of social support, individuals had more comments than organizations (mean 6.0 vs. 2.3, p<0.01), whereas organizations had more likes than 32 individuals (mean 60.3 vs. 26.1, not significant). Likewise, those in survivorship had more likes than those in active treatment (mean 37.8 vs. 16.2, p=0.03). These results demonstrate the feasibility of using social media to study social support among young adults in online cancer communities. Of particular interest is that emotional text was used to communicate about cancer on social media and for seeking social support through social media. These findings likely have implications for YACC, and are particularly important in building the body of literature related to social support on social media among young adults during their cancer experiences. The pilot study revealed some important considerations for future research on social support among YACC. Two specific limitations from the pilot study that are addressed in the current study were 1) the lack of data on personal characteristics (e.g., demographic, caregiver, cancer factors), which limited the ability to understand demographic factors associated with social media and social support, and 2) the rudimentary definition of social support that did not distinguish the types of social support received through social media. Building upon this work, the mixed methods approach described hereafter facilitates the merging of qualitative data collected in Aim 1 with the social media data from Aims 2-3. Additionally, the classification of text according to different functions of social support provides a more accurate depiction of the nature of social support online than simply considering the number of likes and comments on social media posts as was done in the pilot. 33 Mixed Methods Study Design Informed by prior research in adolescent and young adult oncology, cancer caregiving, and social support, a theory-based approach to define social support through social media among YACC during the first six months of caregiving was used. This approach employs a convergent mixed methods design as shown in Figure 3.1. Mixed method study designs stimulate both exploratory (Aim 1) and explanatory (Aim 2-3) opportunities.72, 73 This design ensures the research empowers YACC to describe the process of using social media for social support in their own words (Aim 1, qualitative), while identifying the type, prevalence, and change in social support (Aim 2, quantitative), and modeling possible effects on the change in social support over time (Aim 3, quantitative). To begin, in Aim 1 N=34 purposively sampled YACC were interviewed to explore the process of social media use for social support. Prior to the interview, a brief survey was administered to collect YACC sociodemographics and cancer patient characteristics. After the interviews, all social media posts that were made within the first six months of caregiving were collected. The combined dataset used in Aims 2-3 contained both YACC sociodemographic and cancer patient factors as well as social media data. Wang et al. discovered that different techniques are used in online cancer communities to facilitate emotional (e.g., disclosing fear, sadness, anger) and information (e.g., asking questions) support, the most prevalent types of functional support identified in previous social media research.32, 39 Therefore, in Aim 2 this proposal expands prior research by directly sampling social media posts of YACC to identify the types, prevalence, and change of five types of functional social support (emotional, 34 instrumental, information, companionship, and validation). Social media posts were extracted and manually coded by three research team members according to the definitions of functional social support. Next, the text of social media posts was quantitized into five social support outcome variables which depicted the presence or absence of social support.74 Linear mixed effects regression models were used in Aim 3 to identify differences in the distributions of functional social support by social media platform (Facebook vs. Instagram) and the presence or absence of cancerrelated content (No vs. Yes) over the first six months of caregiving. The qualitative Aim 1 findings refine and explain the results of Aims 2-3 by providing an in-depth description of the process, scope, and boundaries of using social media for social support.73, 75 Findings from this research are synthesized to contribute new knowledge of how social support is sustained over the first six months of caregiving. The results provide a foundation for improving support during the acute transition period to a caregiving role to minimize negative effects of young adult cancer caregiving.24, 76 Participant Eligibility and Consent Participant inclusion and exclusion criteria is listed in Table 3.1. Participants were included who were between six months to five years from the date of the cancer patient’s diagnosis. This timeframe focuses on social support during the initial period when YACC are adjusting to a new cancer diagnosis, a time in which YACC may be particularly vulnerable to social isolation and negative psychological outcomes.20 In addition, those who are caring for a cancer patient with recurrent cancer were excluded because their social support and social media use may differ from those who are experiencing a 35 caregiving role for the first time. One participant who was in a caregiving role for the first time to a patient with recurrent cancer was allowed to participate. The informed consent process occurred with enrolled participants over the telephone prior to the interview. Sampling A stratified purposive sampling approach was used to recruit eligible YACC.75 Based on prior experience recruiting YACC for research,77 and other studies that recruited YACC,21, 23, 78 a three-prong sampling strategy was used to ensure an adequate sample was achieved. First, YACC were recruited through the Huntsman Intermountain Adolescent and Young Adult Oncology Program (HIAYA), an existing program that includes the Huntsman Cancer Institute (HCI), Intermountain Healthcare, and community oncology programs across Utah. Together, HCI and Intermountain Healthcare provide care for over 85% of cancer cases in Utah. HCI has the capacity for this sampling, treating approximately 500 patients ages 15-39 annually. Intermountain Healthcare has a similar number of patients in this age range. Based on national young adult caregiving data, we estimated that approximately 48% of cancer caregivers are in the young adult age range (500 per year in the HIAYA clinics).1 The HIAYA Oncology Program patient navigators reviewed the HIAYA patient database to identify and refer eligible YACC who were willing to be contacted. Because the HIAYA program primarily focuses on young adults, it is likely that most caregivers recruited through HIAYA provide care for a young adult cancer patient. Therefore, the second recruitment strategy expanded and diversified the sample of possible caregiver-patient relationships by recruiting 36 participants through HCI’s social media platforms, which reach over 85,000 individuals. Finally, printed recruitment material was placed in clinical waiting areas, cafeterias, and the Cancer Learning Center Library at HCI, of which approximately 50% of patrons are caregivers. Recruitment occurred from 08/2017-06/2018. Targeted Enrollment Stratified sampling occurred by age to ensure equal representation of YACC in emerging adulthood and young adulthood38 and by gender, because social media use differs between females and males.26, 49 Sampling continued until data saturation was reached,79 and adequate sample size was achieved for power in Aims 2-3. Our original recruitment goal was estimated to be approximately N=40 interviews and recruitment was finalized with N=34 participants (Table 3.2). Power analysis for Aims 2-3 was performed using the Cochran-Armitage test for trend in proportions.80 The estimated sample size of approximately N=14,600 posts (1 post per day x 2 platforms of use x 182.5 days/6 months x 40 participants)47 achieves 99% power to detect a linear trend using a two-sided Z test with continuity correction and a significance level (alpha) of 0.05. Power was calculated using PASS software v.14.0.8. Aim 1 Purpose Aim 1 explored and described how social media is used for social support by YACC. The purpose of Aim 1 was to describe YACC perceptions of the process of using social media during the acute period of transition to their caregiving role, the first six months of caregiving. Aim 1 results provide interpretive context for the results of Aims 37 2-3 and essential information to inform future social media interventions. Aim 1 Data Collection Prior to being interviewed, participants completed a brief survey. The survey included YACC sociodemographics (e.g., age, length and type of caregiver relationship), Social Relationship Index (3 items, α=0.87),81 Duke Social Support and Stress Scale,82 type and frequency of social media use, and cancer patient factors (e.g., diagnosis). Interview topics focused on how and why YACC used social media for social support, social media patterns of use, unintended negative consequences of social support on social media, limits and boundaries of using social media for social support, and a future social media intervention (survey & interview in Appendix A). Interviews were audio recorded, transcribed, and checked for quality. Survey data were entered, cleaned, and managed in REDCap.83 Aim 1 Data Analysis Descriptive statistics (mean, percentage) were generated for the YACC sociodemographic and cancer patient factors from the survey to describe the sample. Data collection and analysis occurred iteratively to provide flexibility and reflexivity in the interpretations of the data. Reflexive and interpretive memos are a key analytical technique in qualitative research,84, 85 and were created throughout the coding process to enhance description and understanding and provided insights into the Aims 2 and 3 analyses. Constant comparative analysis was applied to explore the process of how YACC use of social media for social support, social media patterns of use, and a future 38 social media intervention.86 Two cycles of coding occurred in NVivo 11 (Table 3.3). During first- and secondcycle data analysis, the data were reviewed and emerging codes discussed amongst the coders to gain consensus in how interpretation of the data was developing. First, transcribed interviews were coded via structural coding.87 Structural coding is well suited to mixed-methods research that analyzes transcripts from multiple participants, because it allows the researcher to categorize, compare, and examine relationships among participants. In second cycle coding, a coding scheme was developed via focused coding techniques. All interviews were coded in first and second cycle coding, with a random sample of 10% of the interviews being double coded. The initial kappa score was above the “substantial” range (k=0.61-0.80),88 and the coding scheme was further refined by resolving coding discrepancies, problematic codes, and coding areas with zero or negative agreement. Aim 1 Limitations, Alternative Strategies, and Benchmarks The key assumption of this work is that YACC use social media for social support. While YACC in our sample indicated that they valued social media as a source of social support, this may not apply to all YACC. Furthermore, it is possible that YACC who did not enter the study may use social media differently than those who did, and this may introduce selection bias and limit external validity. For example, while we did not experience this among any of the caregivers who were approached for the study, it is possible there are some YACC who did not want to have their social media posts collected or their interview recorded and thus did not express interest in the study. 39 Positive response bias may exist in that YACC were informed during the consent process that their existing social media posts would be collected, and thus they may have responded more favorably to questions about social media use. As anticipated, recruitment of YACC was a challenge, and despite a multiprong recruitment approach recruitment was finalized beyond the anticipated six-month time frame and with a smaller than targeted sample size. Aims 2-3: Purpose The purpose of Aim 2 was to identify periods within the first six months of the cancer caregiving trajectory (e.g., diagnosis, treatment), during which YACC have different types of functional social support, the extent to which that support was present on social media, and how the prevalence of each type of functional support changed over time. This information expands prior research by identifying the effects of social media platform and cancer-related content on social support, the purpose of Aim 3. Aim 2 and Aim 3 provide essential information on how social support changes over time, by platform type, and with the presence of cancer-related content. Aims 2-3: Data Collection Data collection for Aims 2-3 occurred after participants were interviewed in Aim 1. As mentioned previously, a limitation of the pilot study was that, due to anonymity of public social media profiles and lack of informed consent, characteristics of social media users were not examined. In comparison, a strength of this study is that personal characteristics of the social media users were collected from participants in Aim 1 and 40 merged with the social media data analyzed in Aims 2-3. This approach facilitated a detailed analysis of the effects of social media platform and cancer-related content of posts on social support. This is the first study to incorporate this novel approach to study social support on social media among YACC. Participants were asked to disclose via a screening form (Appendix B): social media platforms of use (e.g., Facebook, Instagram, and Twitter) and social media username from which to extract posts. Participants with private social media accounts were asked for temporary access so that posts could be extracted. We expected about N=14,600 posts (1 post per day x 2 platforms of use x 182.5 days/6 months x 40 participants).47 Given the complexity and large amount of data available, and limitations of web-crawling software, social media posts were manually extracted from Facebook and Instagram only, to construct an analytic database.89 Furthermore, there were social media posts that did not receive an interaction from members of the social network. This “non-response” was included because there are certain social media characteristics (e.g., longer posts) and YACC sociodemographics (e.g., younger age) that may be associated with nonresponse. To ensure high quality content, data cleaning screened for lexical features (e.g., typos, syntactic and semantic complexity).90, 91 Given the large amount of data collected, and the time and resource constraints of the study, a random sample of 10% of social media posts with equal representation from each month after the diagnosis was coded and analyzed. 41 Aim 2: Data Analysis Summary statistics were performed to describe social media posts (e.g., word counts). Social media posts that demonstrated evidence of social support were deductively coded using definitions of functional social support (Table 2.3) through a process called quantitizing,74 wherein text was content analyzed into numerical outcome variables which indicated the presence or absence of each type of social support in YACC social media posts and responses to their posts by followers. To establish interrater reliability of quantitized social support outcomes, social media posts were double coded from 10% of the sample (3 participants). Reliability was high (k=0.95) and thus the definitions of social support were subsequently applied to the sample of social media posts. Summary statistics were calculated for each type of functional social support before hypothesis testing occurred. Based on prior research indicating that emotional and information support are the top reasons for using social media,27, 30, 32, 40, 48 we hypothesized that emotional and information support would be more prevalent than other types of functional support (Hypothesis 1, Figure 3.2).32, 39 Previous research suggests that a cancer caregiver’s social support diminishes over time after the diagnosis;35 thus we also hypothesized that all types of functional support on social media would decline over time (Hypothesis 2, Figure 3.2). To test these hypotheses, detailed summaries of changes in prevalence of each functional support outcome was depicted graphically. Then generalized linear models using the eform command in Stata to transform effect estimates into odds ratios were estimated to depict the odds that social support exchanges would contain each type of functional social support, adjusted for platform type.92 Results show how each type of functional support is 42 sustained over time and are interpreted in the context of qualitative findings from the Aim 1 interviews. Aim 3: Data Analysis To identify potential effects of platform type (Facebook vs. Instagram) and presence of cancer-related content (no vs. yes). on functional social support over time, each social media post was assigned values for functional social support outcomes (present=1, not present=0) and coded for independent variables in Table 3.4. The social support outcome variables were transformed to depict the percentage of responses that contained each type of social support, with the unit of analysis being each original social media post by a YACC. Linear mixed effects regression models were used to compare the distributions of each type of social support offered in response to YACC’s social media posts based on platform type and presence of cancer-related content. Results of these models explain how social media characteristics influence functional social support on social media over the first six months of cancer caregiving. Aims 2-3: Limitations, Alternative Strategies, and Benchmarks In mixed methods, purposive sampling limits generalizability, and this is noted as a limitation for Aims 2-3.75 The information gained by combining YACC perceptions, YACC sociodemographics, and cancer patient factors in social media data analyses outweighs potential threats to external validity that may be present with a small sample size, given that this is one of the first studies to merge YACC personal characteristics with social media data. Interpretations are limited regarding how social support on social 43 media actually benefits YACC. The information gathered in the Aim 1 interviews provides details about the boundaries and limitations social support to situate the results of Aims 2 and 3 in the appropriate context. It is important to note that YACC likely spend time on social media during which they are not actively posting, but rather may be reading posts or articles of other members in their social network. These activities may contribute in meaningful ways to a YACC’s wellbeing (either positively or negatively), but we are limited in our ability to examine this activity on social media in the current study. Nontextual social media data (e.g., Facebook stickers, photos) were not analyzed. Mixed Methods Design and Integration Mixed methods integration occurred at multiple stages throughout the study and at different levels of the approach (Figure 3.4). At the data collection stage, this study employs a convergent mixed methods design, such that the qualitative and quantitative data were collected iteratively to allow for adaptation.93 Specifically, the parallel data collection allowed for reflexivity in how the social media data were collected and prepared for analysis.93, 94 At the data analysis stage, the qualitative and quantitative data were merged, bringing the two databases together for analysis.93 Specifically, Aim 1 interview data refine and explain the results of Aims 2-3 by providing an in-depth description of how YACC use social media for social support.95 Additionally, the results of the Aim 1 qualitative analysis were embedded in the social media analysis in Aims 2-3, meaning that the Aim 1 data revealed how social media platform influenced social support. Thus this was controlled for in the statistical analyses of Aim 2 to minimize bias.93 44 At the interpretation stage multiple mixed methods strategies are employed. While the results of Aim 1 are described separately in a qualitative manuscript, Aims 2-3 use mixed methods integration and interpretation.93, 94 Specifically, representative text of each type of functional social support evident in the content analysis of Aim 2 were integrated with the statistical analyses to develop a narrative of the social media posts. In Aims 2-3, data transformation occurred as social support outcome variables were created from the text of social media posts for quantitative data analysis.74, 93, 94 Throughout the study, the “fit” of the mixed methods data were assessed and confirmed or expanded as appropriate.93 The potential for discordance between the qualitative and quantitative results were minimized by the iterative and reflexive approach taken throughout the study.93 These mixed methods findings, in combination with Aim 1 qualitative results, provide vital information to inform future research and enhance clinical suport for YACC to leverage social media for sustained functional social support. Proposed Manuscripts There are three manuscripts proposed. Each manuscript focuses on the results of a specific aim. Manuscript 1/Chapter 4: Young adult cancer caregiver’s use of social media for social support (Aim 1) Manuscript 2/Chapter 5: Social support exchanges on social media during the first six months of young adult cancer caregiving (Aim 2) 45 Manuscript 3/Chapter 6: Platform type and cancer-related content influence social support on social media among young adult cancer caregivers (Aim 3) 46 Table 3.1 Inclusion and exclusion criteria Inclusion YACC • Ages 18-39 years15 Caregiver • Been in unpaid caregiving role for a cancer patient for at least six months • Currently use Facebook, Instagram and/or Twitter (i.e., within the last week)47 • Understand spoken English Patient • Age 18 or older • Within six months to five years from original cancer diagnosis Table 3.2 Targeted enrollment Emerging Adulthood (18-25 years) Young Adulthood (26-39 years) Total Exclusion • Unable to understand spoken English, primarily use another language besides English on social media • Unwilling to have interview recorded • Unwilling to provide social media access • Provides care in a professional manner • Receiving treatment for recurrent cancer Female 10 10 20 Male 10 10 20 Total 20 20 40 Table 3.3 Aim 1 Qualitative analysis, first and second cycle coding First Cycle Second Cycle Familiarization, develop coding Interpretation, identify Purpose scheme relationships between codes Approach Structural 87 Focused coding 87 Transcribed interviews will be Transcribed interviews will be Analysis unit analyzed line by line analyzed by line by line 1. Read transcripts, develop reflective memos, note first 1. Use first-cycle coding scheme impressions 84 to code all transcripts 2. Create labels for initial codes 2. Track reflexive and interpretive Procedure from the text memos 84 3. Sort labels into categories, 3. Bi-weekly meetings to discuss compare differences coding and interpretation 4. Sort categories into 10-15 meaningful codes Initial, grounded in participants’ Interpretation Abstraction, concept development voice 47 Table 3.4 Aim 3 YACC, cancer patient, and social media independent variables Variables Coding Rationale/Definition Social media use differs by YACC 1 sociodemographics25, 26, 33, 96 Relationship Sociodemographics type may effect social support needs.21 Age Continuous Current age in years Gender Categorical Male, Female, Other Sexual orientation Categorical Straight, gay/lesbian, bisexual, transgender, other Rurality Binary Rural, urban Ethnicity Binary Non-Hispanic White, Hispanic/Latino Race Categorical White, Black/African American, Asian, American Indian/Alaska Native, Native Hawaiian/Other Pacific Islander, Other Living situation Binary By self, with others Financial support Categorical Support self, supported by others, other Annual personal income Categorical ≤$24,999, $25,000-$34,999, $35,000$49,999, $50,000-$74,999, ≥$75,000 Hours worked per week Continuous Number of hours worked at a job per week Health insurance status Binary Insured, uninsured Health insurance type Categorical Employer sponsored, individual, Medicare, Medicaid, Military, Don’t know, Other Language Binary English, other Relationship status Categorical Partnered/married, divorced/separated, widow/widower, single Supportive needs may differ by cancer Patient Factors1 diagnosis or other patient demographics (e.g., age, disease stage). Patient age Continuous Patient current age Caregiving age at Continuous Caregivers age when patient was diagnosed diagnosis Diagnosis Categorical Type of patient’s cancer diagnosis Disease progression Categorical Cancer stage, prognosis Relationship Categorical Type of relationship (e.g., mother-daughter, friends) Length of relationship Continuous Length of patient – caregiver relationship in years Social Relationship Scale Six validated questions measuring Index relationships in three contexts: stress, positivity, negativity. Duke Social Support and Scale Validated tool measuring family and nonStress Scale family social ties and stress82 Longer posts, post frequency, and social Social media 2 media platform may influence social support. characteristics Post length Continuous Character count 48 Table 3.4 Continued Variables Type of post Type of user Platforms of use Frequency of platform use Daily time spent on social media Social media social support2 Likes Comments Shares/retweets Coding Categorical Categorical Binary Rationale/definition Photo, video, other Individual, organization, other Yes/No for Facebook, Twitter, Instagram, Pinterest, Vine, Snapchat, LinkedIn, YouTube, Google+, Other More than once per day – Do not use Likert scale Continuous Measure in minutes per day Social media indicators of social support.49, 96 Continuous Total number of likes per post Continuous Total number of comments per post Continuous Total number of shares/retweets per post Aim 1 survey, includes developmental stage; 2Aim 2 social media data collection; 1 49 Figure 3.1: Mixed methods design Figure 3.2 Aim 2 hypothesized change in functional social support over the first six months of caregiving 50 Figure 3.3: Mixed methods (MM) data integration Figure 3.4: Aims 2 and 3 hypotheses, analysis, and future research domains CHAPTER 4 YOUNG ADULT CANCER CAREGIVERS’ USE OF SOCIAL MEDIA FOR SOCIAL SUPPORT Abstract The purpose of these cross-sectional semistructured telephone interviews was to describe 1) how YACC use social media and 2) what kind of functional social support YACC receive on social media. Eligible YACC were 18-39 years of age, used Facebook, Instagram, and/or Twitter at least once per week, and cared for an adult cancer patient diagnosed six months – five years prior (N=34). Recruitment occurred through oncology clinics in Utah and online advertising by caregiving and cancer organizations from September 2017-June 2018. Interviews were recorded, transcribed, coded and qualitatively analyzed. We found that caregivers were most commonly spouses and on average aged 29 years (range 21-38). Cancer patients were on average 37 years (range 19-76). Four distinct thematic categories emerged about how YACC used social media during a cancer experience: 1) social media is a preferred method for providing updates about the cancer patient; 2) caregivers who post about cancer on social media get support they otherwise would not have access to; 3) posting about cancer online introduces risk for negative consequences (e.g., misinformation, bad advice, loss of privacy); 4) potential 52 for negative consequences influences how some caregivers use social media. All types of functional social support were present on YACC’s social media: information (e.g., feedback from others), emotional (e.g., encouragement), companionship (e.g., feeling less alone), instrumental (e.g., fundraisers), and validation (e.g., reassurance). Posting about cancer on social media often begins as a strategy for YACC to efficiently provide updates about the cancer patient. Caregivers who actively post on social media experience a variety of different functional social supports to which they otherwise would not have access. Yet, this support is sometimes accompanied by potential negative consequences (e.g., loss of privacy, poor advice) which leads some caregivers to avoid social media. YACC who are active users of social media may benefit from using social media during a cancer experience more than passive users. Introduction Addressing the challenges YACC face is of increasing importance, because caregiving roles are shifting to younger generations as the U.S. population ages.2, 16 In the U.S., about half of all caregivers for adult patients are young adults (48%), and nearly a quarter are between 18-34 years (24%).16 Innovative strategies are needed to support young caregivers in their own communities.14 These communities are diverse, ranging from physical clinical space for teens and young adults in hospitals to virtual social media support groups or mobile applications.97 Cancer caregivers report higher burden than caregivers of patients with other health conditions.98 Cancer caregiving is unique because cancer often has a sudden onset, rapid advancement, and life-threatening nature. Unexpected caregiving can disrupt 53 wellbeing and prompt fear of abandonment and uncertainty.12, 18-20 Negative caregiving effects are acute for young adults who are often unfamiliar with severe illnesses, have multiple caregiving responsibilities, and are finishing their education and establishing new careers.22, 37, 38 Young adults are more likely than older caregivers to experience stress, depression, and unmet needs.20 While caring for a loved one with cancer brings challenges, it can be rewarding, enhance relationships, and bring meaning to one’s life.21, 23 Caregivers who feel supported exhibit better coping behaviors and healthier outcomes.3, 5-7 Unfortunately, most resources are targeted toward older caregivers; often support is not provided in the formats and setting desirable to young adults.17, 21 Between 88%-97% of young adults in the U.S. use social media99, 100 to improve social support, gain trust, and maintain relationships.26, 27 When it comes to health, social media is used to find information and seek advice.48, 49 Positive and negative health and social outcomes have been linked to social media use.101, 102 Although limited, recent data suggest that cancer caregivers actively use social media for information and support;103 however, no studies have explained the process of how YACC experience social support on social media during a cancer experience or what types of social support result from their social media use. Our findings provide a novel description of this process. Understanding how social networks are engaged following a cancer diagnosis is the first step for developing future technology-aided social support interventions, which are acutely needed to support the growing millennial and Z caregiver generations. 54 Methods Theoretical foundation The Cancer Family Caregiving Experience Framework posits that cancer caregivers experience an appraisal process where they assess their primary and secondary stressors, then they decide how to cope with these stressors using the resources they have available.35 Given young adults’ widespread use of social media, YACC may use social media as a resource for coping with their stressors. However, this model does not explain or acknowledge how YACC use social media during a cancer caregiving experience or how social media use may serve as a source of social support during the cancer caregiving experience, including what kinds of social support result from social media use. The definitions of functional social support used for interpreting support on YACC’s social media are derived from the Stress and Coping Social Support Theory, which explains that as social needs are met, caregivers’ abilities to fulfill their roles are improved.67, 68 This study was approved by the University of Utah Institutional Review Board. Participants and recruitment Eligible caregivers were 18-39 years, spoke and understood English, used social media at least once per week (i.e., Facebook, Twitter and/or Instagram), and had been caring for a loved one with cancer diagnosed six months – five years prior. Recruitment occurred through flyers, social media advertisements, and in-person referrals. Caregivers were recruited directly through flyers and social media advertisements, and referrals were initiated by discussing the study with cancer patients, who then provided contact 55 information for their primary caregiver to the study team. Flyers were posted in waiting areas, restaurants, a cancer resource library, and wellness center at Huntsman Cancer Institute. Social media advertisements were shared by local, national, and international young adult cancer and caregiving organizations as well as by collaborators at other institutions in Iowa and Massachusetts. In-person referrals occurred through out-patient and in-patient settings via the Huntsman Intermountain Adolescent Young Adult Oncology Program, which serves Utah and the Intermountain West. Of 354 cancer patients screened (through electronic records and in-person screenings), 61 potential caregivers were identified; 13 were deemed ineligible (patient’s time since diagnosis (n=3), age (n=4), recurrent cancer diagnosis (n=1), caregiver did not use social media (n=2), other (n=3)). This left 48 eligible caregivers; n=8 declined, n=6 were unreachable, and N=34 participated (participation rate=70.8%). Enrolled caregivers completed informed consent and a telephone interview. Data collection and management We asked participants both closed ended questions and semistructured interview questions. The semistructured questions were about caregiving experiences, social media use, and future interventions. There were 34 close-ended questions about sociodemographics, the cancer patient, social support, and social media use, which were imported into REDCap for storage and cleaning. Interviews ranged from 41-79 minutes in duration, were audio recorded and transcribed, then read and checked against audio recordings for discrepancies. 56 Sociodemographic, cancer, and social media variables and descriptive statistics YACC sociodemographics included age, gender, ethnicity, race, employment status, insurance status, marital status, educational status, and caring for others besides patient. Cancer patient variables included patient age and relationship to caregiver. Social media use was classified as daily, weekly, monthly, or no use for Facebook, Instagram, and Twitter. Descriptive statistics were performed in Stata 14.2. Interview analysis Transcripts were analyzed in iterative cycles moving from more general descriptive codes that labeled content, to more focused coding that aligned codes by concept and content. In first cycle open coding,87 where descriptive content labels were used, 10% of the interviews were coded individually by ELW and ARW resulting in 321 codes. First cycle generated content and process codes, and second cycle categorized these codes by identifying conceptual and content similarity. In the second cycle focused coding,87 the coding scheme was applied to an additional 10% of the interviews, which were double coded by ELW and ARW. In the second cycle coding, coder agreement, as measured by interrater reliability, was “strong” (κ = 0.88).104 Where codes were found to have zero or negative agreement, the coding scheme was revised through discussion and settled by rater consensus. We created process and analytic memos in NVivo 11 throughout data collection and analysis to note procedural irregularities, meaning of ideas, and interpretations.84, 85 The final coding scheme was applied to all 34 interviews. Final codes related to 57 the process of using social media for social support were identified using a “trinity” approach, a grounded theory technique that organizes central themes with the intention of informing future theory development.105 Our analysis yielded important information about 1) how YACC use social media during a cancer experience, and 2) what kind of social support results from social media use. Results Sociodemographic, cancer, and social media characteristics The average caregiver age was 29.0 years (Standard Deviation (SD): 4.72, range 21-38). Caregivers were primarily female (70.6%), Non-Hispanic White (91.2%), employed (85.3%), insured (97.1%), married/partnered (73.5%), college graduates or higher (53.0%), and caring for others besides the cancer patient (67.6%, Table 4.1). Cancer patients were on average 37.0 years old (SD: 13.77, range 19-76) and most commonly a spouse/partner (52.9%) or a parent (17.6% mothers, 5.9% fathers, Table 4.2). Caregivers used Facebook (79.4%), Instagram (61.8%), and Twitter (2.9%) daily (Figure 4.1). Caregivers had been taking care of the cancer patient for a median of 1.6 years at the time of the interview (standard deviation = 1.1 years). Final thematic categories of how YACC used social media included: 1) Social media is easier than other modes to provide updates about the cancer patient, 2) Caregivers who post social media updates gain support they otherwise would not have access to, 3) Posting about cancer online presents opportunity for negative consequences, and 4) Potential for negative consequences influences how caregivers use social media. Subthemes and exemplary quotes are shown in Table 4.3. 58 Thematic category 1: Social media is easier than other modes to provide updates about the cancer patient Caregivers felt overwhelmed providing updates about the patient to family and friends and this led them to use social media as a tool for updating their loved ones about the cancer patient. Those who used social media for updates felt it was easier logistically and emotionally than other methods such as phone calls and text messaging. Logistically – to save time The time and effort required to provide updates about the cancer patient frustrated caregivers. They felt limited by phone calls and text messaging because these approaches capped how many people they could communicate with simultaneously. Furthermore, the time spent providing updates through traditional communication methods, like email, phone calls or text messages, interfered with their ability to spend time with the patient including completing caregiving tasks (e.g., supporting patient with activities of daily living, doctors’ appointments, and household tasks). This led caregivers to use social media for updates because it enabled them to reach many people with minimal effort and protected their time so they could focus on supporting the patient. Caregivers also believed they could protect the patient from obtrusive requests for updates by posting on social media and that their updates would benefit both themselves and the cancer patients. 59 Logistically – how it happens Caregivers discussed social media posts with the patient before making updates and it was important that these discussions occurred before any posts were made about the cancer patient on social media. Caregivers described consulting patients for permission to post photos, health information, treatment decisions, and about the level of detail to share on social media. Similarly, caregivers felt it was a sign of respect for them to wait to post about major milestones until the patient had made these announcements. Conversations between caregivers and their family members with cancer dissuaded a few caregivers from posting online about the cancer diagnosis. Whether the caregiver posted updates depended on the relationship and the patient’s comfort with social media. For example, a sibling caregiver felt she would use social media differently for loved ones with different social media views. Conversations with patients about posts did not always occur, particularly if the diagnosis was previously disclosed on social media or caregivers felt these discussions were unnecessary because posting about the cancer patient’s experience on social media would be less emotionally burdensome for themselves and the cancer patients. Emotionally Emotional fatigue drove caregivers to post updates about the patient because doing so reduced emotional burden. Being able to disclose updates about the cancer patient online was less difficult than sharing these updates face-to-face because caregivers only had to post information once online as compared to repeatedly talking about what was going on with the cancer patient. An unintended consequence of updating 60 on social media was the caregiver feeling like they needed to provide emotional support to others that interacted with their social media post. A caregiver described how this happened saying, “It’s super draining because you find yourself in the position to comfort them when they should be comforting you.” Social media updates were perceived as low stress because they provided flexibility in response time which then allowed caregivers to better accommodate the patient’s needs and their other responsibilities because they were not spending as much time sharing updates with others. However, some caregivers still felt emotionally burdened with social media updates. One caregiver lamented, “Sometimes it would be overwhelming…I just felt like it was another thing that I had to do to keep up on.” Thematic category 2: Caregivers who post social media updates gain support they otherwise would not have access to Posting about the cancer experience online facilitated connections with others and helped caregivers feel like they were “giving back” to their online communities. Caregivers posted a variety of different updates, and not all were focused on the cancer patient. Some updates shared information about the caregivers, the family unit, or other noncancer related happenings. Connections with others Caregivers found support by connecting with others in similar situations, either through their personal networks or cancer-related groups. Some caregivers chose to expand the breadth of these connections by making public updates. Publicizing their 61 experience enabled these caregivers to surpass the reach of their offline support systems. Broadening social networks, whether publicly or privately, facilitated disease and agespecific connections that caregivers believed they otherwise would not have experienced. These connections were often preferred to in-person support groups because they permitted tailored support. A caregiver explained the benefit of these connections, “Being a caretaker to a cancer patient could be very isolating at times, and not a lot of people understand what it means or what you're going through. And so, to speak to somebody who is going through a familiar experience was comforting.” “Giving back” Caregivers felt rewarded when they shared advice, knowledge or skills, and their cancer journey with newly diagnosed patients and caregivers. They believed that sharing caregiving experiences brought comfort to others during difficult cancer or noncancer experiences because they were demonstrating strength and resilience. This notion was validated in the feedback they received on social media when they received comments about how they had impacted other people’s lives and their cancer journeys. Caregivers commonly posted about the patient’s diagnosis and treatment to raise awareness of cancer in their social circles. They participated in cancer awareness campaigns, discussed cancer symptoms, and promoted fundraisers. Posting photos or videos of awareness events was perceived by caregivers as a way to advocate for cancer prevention. 62 Thematic category 3: Posting about cancer online presents an opportunity for negative consequences Caregivers acknowledged the potential negative consequences of posting about cancer on social media, including misinformation, uncomfortable responses and posts, and limited social cues. Misinformation Misinformation was perceived as an expected consequence of using social media, and caregivers largely believed good intentions precluded misinformed feedback. One caregiver stated, “I understand the people [who] are doing that are doing that out of love and that they really are misinformed and they don’t know that it doesn’t help.” Caregivers were accustomed to misinformation because they navigated it on social media outside of a cancer context, and they had learned ways of coping with misinformation by ignoring it or researching information that they did not perceive as being trustworthy. Uncomfortable responses and posts Uncomfortable responses and posts occurred when individuals in the social network made uneven comparisons, provided unsolicited advice, minimized the caregiver’s experience, made insincere comments, or provided incongruent support. Caregivers were frustrated if they perceived others did not understand the severity of their situation. A caregiver explained, “…it’s like the comparison game…like, ‘Oh, I’m so sorry. I know exactly what you’re going through’ but it would be like their dog passed away…It’s a different level of severity.” Caregivers perceived responses that minimized 63 their experiences to be unsupportive and felt disappointed with insincere responses. While responses to caregiver’s posts were the most common source of discomfort, some caregivers were upset when they passively viewed cancer-related posts from other caregivers or patients, and this was particularly the case when it came to posts about bereavement or end-of-life which caregivers interpreted as being scary. Limited social cues Social media diminished social cues that occur during face-to-face interactions (e.g., body language), which sometimes made it difficult for caregivers to manage uncomfortable situations on social media. For example, one caregiver was caught off guard by comments on her post, “It’s hard for me to see something on social media like… ‘Oh I hope your sister’s doing good, unfortunately, my sister or my mom didn’t make it’. For me, it’s really uncomfortable. You’re not in front of that person, but I mean, what do you comment back? Like, I’m sorry? That doesn’t really sound very sincere.” Others felt the limited interaction on social media led to misunderstandings and made it difficult to personalize posts about their cancer caregiving experience or the wellbeing of their family member with cancer. Some caregivers used strategies to ameliorate the lack of social cues, like making a photo or video post. Thematic category 4: Potential for negative consequences influences how caregivers use social media While the majority of caregivers accepted negative consequences as an expected outcome of using social media, this led some caregivers to forego posting about cancer 64 on social media. The most common reason for not posting about cancer online was for privacy, either because the patient did not want their diagnosis to be shared publicly or because the caregiver felt cancer was a private matter. Safety, fear of judgment, and criticism also dissuaded caregivers from posting online. Caregivers who did not post on social media still used social media for cancer-related purposes, like viewing the posts of other cancer patients or caregivers, interacting anonymously on social media groups, and providing covert updates about the cancer patient. For example, a caregiver posted pictures to update people who know, offline, about the cancer diagnosis. She said, “We’ll post more stuff with my mom, like pictures, because it’s kind of a way for family to know she’s doing good.” Passive users of social media, those who did not post about the patient having cancer on social media, described gaining cancer-related information, connecting with others who are going through a cancer experience, and viewing how others are dealing with cancer. Presence of different types of social support Caregivers had all forms of functional social support on social media. Definitions and illustrative quotes for different types of social support that resulted from YACC posting updates about the cancer patient are shown in Table 4.2. Caregivers received feedback and information, encouragement and positive responses, donation of money and meals, and felt loved and cared for through their social media use. Although caregivers initially used social media to provide updates on the cancer patient, social media became an important social support resource for themselves as well. A caregiver described, “We set up a Facebook page for [the patient] for support…So when I [posted on this 65 Facebook page] this time, it was just a bad day. I guess it was kinda me shouting out like, “I’m frustrated, and I’m mad” and so, when people acknowledged it or just showed the thumbs up…it was kinda nice...” While few caregivers described making posts that focused on themselves, they felt supported when members of their social network interacted with their social media posts about the cancer patient. For example, a caregiver described, “You’re feeling like you’re just sharing information, but also, you need something back from social media. That’s why we [post about cancer on social media], to get a little bit of a response.” Social support on social media helped caregivers feel less alone. A caregiver described this as, “I think [a caregiver Facebook group] helps me not feel completely alone because cancer treatment for my wife has been very isolating [for me]. Discussion For most caregivers, the process of using social media for social support as a YACC is preceded by requests for updates about the cancer patient. As requests for updates become more time consuming and emotionally burdensome caregivers turn to social media as a practical tool. Separate pages are created, or a caregiver’s personal pages are used for providing updates, and support is received as a result of these updates. While the majority of this support is perceived as positive, negative consequences are possible. Most caregivers are accustomed to negative consequences on social media in noncancer contexts. However, for a minority of caregivers, patient preferences, privacy concerns, misinformation, and uncomfortable responses may lead them to limit or avoid posting cancer updates on social media. This novel description fulfills an acute need for 66 updated information about how YACC use social media for social support, a critical step in the development of future supportive interventions for YACC. Our results can inform future cancer caregiver theory development by explaining how the process of social media use and subsequent social support is part of the young adult caregiver experience.35 In the current Cancer Family Caregiving Experience Model, a stress cycle occurs in which caregivers evaluate their primary (e.g., patient illness, functional dependence) and secondary (e.g., changes in relationships, feelings of isolation) stressors. 35 Evaluation of these stressors initiates the appraisal process, wherein the caregiver determines what these stressors mean for them and how they will cope.35 Until now, no studies have articulated the process by which YACC use social media as a tool for coping with primary and secondary stressors and how posting about the cancer experience on social media leads to social support. Our results suggest that as the stress of providing updates about the cancer patient becomes increasingly burdensome, caregivers use social media to manage this social stress. Interestingly, our findings also suggest that social media is an important source of social support for both active (i.e., those who openly post about cancer) and passive (i.e., those who do not post about cancer) users. Online communities, like those discovered on social media, are a critical component of young adult caregivers’ social context that has not previously been included in the caregiving model. These findings provide new information about the role of social media in the appraisal process and subsequent social support. Our participants initiated social media use as a practical tool but continued using it to complement their offline social networks. Prior research shows that cancer caregivers use social media for getting information and emotional support,40 two of the 67 five types of functional social support defined in the Stress and Coping Social Support Theory.67, 68 Caregivers in our study also gained information and emotional support, but unlike existing literature, they received validation, companionship, and instrumental support. Given young caregivers’ competing demands and high risk for isolation during a cancer experience,21 social media may be strategically used for less common types of social support. For example, caregivers needing financial aid, a type of instrumental support,67, 68 may have a broader reach with a fundraiser shared on social media than a one-time fundraising event. The nature of cancer caregiving might present a unique situation for social support online, which may explain why validation, companionship and instrumental support is typically not described in studies of other diseases. Future research is needed to assess the extent to which presence of all functional support types on social media is unique to YACC. Negative consequences of using social media described by our participants merit attention. As described in the Stress and Coping Social Support Theory, support that originates from inappropriate sources or in undesired forms can be detrimental to coping.67, 68 There is evidence of the negative impact of unsolicited advice and misinformation on social media,48, 52-54 and we found examples of this in our interviews. These situations were interpreted as negative consequences of using social media. Because social media use has been associated with negative outcomes like depression,101 which is also common among cancer caregivers,20, 23 caregivers using social media for updates may benefit from clinical guidance on best practices for using social media during a cancer experience to avoid exacerbating negative mood. There was a tendency among participants who used social media passively before 68 the cancer diagnosis to hold more negative expectations for using social media as a cancer caregiver. In contrast, active social media users were adept at using social media during a cancer experience. Passive use has been associated with negative outcomes (e.g., depression, negative rumination) in prior social media research.102 This suggests that certain patterns of social media use may be more and less effective for eliciting social support on social media during a cancer experience. More work is needed to explore and define these patterns before guidelines can be developed and recommended to caregivers by supportive care professionals. This study offers novel information but has limitations. Our results may overrepresent the experiences of spousal caregivers and underrepresent the experience of caregivers with other relations to patients. While qualitative research provides a rich description of a phenomenon and can inform theoretical development, our sample is small and other caregivers may experience the process of using social media for social support differently. Because social media use differs by race,100 future research that describes the experiences of racial and ethnic minority cancer caregivers is needed. Lastly, individuals may overestimate the positive outcomes of using social media. Therefore, without objectively measuring social support from caregivers’ social media feeds we cannot determine the extent to which caregivers’ perceived social support reflects reality and we do not know the extent to which they distinguish between different types of support or what support types are considered most valuable to caregivers. Future research should use robust measures of social support to explain changes in support online and identify factors that influence how social support is sustained over time. 69 Conclusion YACC make up an increasing proportion of all caregivers in the U.S. As these individuals increasingly take on caregiving roles there is an imperative need to improve and expand the reach of supportive care resources. To fully support YACC, services are needed in formats that are convenient for them as they balance their caretaking duties with their daily lives. Prior research shows that young people may prefer accessing resources in online formats because this can be easier to do in terms of time and comfort level.17 However, there is still much to be learned about the strengths and challenges of supporting YACC through online technologies like blogs, information resources, provider portals, and social media. By describing how YACC use social media for social support, this study provides critical information for adapting existing cancer caregiving theory and informing the development of these future supportive care resources. 70 Table 4.1. Sociodemographic characteristics of YACC (N=34) N % Age at Interview 18-24 4 11.8 25-29 13 38.2 30-34 9 26.5 35-39 8 23.5 Gender Female 24 70.6 Male 10 29.4 Ethnicity Non-Hispanic White 31 91.2 Hispanic/Latino 3 8.8 1 Race White 31 91.2 Black/African American 1 2.94 Asian 0 0.0 AI/AN 2 5.9 NH/PI 0 0.0 Other 3 8.8 Employment status Employed 29 85.3 Unemployed 5 14.7 Insurance status Insured 33 97.1 Uninsured 1 2.9 Marital status Married/partner 25 73.5 Single/divorced/widowed 9 26.5 Educational status High school or less 3 8.8 Some college 13 38.2 College graduate or more 18 53.0 Caring for others besides patient2 Yes 23 67.6 No 11 32.4 1 Totals do not equal 100% because participants could select more than one option 2 Options included a child under age 18, child over age 18, parent, spouse, or other 71 Table 4.2. Characteristics of cancer patients (N=34) Patient Age at Interview 18-29 30-39 40-49 +50 Relationship with caregiver Spouse/partner Mother Father Sibling Other (e.g., cousin, child) Time since diagnosis 6 months to less than 1 year 1 years to less than 2 years 2 years to less than 5 years N % 7 19 1 7 20.6 55.9 2.9 20.6 18 6 2 5 3 52.9 17.6 5.9 14.7 8.8 12 16 6 35.3 47.1 17.7 A. Logistically – to save time 1) Social media is easier than other modes to provide updates about the cancer patient C. Emotionally B. Logistically – how it happens Subthemes Thematic category • • • • • • • • • • “All those emotions come up every time I have to go over it again…I couldn’t be on that emotional level, that high emotion all the time, every day, with every single person…so [social media] was a way for them to know without me having to tell them.” “I could go and look at [the responses] at a later date, at all of them compiled together, instead of looking at every single one, every time they came across.” “I definitely talked to him about doing it before I did it because I didn’t know if he would be okay with it or not.” “He definitely wanted me to not get too personal with it.” “When it was bigger news, I would have to wait until she posted it so I wasn’t stealing her thunder.” “She specifically asked that we do not, um, publicize her cancer on Facebook.” “…If it were my spouse, that would be different, it would depend on what he would want me to do…but, if it were my mom, I’m sure she’d love for me to post about [cancer].” “…your phone only allows you to mass text so many people on one message, so it was a lot of work...with Facebook…I just said, I’m creating a separate group. If you want to be updated, I’ll add you...” “In a tense situation that interferes with your ability to actually…care for [the patient] because you’re taking the time to speak with [loved ones] about it.” “I wanted to update everybody [on social media], so that he…could rest and not have people bugging him.” Exemplary quotes Table 4.3. Thematic categories, subthemes, and exemplary quotes for functional social support types on social media among YACC 72 A. Misinformation • “…there's definitely bad advice” • “I think [our page] has created a support group that has impacted other people’s lives and given them some hope.” • “I get so many thank-yous from people and so many; you have no idea how much you’ve changed our cancer journey.” • “September was ovarian cancer awareness month, and so I…found a little thing online that, these are the symptoms of it and everything… I posted [that] for the first time in September.” B. Giving back 3) Posting about cancer online presents opportunity for negative consequences • “It was comforting to know that someone’s already gone through this, and this is how you get through it yourself.” • “This page we’ve opened it up to the whole public. It wasn’t just our friends and family getting on it, but people that we had never met in our lives before.” • “…People from all over the world were praying for us and sending us good thoughts, and that was pretty cool to see.” • “…You [found] people that were in the exact same spot you were in. Where if you went to [a support group] in person, you’re only gonna have a handful of people, and there’s no telling what they’re facing or what part they’re in.” • “I never expected people to help or to donate their time or money or other things like that, that were a huge help, but I never would have asked for and never would have expected.” A. Connections with others 2) Caregivers who post social media updates gain support they otherwise would not have access to Exemplary quotes Subthemes Thematic category Table 4.3. Continued 73 4) Potential for negative consequences influences how caregivers use social media. Thematic category Table 4.3. Continued B. Doesn’t influence social media use A. Influences social media use • • • • • “[Bad advice] hasn't really been a problem…just something you take with a grain of salt.” “It was just right after her diagnosis, and she just said like “No Facebook”. Like, “I don’t want anyone putting anything on Facebook about this.” “there’s two melanoma groups that I like have on my feed that I look at articles.” “She didn’t want anybody at work to find out because she didn’t want people to be staring at her chest, asking her, making comments.” “I have always in all honesty preferred to talk to someone in person, you can engage them, you can see their emotions, you can see and feel what they’re going through…So on this site that we created for [the patient] there’s a lot of videos…doing it that way it’s almost, well it’d be like you [were in] person, because now you’re seeing all the emotions.” • “…Not that it’s bad that people are like, ‘You’re my hero! You’re my inspiration!’ But it doesn’t feel like it has much merit when those comments are on every kind of post.” • “…when you tell some news, and to you it feels absolutely devastating, and somebody downplays it…” • “…sometimes it was kinda scary because some [other families]…went through similar things that we have, and then things got worse versus better.” B. Uncomfortable responses & posts C. Limited social cues Exemplary quotes Subthemes 74 75 Knowledge, advice, alternative action Availability of people to spend time with Transportation, household chores, child care, financial support Feedback, social comparison Sympathy, caring, acceptance Companion ship Instrumental Validation Emotional Information Table 4.4. Exemplary quotes for functional social support types on social media among YACC Support Definitions30,31 Exemplary quotes • “I was hoping to get feedback or information from others [on Facebook], which that’s exactly what it did.” • “I would get a lot more information [thorough forums] than I did going to the few doctors’ appointments.” • “Yeah…just because people post research papers [on Facebook groups].” • “I think social media is important when it comes to support and encouragement and just never giving up. So when I see [social media] posts about individuals going through treatment, there are thousands of encouraging comments and prayer comments and stuff like that. And I think that’s important when it comes to an illness like this.” • “I think it’s great, especially in these times to get, the more positive responses you can get, the more people that are on your team, it just makes it feel easier.” • “Without having the Facebook page it wouldn’t be as outreaching, we wouldn’t have got the amount of support that we’re getting now.” • “For sure I feel like it’s probably more responsive on Facebook than on my text messaging.” • “I never expected people to help or to donate their time or money or other things like that, that were a huge help, but I never would have asked for and never would have expected.” • “But I think having it on social media kind of broke that barrier of people that you’re not as close with or you don’t see on a daily basis, for them to feel more comfortable now to come to you and say here’s this. Or what can I do? Or you know, [social media] helped them connect with me and me with them, without it being super weird I guess.” • “And through social media he was able to pay a lot of the up-front costs that come from cancer, just through that. People are really supportive in that sense.” • “There was tons of love, tons of support of, ‘Hey, remember, you guys have got this. You’re strong.’” • “People even from the melanoma pages that I would comment or post in the private groups would follow our page to see what was happening…I guess in a way [their responses] made – it definitely made me feel special.” 76 Frequency of commonly used social media platforms 100 90 Proportion using 80 70 60 50 40 30 20 10 0 Facebook Daily Instagram Social media platform Weekly Monthly Twitter Do not use Figure 4.1: Frequency of commonly used social media platforms 77 CHAPTER 5 SOCIAL SUPPORT EXCHANGES ON SOCIAL MEDIA DURING THE FIRST SIX MONTHS OF YOUNG ADULT CANCER CAREGIVING Abstract YACC are a growing population in the U.S. Their support needs differ from those of older adult caregivers because they experience cancer caregiving at a unique developmental stage of life when they often have multiple caregiving roles. We explored the use of social media for YACC exchanges of social support with their followers in a cross-sectional retrospective mixed methods study. Eligible YACC were 18-39 years, used Facebook and Instagram at least once per week, and cared for an adult cancer patient diagnosed six months-five years prior (N=34). Recruitment occurred through oncology clinics in Utah and online advertising by caregiving and cancer organizations from September 2017-June 2018. Each participant completed a semistructured interview which was recorded, transcribed, coded and qualitatively analyzed. Then, 10% of the social media posts exchanged, meaning they were posted by either the YACC participant or their followers, during each of the six months after diagnosis were randomly sampled and content analyzed for five types of social support (emotional, information, validation, companionship, and instrumental). Descriptive statistics were calculated for 78 sociodemographic factors, cancer factors, social media platform, and support types. Generalized linear regressions were fit to portray changes in the prevalence of posts containing each social support type over the six-month period following the patient’s cancer diagnosis. A convergent mixed methods design facilitated integration of the qualitative feedback about changes in social support during the first six months of caregiving to explain the prevalence of social support types in the quantitative analysis. Caregivers were primarily ages 25-29 (38.2%), female (70.6%), Non-Hispanic White (91.2%), employed (85.3%), and caring for others besides the cancer patient (67.6%). Cancer patients were predominantly ages 30-39 (55.9%), a spouse/partner (52.9%) or a parent (17.6% mothers, 5.9% fathers). We analyzed N=2,090 social media posts, of which 26.9% were made by YACC participants and 73.1% were made by followers. Over one-third of posts were cancer-related (36.8%). In aggregate, the most common type of social support exchanged was emotional (63.3%), followed by information (27.7%), validation (15.3%), companionship (5.7%), and instrumental (1.3%). When controlling for platform, over the six-month period following the patient’s diagnosis the odds of posts containing emotional support decreased significantly (adjusted Odds Ratio [aOR]: 0.90, 95%CI 0.85-0.94, p<0.001), while the odds of information (aOR: 1.15, 95%CI 1.09-1.21, p<0.001) and companionship (aOR: 1.12, 95%CI 1.02-1.24, p=0.02) support increased. Qualitative feedback emphasized the high prevalence of support in the first month after the cancer diagnosis, and described factors that influenced future social support exchanges like active vs. passive use of social media, social network composition, and suboptimal support. Social media is an important source of social connection for YACC, particularly for exchanging emotional and information support, 79 the two most prevalent types of support exchanged by our participants and their followers. Supportive care personnel may consider evaluating YACC’s unmet supportive needs to determine how social media use may facilitate or hinder their social support. Future work is needed to explore the cost and accessibility of providing patient-centered counseling to YACC on social media-based strategies for improving their social support. Background Young adults, ages 18-39, are increasingly adopting cancer caregiver roles.106 As of 2015, there are an estimated 1.46 million YACC in the U.S.1 YACC enter a caregiving role during a transitional stage of life.37, 38 As new caregivers, they must integrate caregiving duties with existing responsibilities of young adulthood, like completing their education, starting new careers, and potentially caring for young children,21 areas in which older caregivers may be more established or are no longer relied upon. Caregiving may also be new territory for YACC who are often unfamiliar with severe illness.18, 107 While caring for a loved one with cancer can have positive aspects,21, 23 YACC are at particular risk for negative effects of caregiving.21 For example, YACC often have multiple and cross-generational caregiving responsibilities (e.g., children and parents), and they may experience role reversal when they care for a parent or grandparent who once took care of them.21 These challenges have potential negative consequences,18 including that YACC may put aside personal aspirations and delay key developmental milestones, like attaining financial independence, when they take on the cancer caregiving role. Caregiving as a young adult also limits social engagement, which may lead to feelings of isolation and depression.19, 21, 22 Compared to older cancer caregivers, 80 YACC report greater stress and depression during the first six months of caregiving.100 Mitigating the social burden of caregiving among YACC is a national priority.14, 35, 106 Having access to robust social support may alleviate negative impacts of caregiving on YACC’s distinctive social needs. Previous research suggests that social support mediates burden among caregivers of patients with functional impairment,5 and is negatively related to depression, loneliness, and burden among cancer caregivers.5-7, 108 Likewise, social support has protective effects on health,26, 109, 110 and may shield YACC from becoming overburdened when they find themselves in an unexpected caregiving role.12 Given the uniqueness of their caregiving role and their likelihood of negative psychosocial effects compared to older cancer caregivers,18 YACC require innovative solutions for engaging their social networks for support. YACC are part of the millennial and Z generations,45 often referred to as digital natives.111, 112 In the U.S., an estimated 88%-97% of these young adults use social media,99, 100 and they do so to gain social support, garner feelings of trust, and establish and maintain close relationships.26, 27 When it comes to their health, young adults use social media to provide updates about personal health information, request guidance or feedback about disease, and receive information and emotional support.48, 49, 99 In a digital age where online social networks are a natural extension of YACC’s existing communities, social media is likely a natural resource for social support. However, the limited literature available on YACC’s social media use has not assessed the types and prevalence of social support that are exchanged on social media between YACC and their followers. Detailing these social support exchanges is the next immediate step for helping supportive care personnel aid YACC in the digital age. 81 A growing body of literature supports studying the types of social support on social media among YACC. In 2015, 84% of cancer caregivers expressed interest in using social media as a resource for coping with cancer.40 Of these, 54% endorsed a need for emotional support and 77% endorsed information support.40 Cancer caregivers also endorse the use of social media to help coordinate care and avoid conflicts with the cancer patient.40 For example, social media content on Twitter about young adult cancer in general encompasses a range of emotional expressions including hope, sadness, fear, humor, and anger,27 demonstrating that social media is used as an outlet during cancer experiences. Prior research suggests that users who seek emotional support tend to remain involved in online social communities longer than those who seek information support.32 While this research clearly demonstrates a need for social support and a preference among cancer caregivers for using social media to meet these needs, it does not identify the prevalence or types of support exchanged on social media. Furthermore, the broad age range of caregivers in prior studies makes it difficult to determine the type of support young caregivers are exchanging on social media. This study addresses this gap in the literature by evaluating five types of social support exchanged on social media with a specific focus on young caregivers, ages 18-39 years. We collected and analyzed content from YACC’s personal social media pages for the immediate six months following their loved one’s cancer diagnosis. We defined any post by a YACC and any comment in response to a post as an “exchange,” and all exchanges were coded for types of social support. As there were some posts that did not contain social support, these were excluded from our analyses as described in the following methods. Our study was guided by the Stress and Coping Social Support theory 82 which posits five different actions that individuals perform to show support and help relieve stress, which are classified into five categories: emotional, information, companionship, validation, and instrumental.67, 68 Definitions of functional social support and examples of their applications to social media are shown in Table 5.1. In this study, functional social support refers to the specific actions that are provided by YACC’s followers on social media.9 Each type of functional social support is intended to address different social needs.67 Theoretically, as social needs are met a YACC’s ability to cope with their caregiving role improves, which in turn likely mediates the relationship between the stress of their caregiving role and their health.67 Our purpose was to determine the extent to which five types of functional social support were exchanged between YACC and their followers by quantifying the types and prevalence of posts containing each type of social support on their personal social media pages during the first six months of caregiving. We learned in prior interviews with YACC that they disclose their caregiving role on social media at varying levels, with some YACC never disclosing their family member’s cancer diagnosis online even though social media is a key source of connection in their lives (see Chapter 4, Results). Therefore, to quantify the exchange of social support we determined that posts on the YACC’s social media pages were eligible for this study regardless of whether they were cancer related or not. Our hypothesis was two-fold. First, we expected that emotional and information support would be the most prevalent types of social support exchanged on YACC’s social media pages over the six months following their loved one’s cancer diagnosis. Second, we expected the prevalence of posts containing all types of social support to decrease over time. 83 This manuscript provides a first look at the role of social media as a venue for social support exchanges between YACC and their social media followers. Given the risk of negative psychosocial health outcomes in the first six months of caregiving for YACC,20 evaluating the types, prevalence, and change in functional social support on social media over the first six months of cancer caregiving among YACC is a first step towards developing social media-based supportive care interventions to support this growing cohort of cancer caregivers. Knowing the types of social support exchanged on social media may help supportive care personnel identify how to aid YACC in seeking support that they do not have access to otherwise.113 Methods This work is part of a larger research study on young adult cancer caregiving in Utah, and was approved by the University of Utah Institutional Review Board. Participants and recruitment Recruitment occurred through flyers, social media, and in-person. Flyers were posted in clinic waiting areas, restaurants, and at a cancer resource library and wellness center at Huntsman Cancer Institute in Salt Lake City, Utah. Social media advertisements were shared by regional, national, and international young adult cancer and caregiving organizations. In-person referrals occurred through in-patient and out-patient settings via the Huntsman Intermountain Adolescent Young Adult Oncology Program, which serves Utah and the Intermountain West and is described elsewhere.114 Cancer patients, who were diagnosed at between six months to five years prior and currently ages 18 and older, 84 were asked to refer their primary caregiver. Eligible caregivers were currently aged 18-39 years, spoke and understood English, used social media at least once per week (i.e., Facebook, Instagram), and had been providing “hands on” care (e.g., not at a distance) to a loved one with cancer for at least six months. Of 354 cancer patients screened, 61 potential caregivers were identified, 13 were deemed ineligible based on patient’s time since diagnosis (n=3), age (n=4), recurrent cancer diagnosis (n=1), no use of social media (n=2), or other (n=3). This left 48 eligible caregivers; n=8 declined, n=6 were unreachable, and N=34 participated (participation rate=70.8%). Enrolled caregivers completed informed consent and a semistructured telephone interview. Qualitative data collection and management Participants were interviewed by phone about their caregiving experiences, social media use, and preferences for future uses of social media in healthcare settings. The interviews lasted 41-79 minutes, were audio recorded and transcribed. Transcripts were read and quality checked against the audio files prior to analysis. In addition to openended interview questions, 34 close-ended questions were asked about sociodemographics, the cancer patient, and social media use. Survey data were imported into REDCap for storage and cleaning. Sociodemographic and cancer variables Sociodemographics included age at interview, gender, ethnicity, race, employment status, insurance status, marital status, educational status, and caring for others besides patient. Cancer patient variables included patient age at interview, 85 relationship with caregiver, and time since diagnosis. Mixed methods data integration Mixed methods integration occurred at multiple stages throughout the study and at different levels of the approach. We employed a convergent mixed method design.29 Qualitative and quantitative data were collected iteratively to allow for reflexivity in how the social media data were collected and prepared for analysis. During data analysis qualitative and quantitative data were analyzed together so that interview data explained the boundaries of social media for exchanging social support during a cancer experience and how social support exchanges changed over time. Finally, integrating the results of the qualitative and quantitative data enhanced interpretation of the findings in that qualitative data helped to contextualize the quantitative findings. Representative quotes related to change in social support on social media over the first six months of caregiving are presented in combination with the statistical results. Qualitative analysis Using NVivo 11, open coding was applied to 10% of the interviews by ELW and ARW resulting in 321 codes.115 A coding scheme was created in which similar codes were grouped together based on content. The coding scheme was reviewed by all members of the research team and revised after feedback. In second cycle focused coding,115 the revised coding scheme was applied to an additional 10% of the interviews, which were double coded by ELW and ARW. Interrater reliability was “strong” (k=0.88).104 To finalize the coding scheme, codes with zero or negative agreement were 86 refined to minimize discrepancies. The final coding scheme was applied to all 34 interviews. Responses to the question “Has the way you use social media changed at all since you started taking care of someone with cancer?” were analyzed and are presented in combination with the statistical results below to explain why changes in social support occurred over time after the cancer diagnosis and how this influenced YACC’s perceived well-being. Qualitative feedback about changes in social support during the first six months of caregiving that arose in other sections of the interview were also identified via analytic memos, which were created throughout data collection and analysis.84, 116 Qualitative data were synthesized for integration with the quantitative social media data to provide a narrative about the change in social support exchanges from the perspective of YACC, as is often done as part of data integration in mixed methods studies.74, 93, 94 Social media data collection and management Social media data were extracted from each participant’s individual Facebook and Instagram pages manually following the interviews. Social media data were collected from a six-month period, beginning at the date of patient diagnosis. For this analysis, 10% of the social media posts from each participant were randomly sampled, with equal representation from each month after diagnosis, resulting in n=2,298 social media posts in the sample from n=33 participants. One participant had not created any social media posts during the six-month period of data collection even though they regularly logged onto and viewed social media sites. A coding scheme was applied to these posts to indicate: year of post, post origin (whether the post was made by the YACC vs. follower), type of post (original post, response post), cancer related post (yes, no), visual content, 87 indicating a photo or video post (yes, no), shared post, indicating the post was shared to the YACC’s social media page by a follower (yes, no), and platform (Facebook, Instagram). The total number of likes, comments, shares, and word count for each post were also recorded. Data transformation, a mixed-methods technique, was performed to create social support outcome variables from the text of social media posts.74, 93, 94 For this data transformation we used definitions from the Stress and Coping Social Support Theory to code for the presence (yes, no) of five types of functional social support: emotional, information, instrumental, companionship, and validation.67, 68 We coded any post on the caregiver’s “wall” or “feed” frame during the six-month time, so the support could have been directed toward and/or created by the YACC, patient, a follower, and/or family unit. Double coding occurred on 10% of the sample resulting in high interrater reliability (k=0.95). The coding scheme was then applied to the remaining posts by ELW, ARW, and TN. During coding there were n=188 posts that did not contain a codable expression of social support (e.g., posts with hidden content), and n=20 shared posts which predated the cancer diagnosis. For consistency, these n=208 posts were excluded from further analysis resulting in a final sample of N=2090 posts. Social media analysis Descriptive statistics were calculated for the sample of social media posts using Stata 14.2. The number and proportion of posts containing each type of social support were summarized across all participants and by month since diagnosis to assess the prevalence of posts containing each type of social support. Generalized linear models 88 were fit for each type of social support over the six-month period with a binomial family and logit link using the eform command to produce adjusted odds ratios (aOR) and 95% Confidence Intervals (CI), with significance set at p<0.05.117 These models were adjusted for platform type as recent research suggests that social support exchanges differ by platform.118 For this analysis we chose to analyze social support exchanges on an aggregate level to quantify a broad summary of changes in social support over the first six months of caregiving. Results YACC sociodemographic and patient cancer characteristics Average caregiver age was 29.0 years (Standard Deviation (SD): 4.72, range 2138, data not shown). Caregivers were primarily female (70.6%), Non-Hispanic White (91.2%), employed (85.3%), insured (97.1%), married/partnered (73.5%), college graduates or higher (53.0%), and caring for others besides the cancer patient (67.6%, Table 5.2). Cancer patients were on average 37.0 years (SD: 13.77, range 19-76, data not shown) and most commonly a spouse/partner (52.9%) or a parent (17.6% mothers, 5.9% fathers, Table 5.3). YACC perspective on changes in social support exchanges on social media over the first six months of caregiving The following themes were synthesized through qualitative analysis of interview data about the change in social support exchanges on social media over the first six months of caregiving. These findings provide a narrative of YACC’s perspective of 89 changes in social support on social media during this time period. Spike in the first month of caregiving in the extent of social support exchanged online YACC described the first month after the cancer diagnosis as an intense time of communication with their social networks, and their social media pages reflected this showing a substantially higher number of posts exchanging social support during the first month (Figure 5.1). Caregivers had interesting insights as to why their social networks engaged with their initial updates about the cancer patient. For example, followers provided social support on social media because they identified with what the YACC was going through, a form of validation support. A participant described this as, “When you see a post [with] something sad, like “Hey, I have cancer” or “Hey, my brother has cancer” … everyone knows that’s something that could happen to anyone. They engage in those posts because they have feelings for it. It relates to them in some way, shape, or form.” A decline in support was commonly noted in the subsequent months, described by a participant who said, “There’s always like a real fervor right at the beginning and then after three months or so it kind of dies down a little bit…we didn’t need as much support.” Active versus passive use influences the extent of social support exchanged online YACC believed that the frequency of their updates influenced the amount of social support they received on social media. By increasing the frequency of their posts 90 YACC were connecting more with members of their social network by sharing information about the cancer patient, the patient’s prognosis and treatment, and how they as a family unit were coping with cancer. In contrast, YACC felt that when their posts were less frequent or less “shocking” they received less social support. One caregiver described this saying, “When he was first diagnosed, and probably within those six months, I gave a lot of updates on social media. I connected people and it allowed them to know what was going on, and so, then people wanted to help. But, when I don't post because either we're just in this lull or [there’s] not really anything to update, it sounds evil, but people kind of forget about you, even though you're still kind of going through stuff. It's not like the crisis mode of going through stuff.” After treatment becomes routine, caregivers post updates about the cancer patient less frequently either because they are more accustomed to managing caregiving responsibilities, or because they did not experience supportive responses from prior posts. Furthermore, caregivers tended to make updates on their social media platforms based on the patient’s status not their own wellbeing, so the support they exchanged was often directed toward the patient’s wellbeing or the family unit as a whole. YACC who passively used social media for their caregiving role– that is they did not post about cancer on social media – believed that their followers were unable to provide social support related to the diagnosis. YACC who generally felt uncomfortable posting about private happenings in their life did not feel hindered by their lack of social support exchanges on social media. Yet, this was not the case for all participants; some YACC chose to “go outside their comfort zone” and commented that their willingness to open up to their followers in a new way by actively posting about cancer influenced the 91 social support exchanges they experienced on social media. A caregiver described this as, “It was so hard at the beginning…I was just too overwhelmed and worn out with everything else going on to give two craps about what was going on with social media… [social media has] kind of changed me a little bit I guess, helped me be a little bit more open…it breaks the ice a little bit... it was never something I used on a real personal level like this before. So maybe that’s why it was a little uncomfortable too, because I’d never really gotten under that superficial level…and this was definitely a personal deep part of our lives that we shared on Facebook.” YACC who were passive users noted that they gained cancer related information support on social media in more private ways by following cancer and/or caregiving advocacy groups on social media without disclosing the cancer patient’s diagnosis publicly. Social network composition influences the extent of social support exchanged online Interestingly, caregivers noticed declines in social support, primarily from followers they considered acquaintances, shortly after the initial diagnosis. One YACC described this saying, “The casual people that are your casual friends on social media that spike interest when something crazy happens and then they just fall away…” In contrast, close family members and friends continued to provide social support both on social media and offline, because they knew the details of the YACC and patient situation. These close ties were available to provide support regardless of the time since diagnosis. Additionally, YACC who cultivated a larger number of followers during this time period believed that by increasing their number of followers they had more 92 opportunities to exchange social support. One caregiver described this increase saying, “Having a [Facebook] page allows [information] to spread, it’s not just going to people here in the state of Utah, it’s going throughout the nation. It’s gone throughout the world. I mean, we had people from France, Australia, England…Without the Facebook page it wouldn’t be as outreaching, we wouldn’t have got the amount of support that we’re getting now… it gives [us] that light, it gives [us] that hope. You can see all the success stories and you no longer become a statistic.” Suboptimal social support leads to less frequent exchanges Multiple participants revaluated relationships with their followers who failed to respond in a supportive way or provided incongruent support exchanges on social media (e.g., individuals provided information support when emotional support was sought after). One caregiver described this saying, “The crazy thing is cancer definitely shows your true friendship and it shows your true colors.” Individuals who posted about cancer on social media and didn’t receive positive feedback felt frustrated with their social network, and this led them to post less frequently about cancer. A caregiver described her frustration with what she perceived were insincere responses to her social media posts, “…after [I posted about] her diagnosis there were people that I hadn’t talked to in a really long time [who] would react with a crying face or something like that and I would never hear anything else. There were no texts. No phone calls. I didn’t get a message. Like there was no follow-up.” YACC who experienced these negative exchanges acknowledged that by not posting on social media about their cancer experience there was less potential for them to exchange both positive and negative social support with 93 their followers. YACC social media characteristics Of the N=2090 posts analyzed, there was an average of 63.3 posts per participant during the six-month time frame, ranging from 1-359 posts (median=29; SD 88.1, Table 5.3). Total number of likes per post ranged from 0-659 likes with a median=119 (mean=166.4 posts, SD 193.9). Total number of comments per post ranged from 0-105 with a median=14 (mean=30.0, SD 33.3). Total number of shares per post ranged from 01,500,000 with a mean of 3498.6 (SD 53,814.4); however, 98.9% of posts had fewer than five shares and the median was 0 shares. The average word count of posts was 13.5, ranging from 0-848 words (median=6, mean=13.5, SD 34.2). Social media posts ranged in date from 2014 to 2018. 26.9% of posts were made by YACC participants and 73.1% were made by their followers. Only 14.5% of posts were original posts, meaning that the YACC posted the content to their own page, compared to the majority of posts which were responses to these initial posts (85.5%). Over one-third of posts were cancer-related (36.8%) while the remaining posts were general exchanges of social support. Less than half of posts were associated with photo or video content (42.9%). “Shared” content was contained in 9.3% of posts, meaning a post was shared to the YACC’s social media page by a follower or the YACC. Three quarters of posts were made on Facebook (75.4%); 24.6% were from Instagram. 94 Summary of five types of functional social support exchanged between YACC and their followers on social media during the first six months of caregiving Overall, the total number of posts per month declined in the six months following the patient’s diagnosis, with a marked decrease after the first month (Figure 5.1). In Table 5.4, overall the most common type of social support exchanged (e.g., any post or response made between YACC and their followers) was emotional support, which was exchanged in 63.3% of all posts (average 220.5 posts per month, SD 105.8), followed by information (27.7%, average 96.7 posts per month, SD 26.0), validation (15.3%, average 53.2 posts per month, SD 31.1), companionship (5.7%, average 20.0 posts per month, SD 5.7), and instrumental support (1.3%, average 4.7 posts per month, SD 4.5). Change in social support exchanged on social media during the first six months of caregiving The number of social media posts exchanged between YACC and followers decreased for all types of social support over the first six months of caregiving (Figure 5.2). However, the proportions of social media posts with social support varied as shown in Figure 5.3. When controlling for platform, over the six-month period following the patient’s diagnosis the odds of posts containing emotional support decreased significantly (aOR: 0.90, 95%CI 0.85-0.94, p<0.001), while the odds of information (aOR: 1.15, 95%CI 1.09-1.21, p<0.001) and companionship (aOR: 1.12, 95%CI 1.02-1.24, p=0.02) support increased (Table 5.5). 95 Discussion This study evaluated the prevalence of functional social support exchanged on social media among YACC and their followers during the first six months following the cancer patient’s diagnosis. Semistructured interviews provided caregivers’ perspective on the factors that influenced social support changes. Similar to studies of social support on social media among other populations,119 emotional and information support were the most common support types exchanged between YACC and their followers. While we hypothesized that all types of social support would decline over the first six months of caregiving, the proportions of information and companionship support, two of the five types of social support we studied increased during this time period. Compared to other types of support on social media, emotional and information support sustained higher prevalence. This finding reinforces our prior findings that social media platforms may provide a useful structure for YACC to exchange emotional support and information versus other strategies like face to face, over the phone, text messaging, or email (Chapter 4, Results). The decline in emotional support exchanges over the sixmonth period may represent YACC’s adjustment to their caregiving role during this period, but more research is needed to determine why this decline occurs and whether it is an intentional or unexpected change. Caregivers explained that declines in social support occurred after the initial period of adjustment to the cancer experience had subsided. Caregivers recognized that actively posting about their cancer experience enabled their followers to be more responsive on social media. While some caregivers expected a decline in support, others perceived a decline in social support as being forgotten about. When caregivers did not 96 receive the support they hoped for on social media they tended to post less and acknowledged that this meant their followers had fewer opportunities to engage with their social media posts to provide support. Unfortunately, cancer caregivers are not formally recognized in clinical cancer care settings, meaning that their personal care needs often go unmet.120 For example, despite their critical need for help managing emotional stress only 29% of cancer caregivers report being asked by healthcare providers about their own care needs.120 At the same time increasing the availability of supportive care personnel (i.e., social workers, patient navigators) in young adult oncology programs is a critical need across the U.S.121 Our findings highlight an opportunity for supportive care personnel to begin addressing unmet social support care needs among YACC. For example, given the sustained prevalence of emotional support and information exchanged on social media, supportive care personnel may consider evaluating the usefulness of social media as a tool for YACC who express frustration or discomfort communicating their social support needs. An evaluation of this type would require that supportive care personnel recognize and establish that some YACC are not comfortable disclosing cancer information online either due to their own discomfort or reticence from the cancer patient about having their diagnosis disclosed online. This evaluation could occur in the initial stages of the cancer diagnosis when the caregiver and patient are beginning to communicate with their social networks about their support needs. Evaluating the usefulness of YACC using social media to communicate their support needs may help them utilize their social network to the fullest extent possible. Future implementation research is necessary for determining optimal approaches for supportive care personnel to partner with YACC in evaluating the 97 role of social media as a resource for social support. The National Health Information National Trends survey indicates that 64% of American adults highly trust information found on the internet or social media.122 Likewise, our findings suggest that YACC actively use social media to exchange information support, likely in attempts to fulfill the unmet information needs they experience.20 Many hospitals, nonprofits, and cancer advocacy groups have already recognized the opportunity to use social media as a platform for reaching young adults experiencing cancer. We reported in an earlier study that 32% of Instagram posts about young adult cancer originate from these organizations.123 While organizations have an exciting opportunity to effectively disseminate caregiver related information and resources through social media, information seeking on social media brings about potential challenges given the widespread dissemination of misinformation on these platforms.124 Access to information on social media and other internet sources has drastically altered the accessibility of health information in recent decades, yet future studies assessing how YACC evaluate and judge the quality and credibility of cancer and caregiver related information found on social media is urgently needed. The lower prevalence of instrumental, companionship, and validation support we found is likely because certain types of social support do not yield themselves to online exchange. Caregivers who experience unfulfilled support needs in these areas may do so because they do not openly seek these supports, and because it is difficult for their social networks to provide this kind of support on social media. While some YACC may already have adequate offline support in these areas, others may need to seek these social resources in other ways. Supportive care personnel may be able to help YACC more 98 effectively utilize their social networks by coaching them on ways to seek support while recognizing and discussing that unfulfilled support needs in these areas may require the YACC to seek these types of support offline. Offline resources available for YACC may include instrumental (e.g., institutional grants, government assistance, fundraisers), companionship (e.g., local events for young cancer patients and caregivers), and validation (e.g., connecting to caregiving or young adult cancer organizations either through social media, other online formats, or in person) supports. Still, if YACC are unaware of these resources they may continue to be underutilized. There is a notable drop in the prevalence of posts containing social support exchanges on social media after the first month of caregiving, and this phenomenon has been previously described in the caregiving literature and was noted by caregivers in our study.24 Our prior interviews with YACC explained that an initial surge in support occurs when their followers first learn about the YACC’s caregiving role, but then support declines as the caregiver becomes more accustomed to their role and seeks less support, has their support needs fulfilled, or foregoes needed support on social media because their followers have not been responsive. YACC may benefit from learning about this potential decline in support and being coached about how to engage in their social network, or alternatively about resources that are available to fill the void of support that networks are not equipped to sustain over longer periods of time. While we did not assess it in this study, it is likely that followers exchange support with YACC in response to an update about the cancer patient. While YACC may feel supported by these exchanges, their followers may not be aware of the YACC’s own wellbeing, thereby limiting followers’ responsiveness to the support needs of YACC. Furthermore, it is possible that 99 posting too frequently, or not frequently enough, leads to fewer opportunities for support exchanges, and the point at which posting frequency leads to loss of social support among cancer caregivers should be a focus of future research. This study provides an overview of social support exchanges on social media between YACC and their followers. Although we were unable to assess how factors like engagement, composition, and size of social networks influence exchanges of social support, prior research suggests that these factors may be important determinants of a social network’s responsiveness to support requests as the nature of relationships on social media may preclude the type of support that followers are equipped to provide to the caregiver.125 Our sample of posts reflect substantially lower word count, yet higher number of likes and comments per post than prior analyses of Instagram posts related to young adult cancer,123 and reasons for this should be explored in future research. Future work should also assess the dynamic exchanges of support between YACC and their followers through the use of social network analysis to determine the extent to which social networks are responsive to YACC requests for specific types of support.97 Our study is limited in that the sample of social media posts were extracted from a small group of participants recruited primarily in Utah and the Intermountain West. In mixed methods, purposive sampling may limit generalizability.75 However, the information gained by integrating YACC’s perceptions with social media post analyses outweighs potential threats to external validity that may be present with a small sample size. Due to the retrospective nature of the interviews, our interpretations of social support may not accurately reflect how YACC felt when they were newly caring for a cancer patient. However, the narrative synthesis of the interview and the social media 100 data are intended to situate the results in the appropriate context of how YACC interpreted their exchanges of social support with their social media followers. Furthermore, YACC likely spend time on social media during which they are not actively posting, but rather may be reading posts or articles from other followers. These activities may contribute in meaningful ways to a YACC’s wellbeing (either positively or negatively), but we did not examine these activities in the current study. Nontextual social media data (e.g., Facebook stickers, photos) were not analyzed. A further limitation of this work is that we grouped Facebook and Instagram data together for our analyses, and social support may differ by platform type, meaning that future work should aim to distinguish the prevalence of social support exchanges by platform type because some platforms may be better equipped for certain types of support. While the analysis of individual variation in the exchange of social support over time using mixed effects models would provide detail on individual changes in social support, the dynamic nature of social media posts precluded this approach for the current analysis. Future studies that are designed to identify individual-level variation in social support exchanges on social media among YACC are needed. A final consideration is that, for a variety of reasons, some YACC do not disclose their family member’s cancer diagnosis or their role as a cancer caregiver on social media. For these individuals, the extent to which the social support exchanged with their followers during their initial caregiving relates to or benefits them in their role as a cancer caregiver is unknown. In summary, social support exchanges on social media may help YACC facilitate specific types of social support to address different needs.113 Our findings explore the changes in types of functional social support exchanged between YACC and their 101 followers on social media over the first six months after the cancer patient’s diagnosis. Understanding the extent of social support on social media during this time frame is a crucial step to inform targeted supportive care interventions for YACC who use innovative new technology in the era of social media. Interventions that provide coaching to YACC on how to use social media to optimize social support are needed to address the higher levels of unmet need in these areas expressed by YACC compared to older caregivers. Social media-based interventions for YACC would most likely benefit from having offline components such as the inclusion of supportive care personnel, like a patient navigator or social worker. Furthermore, the effect of social support through social media on YACC’s health outcomes has yet to be studied and this is a critical area of future research. 102 Table 5.1. Functional social support definitions and social media examples1 Emotional Sympathy, caring, Comments including: “You are amazing!” acceptance “Keep smiling, stay positive!” Instrumental Transportation, Links to scholarships, fundraising. Posts household chores, child seeking childcare, transportation. care, finance Information Knowledge, Links to articles/videos related to information, advice, caregiving or cancer questions. alternative action Companionship Availability of persons Links to YACC support groups, chats, or to spend time with events. Posts planning activities with others. Validation Feedback, social Comments such as “My mother was comparison diagnosed with breast cancer, too”. 1 Social media posts could be directed at the cancer patient, young adult cancer caregiver, and/or the family unit. 103 Table 5.2. Sociodemographic characteristics of YACC (N=34) Age at interview (years) 18-24 25-29 30-34 35-39 Gender Female Male Ethnicity Non-Hispanic White Hispanic/Latino Race1 White Black/African American Asian AI/AN NH/PI Other Employment status Employed Unemployed Insurance status Insured Uninsured Marital status Married/partner Single/divorced/widowed Educational status High school or less Some college College graduate or more Caring for others besides patient2 Yes No Cancer patient age at interview (years) 18-29 30-39 40-49 50-76 Relationship with caregiver Spouse/partner Mother Father N % 4 13 9 8 11.8 38.2 26.5 23.5 24 10 70.6 29.4 31 3 91.2 8.8 31 1 0 2 0 3 91.2 2.94 0.0 5.9 0.0 8.8 29 5 85.3 14.7 33 1 97.1 2.9 25 9 73.5 26.5 3 13 18 8.8 38.2 53.0 23 11 67.6 32.4 7 19 1 7 20.6 55.9 2.9 20.6 18 6 2 52.9 17.6 5.9 104 Table 5.2 Continued N % Sibling 5 14.7 Other (e.g., cousin, child) 3 8.8 Time since diagnosis 6 months to less than 1 year 12 35.3 1 years to less than 2 years 16 47.1 2 years to less than 5 years 6 17.7 1 Totals do not equal 100% because participants could select more than one option 2 Options included a child under age 18, child over age 18, parent, spouse, or other 105 Table 5.3. Characteristics of social media posts (N=2,090 posts from N=33 YACC) Range Mean Median Standard deviation 1-359 63.3 29 88.1 Total number of social media posts 0-659 166.4 119 193.9 Total number of likes per post 0-105 30.0 14 33.3 Total number of comments per post 0-1,500,000 3498.6 0 53,814.4 Total number of shares per post 0-848 13.5 6 34.2 Word count per post N %1 Year of post 2014 6 0.3 2015 121 5.8 2016 248 11.9 2017 1607 76.9 2018 108 5.2 Post origin YACC participant 563 26.9 Follower 1527 73.1 Type of post Original post 303 14.5 Response post 1787 85.5 Cancer related post Yes 769 36.8 No 1321 63.2 Visual content (video or photo) Yes 896 42.9 No 1194 57.1 Shared post Yes 195 9.3 No 1895 90.7 Platform Facebook 1575 75.4 Instagram 515 24.6 1 Percentages may not equal 100% due to rounding 106 Table 5.4. Prevalence of five types of functional social support exchanged between YACC and their followers on social media during the first six months of caregiving (N=2,090 posts, N=33 YACC) Mean number Total number of posts per of posts1 %1 month SD Emotional 1323 63.3 220.5 105.8 Information 580 27.7 96.7 26.0 Validation 319 15.3 53.2 31.1 Companionship 120 5.7 20.0 5.7 Instrumental 28 1.3 4.7 4.5 1 Totals to not equal N=2,090 or 100% because n=280 posts contained more than one type of social support. Table 5.5. Generalized linear models evaluating the odds of posts containing each type of social support over the six-month period, adjusted for platform type (N=2,090 posts, N=33 YACC) Type of support aOR 95%CI p-value1 Emotional 0.90 0.85 – 0.94 <0.001 Information 1.15 1.09-1.21 <0.001 Instrumental 0.92 0.75-1.14 0.44 Companionship 1.12 1.02-1.24 0.02 Validation 1.04 0.97-1.11 0.22 1 Adjusted for platform type, bold indicates statistical significance at p<0.05. 107 Figure 5.1: Total number of social media posts exchanged between YACC and followers during the six months following the cancer patient’s diagnosis There were n=280 posts which contained more than one type of social support. Figure 5.2: Change in the total number of posts containing each type of social support over time since patient diagnosis (N=2090 posts, N=33 YACC) 108 There were n=280 posts which contained more than one type of social support. Figure 5.3: Change in the total proportion of posts containing each type of social support over time since patient diagnosis (N=2090 posts, N=33 YACC) 109 CHAPTER 6 CANCER RELATED CONTENT AND SOCIAL MEDIA PLATFORM INFLUENCE YACC SOCIAL SUPPORT ON SOCIAL MEDIA Abstract YACC use social media during their cancer experiences to seek social support from online communities. We studied the extent to which social media platform type and presence of cancer-related content in social media posts influenced the social support in responses to YACC’s posts during the first six months of caregiving. Eligible YACC aged 18-39 were caring for a cancer patient diagnosed between six months to five years prior, spoke English and used social media at least once per week. YACC completed an interview and survey, then gave permission for their social media posts to be collected retrospectively from six months following the patient’s diagnosis. We collected original posts from YACC’s Facebook and/or Instagram profiles and all responses to these posts. The total number of likes, comments, and shares, for each original post (e.g., the post which the response was made to), and word count for each response was recorded. Original posts and responses were manually coded for the presence of five types of social support (emotional, information, validation, companionship, instrumental) with definitions derived from the Stress and Coping Social Support Theory. Support variables 110 were transformed to depict the proportion of responses per post that contained each type of social support. Using linear mixed-effects models, we compared the distributions of each type of social support offered in response to YACC’s original posts by platform type (Facebook vs. Instagram) and cancer-related content (no vs. yes). Summary statistics were calculated for YACC sociodemographics, patient cancer factors, and social media use. We analyzed N=1,527 social media responses from N=33 YACC. YACC were 70.6% female, 91.2% non-Hispanic White, and 85.3% employed. YACC were most commonly caring for a spouse/partner (52.9%), or a parent (23.5%). The most common type of social support offered was emotional (72.1%), followed by information (20.4%) and validation (17.2%). There were significantly more responses that contained emotional support on Instagram compared to Facebook (β = 0.25, Standard Error (SE) = 0.09, p = 0.007). There were significantly more responses with cancer-related content that contained validation support (β = 0.20, SE = 0.07, p = 0.002), but fewer responses containing emotional (β = -0.17, SE = 0.07, p = 0.02) and instrumental support (β = 0.06, SE = 0.02, p = 0.001) compared to posts without cancer-related content. Future interventions for YACC should consider how platform features and user characteristics (e.g., close vs. weak ties) change the nature of social support offers. Clinicians and supportive care personnel can guide YACC on strategies that are most likely to result in supportive social media responses. For example, YACC may benefit from guidance on which platforms are set up for users to conveniently offer different types of support, and caution about the overuse of cancer-related terms which may lead to diminishing social support over time. 111 Background Of the 40 million caregivers in the U.S. today, one in four is of the millennial generation.126 Compared to older generations, young adult caregivers are more diverse in terms of ethnicity and sexual and gender identity, more often employed, and equally male and female.126 Millennial caregivers spend 21 hours per week caring for their loved one on average, and over half are alone in their caregiving duties.126 About 1.46 million millennial caregivers take care of a loved one with cancer.6 Millennial caregivers, including young adult cancer caregivers (YACC), are known for their expert use of technology for social connectedness. A large body of evidence shows that social support has protective effects on health, and this is the case for cancer caregivers as well.9-12 YACC are extending their social support networks through social technologies. Social media, like Facebook and Instagram, have fundamentally altered social connection, especially among young adults.26, 33 Individuals experiencing chronic diseases, like cancer, often turn to social media to connect with communities of support, find and share information about the condition, and learn about resources that are available to them.127 Social media persists across many aspects of young adults’ lives, and are increasingly used by young patients and caregivers for sharing their cancer experiences.123, 128, 129 Using social media can improve social support, produce feelings of trust, and help users create and maintain new relationships.26, 27 YACC may be shielded from becoming overburdened in their new caregiving role when they engage with online networks for social support,3, 12 but little is known about factors that influence social support on social media among YACC. Organizations like the National Cancer Institute and LIVESTRONG Foundation 112 emphasize the need for interventions that use new technology, and recognize that social media can be a tool to support YACC.14, 15 Eighty-four percent of cancer caregivers express interest in using social media as a resource for dealing with cancer, 54% endorse a need for emotional support and 77% informational support.40 Cancer caregivers also support the use of social media to help coordinate care and avoid conflicts with the cancer patient.40 While there is support from national organizations and cancer caregivers themselves for using social media to enhance cancer caregiving experiences, more information is needed about factors that influence YACC’s social support on social media to inform the development of future supportive resources. Social media platforms have different audiences and purposes, which may influence the types of social support generated by their users.118 For example, paralinguistic digital affordances (PDAs),130 such as likes, upvotes, and comments, are considered features of social support, but these features appear differently by platform. While PDAs are relatively straightforward to view and track on Facebook, Instagram is primarily focused on visual communication with accompanying PDAs, and Twitter emphasizes connections between individuals but still contains PDAs. Other social media platforms are even more diverse, like LinkedIn, which is primarily for professional connections and Pinterest for sharing and posting, or “pinning” images and videos. These platforms are less focused on tracking PDAs. PDAs have many different interpretations depending on the content and context, but are often described by social media users as signs of status, emotional support, and social connectedness.130 In research with young adults, Hayes et al. describe how seeking some types of social support is considered inappropriate on certain platforms – for example, instrumental support (e.g., fundraising, 113 babysitting) is inappropriate on LinkedIn, while seeking emotional support using “sad” or “gloomy” photos was viewed as undesirable on Instagram.118 The unique structure of each social media platform further influences the relationships users can establish and maintain within those social spaces,118 which may impact the type of support available in that community. Comparison of social support between social media platforms is needed because most literature is limited to Facebook, which is not generalizable to other platforms.131 Whether a post contains cancer-related content or not may also influence the presence of social support in responses to YACC’s posts. Twitter posts related to young adult cancer with detailed experiences, religious references, and empathy were the most likely to receive PDAs, whereas those expressing sadness and anger were less likely to be shared.27 Moreover, users who seek emotional support tend to remain involved in online social communities longer compared to those seeking information support, because engagement differs by these support types.32 Both social media post length and emotional content influenced social support in two prior studies of young adult cancer-related content on social media.123, 132 While these studies suggest that cancer-related content prompts emotional support, no prior studies have assessed how posts with cancer-related content compared to posts that do not contain cancer-related content differ in the types of social support responses they garner. In this study, we evaluated the distribution of social support in responses to YACC’s social media posts by social media platform (Facebook vs. Instagram) and cancer-related content (no vs. yes). We were interested in the support provided by YACC’s followers, so for this analysis we analyzed only responses to their original posts. 114 Because there is very little research in this area, our statistical analysis was exploratory but was driven by the findings from our prior work.123 For this study we had an exploratory research question and a hypothesis. Our research question was “How does platform type influence the social support contained in responses to YACC’s social media posts?” We hypothesized that if a YACC’s social media post contains cancerrelated content, then the responses to their post will contain more social support compared to posts without cancer-related content. Methods Participants and recruitment YACC were recruited through flyers, social media, and in-person referrals. Flyers were posted in waiting areas, restaurants, the cancer resource library, and the wellness center at Huntsman Cancer Institute in Salt Lake City, Utah. Social media advertisements were shared by regional, national, and international young adult cancer and caregiving organizations. In-person referrals occurred through in-patient and outpatient settings via the Huntsman Intermountain Adolescent Young Adult Oncology Program.133, 134 Eligible cancer patients, who were diagnosed between six months to five years prior and currently ages 18 and older, were asked to refer a caregiver. Eligibility criteria for caregivers included being 18-39 years old, English fluency, use of social media (i.e., Facebook, Instagram) at least weekly, and caregiving for a cancer patient for at least six months. Of 354 cancer patients screened for eligibility, 61 potential caregivers were identified, 13 were deemed ineligible based on patient’s time since diagnosis (n=3), age (n=4), recurrent cancer diagnosis (n=1), no use of social media (n=2), and other 115 (n=3). This left 48 eligible caregivers; n=8 declined, n=6 were unable to contact, and N=34 participated (participation rate=70.8%). Enrolled caregivers completed informed consent, a semistructured telephone interview, and provided access to their personal Facebook and Instagram feeds so that the research team could extract their posts and responses to their posts for a six-month period starting on the date of the patient’s diagnosis. This study was approved by the University of Utah Institutional Review Board. Social media data collection and mixed methods integration We used a three-phase exploratory sequential mixed method study design (Figure 6.1).29 The textual social media posts that were extracted from each participant’s individual Facebook and Instagram pages were transformed through manual coding, or quantitizing,74, 93, 94 so that posts could be analyzed using statistical software. We used definitions from the Stress and Coping Social Support Theory to code for the presence (yes, no) of five types of functional social support: emotional, informational, instrumental, companionship, and validation.67, 68 Double coding occurred on 10% of the sample with high interrater reliability (k=0.95) across all variables. The coding scheme was then applied to the remaining posts. We randomly sampled 10% of each participant’s posts, with equal representation from each month after diagnosis, resulting in n=2,298 social media posts in the sample from n=33 participants (n=1 participant had zero social media posts during the six-month period of data collection). During coding, there were n=188 posts that did not contain a codable expression of social support (e.g., posts with hidden content), and n=20 shared posts which predated the cancer diagnosis. For 116 consistency, these n=208 posts were excluded from further analysis resulting in a final sample of N=2090 original posts. For this analysis we limited our data to N=1,527 responses. Sociodemographic, patient cancer factors, and social media variables Sociodemographic variables included age at interview, gender, ethnicity, race, employment status, educational status, and caring for others besides the patient. Cancer patient variables included cancer patient’s age at interview, cancer patient’s relationship with caregiver, and time since cancer patient’s diagnosis. The dataset also included social media variables indicating the presence of cancer-related content in the original post (yes, no) and the type of platform (Instagram, Facebook). The total number of likes, comments, and shares, for each original post (e.g., the post which the response was made to), and word count for each response was recorded. Social support outcome The social support outcome variables were originally binary, indicating the presence or absence of each type of social support. In the original data structure, each row represented either an original social media post or response made by YACC or a follower. For this analysis we focused on responses to original social media posts. To do this, we obtained the proportion of responses per post that contained each type of social support. For example, a single post could generate 50% emotional responses, 30% information responses, and 10% validation responses. Each row in the final data structure represents a single social media post, and includes continuous social support outcome 117 variables with separate variables for each type of social support. Data analysis Descriptive statistics were calculated for YACC sociodemographics, cancer patient factors, and social media characteristics. The prevalence of social media responses that contained each type of social support was summarized. We used linear mixed-effects models to assess the proportion of support type per post while controlling for repeated measures and to allow for individual variation in the number of social media posts.135 We estimated the fixed effect of either platform or cancer-relatedness while controlling for random effects of individual variation between participants. Estimated unstandardized beta coefficients and standard errors were used to evaluate whether 1) posts from Instagram had responses with each type of social support compared to Facebook, and 2) posts with cancer-related content had responses with each type of social support compared to posts with no cancer-related content. Finally, we created figures for each type of social support by platform type and cancer-related content. The figures depict the proportion of social media responses per post that contain each type of social support by platform and whether post was cancer-related. Analyses were performed in Stata 14.2 and R software version 3.6.1, with the significance level set at p < 0.05. Results YACC sociodemographic and cancer patient factors Age groups for N=34 participants included: 18-24 years (11.8%), 25-29 years (38.2%), 30-34 years (26.5%), and 35-39 years (23.5%). There were more female than 118 male caregivers (70.6 vs. 29.4%). The vast majority of caregivers were non-Hispanic White (91.2%), and employed (85.3%). Over half of caregivers were college graduates (53.0%) and were also caring for a child under age 18 (61.8%). Three-quarters of cancer patients were also young adults, with 20.6% being 18-20 years and 55.9% being 30-39 years. Cancer patients tended to be spouses/partners (52.9%) or parents (23.5%), and 82.4% were diagnosed between six months to less than 2 years prior (Table 6.1). Social media factors There were N=1,527 social media responses from N=33 participants. Of these responses, 70.3% were from Facebook, and 29.7% were from Instagram. There were between 1-343 social media responses per participant, with an average of 47.7 responses (SD = 72.6) per participant over the six-month period. Responses varied in terms of the number of likes, comments, shares, and word count (Table 6.2). Over one-third of posts contained cancer-related content (38.3%), and 38.1% contained visual content, either a video or photo. In Table 6.3, regardless of cancer-related content or platform type, emotional support was most commonly present in responses to YACC’s posts (72.1%) followed by information (20.4%), validation (17.3%), companionship (3.1%), and instrumental (1.1%). Exploratory evaluation of social support by social media platform We compared the distribution of each type of social support contained in responses to YACC’s posts on Facebook versus Instagram (Table 6.4). Figures 6.2 119 through 6.6 show the number of social media posts that contain a given proportion of responses with each type of social support, by platform. There were significantly more responses that contained emotional support on Instagram compared to Facebook (β = 0.25, Std. Error 0.09, p = 0.007), but nonsignificant differences by platform for responses that contained the other types of support. Exploratory evaluation of social support by cancer-related content Figures 6.7 through 6.11 show the number of social media posts that contain a given proportion of responses with each type of social support, by cancer-related content. Compared to social media posts without cancer-related content, there were significantly fewer responses to posts with cancer-related content that contained emotional (β = -0.17, Std. Error = 0.07, p = 0.02) and instrumental support (β = -0.06, Std. Error = 0.02, p = 0.001, Table 6.4). There were significantly more responses to posts with cancer-related content that contained validation support (β = 0.20, Std. Error 0.07, p = 0.002), compared to posts without cancer-related content. Cancer-related content did not significantly influence the distribution of responses with information and companionship support compared to posts without cancer-related content. Discussion We compared the distributions of different types of social support contained in responses to YACC’s posts during the first six months of caregiving. This information is needed to inform future research about YACC’s use of social media at the beginning of 120 their caregiving role. These findings can also guide clinicians on how YACC are using social media to shed light on strategies that are most likely to result in social support through social media use. There were significantly more responses with emotional support on Instagram compared to Facebook, but no appreciable differences for other types of support. Among young adults, Instagram is perceived to be a less desirable platform for emotional support;118 however, our findings suggest this may not be the case during especially challenging circumstances, like cancer caregiving. It is possible that the presence of an image or video, which is required on Instagram posts, invokes emotional engagement. However, the presence of a video or image alone was not significantly associated with social support (data not shown), suggesting factors aside from visual appeal drive the differences in emotional support by platform. Another explanation for this difference is that the nature of relationships differs by platform.118 As such, YACC may cultivate more emotionally supportive relationships through Instagram if they use Instagram to connect with their closest friends and family members, compared to relationships on Facebook which tend to consist of weaker ties, who may not be able or willing to provide direct emotional support.118 Relational closeness, or the strength of ties, is an essential area for future research about YACC’s social media use because relationships influence how PDAs are interpreted, potential health benefits, and the type of social support exchanged between users.118, 136 While social network analysis is beginning to be applied to study social media use among young adult cancer patients,128, 137 we are aware of no prior application for studying relational ties and social support in YACC’s networks. We found significantly more validation support for posts with cancer-related 121 content compared to posts without cancer-related content, but fewer responses to posts with cancer-related content had emotional and instrumental support. While it is somewhat surprising that cancer content did not result in more responses with other support types, it is possible that “diminishing returns” may lead to lower engagement.118 In other words, when YACC post too frequently about the cancer experience their social networks may become less responsive. We previously found evidence of diminishing number of posts in exchanges of social support on social media between YACC and their followers, which peaked in the first month after diagnosis and declined over time (see Figure 5.1). Furthermore, YACC’s online networks are likely comprised of other young adults who may lack experience coping with severe illnesses,13 and thus may not feel confident about how to respond to a YACC’s post about their cancer caregiving experience. Although there were no significant differences in the distribution of responses to YACC’s social media posts that contained information support by platform or cancerrelated content, information support was the second most common type of support in responses to YACC’s social media posts. The high prevalence of information support on social media is notable and has been previously documented.138-140 The rise of misinformation exchanged on social media, both benign and severe in content, is a growing public health concern,138 and warrants discussion within the context of our findings. When seeking information support online, users may be influenced by internal (e.g., health literacy) and external factors (e.g., information quality, inconsistency between sources).141 Similarly, YACC’s information support is permeated with misinformation, as we previously documented (see Table 4.3). When YACC go to social media seeking information their support networks may not be equipped with the 122 knowledge or resources to meet these requests. Alternatively, YACC may not seek information, but still receive unsolicited information because that is the type of support their social networks are able to provide on social media mediums. Future research is needed to explain the concordance of information support seeking with support that is provided, because when online networks are responsive to information requests users feel more confident in their social network’s ability to provide informational support,50 and this may improve YACC’s coping. Furthermore, seeking and utilization of online information, particularly that which relates to complex health information, can be convoluted. When YACCs gain information support on social media, they may not be prepared to interpret that information, evaluate its quality, or transform it into actionable knowledge. Prior research suggests that exposure to high-quality health information online is associated with higher health literacy among adolescents,142 and this may be the case for YACC as well. We found no differences in the distributions of posts with companionship support based on platform or cancer-related content, and this is likely because companionship is a less prevalent type of support exchanged on Facebook and Instagram. However, a potential area for future research should be evaluating whether companionship is more common on other platforms, like SnapChat or in messaging tools like WhatsApp, iMessage, and Facebook Messenger. A major focus of accelerating social media research in the health sciences has been on diversifying the social media platforms studied.131 These findings contribute meaningful information about platform differences that can shed light on the development of future social media interventions and supportive services for YACC. 123 Based on this work, our recommendations for future research include studying the impact of the social media audience on social support, and how high- and low-quality information support varies across platforms. Previously, factors such as the type of cancer diagnosis, cancer treatment, hospital admission (e.g., public vs. private hospital), and diagnosis stage, have been identified as mediators of caregivers’ unmet needs, and may also influence YACC’s social support on social media.20 While important for future research, these factors were outside the scope of this analysis. Other factors that may influence YACC social support on social media include gender, given that more female social media users may exchange emotional support than male users;49 however we found no differences by gender (data not shown). Given the paucity of research studying how sociodemographic characteristics influence social support on social media it is imperative that future researchers collect and analyze variables like gender, race/ethnicity, and income that are associated with heightened caregiver burden among general caregiver populations.34 Our findings also have implications for clinical providers and supportive care personnel (e.g., social workers, patient navigators) who interact with YACC. Most importantly, this research suggests that YACC actively use social media as a resource for social support. Taking this support seeking behavior into consideration suggests that providers advise YACC about strategies for using social media to their benefit, like encouraging emotional support through Instagram, while cautioning YACC who post about cancer online about the overuse of cancer-related content in their posts. Given the relatively high prevalence of information support on social media, it is imperative that clinicians and supportive care personnel provide YACC with recommendations for high- 124 quality online resources to use for information support.143 Clinicians can also underscore to YACC that online communication may differ from face-to-face interactions, especially because nonverbal cues (e.g., nodding in agreement, personal touch, etc.) are absent in online interactions,31 and so online support may benefit coping differently than face-to-face support.30 It may be beneficial for clinicians to explain how being physically separated during an online interaction can provide YACC opportunities for deeper self-reflection when discussing their cancer experience.30 At the same time, for some YACC posting about their caregiving experiences online may make them more attuned to caregiving demands and stress, and so emphasizing the availability of in-person supports is essential. Our findings should be interpreted within the context of some methodological limitations. In mixed methods, purposive sampling may limit generalizability, and this is noted as a limitation.75 Our sample contained more posts on Facebook than Instagram, but this reflects the disproportionately higher use of Facebook than Instagram in general.33, 99 We also had a relatively homogenous sample, primarily consisting of urban non-Hispanic White heterosexual women, meaning that the experiences of minority YACC groups may be underrepresented and fundamentally different in terms of social media use for social support compared to our sample. Conclusions To support YACC now and in the future, supportive resources are needed in the spaces, online or physical, where young adults exist.14 With that recognition, our goal was to analyze social media characteristics that influence YACC social support on social 125 media. Therefore, we compared the distributions of posts with responses that contained five types of social support on Instagram versus Facebook, and with cancer-related content versus no cancer-related content. This study contributes to growing evidence that social media plays a vital role in social support for YACC during the initial six months following a loved one’s cancer diagnosis. Future research is needed to fully describe the use of social media among YACC, starting with social network analyses of ties and relationships between users who respond to YACC posts. A continued focus on social media platform variability and the congruency between social support that is sought and support that is received on social media is also necessary for advancing this field of study. Finally, this work establishes a basis for development of future innovative interventions that capitalize on the growing popularity of new social technology. 126 Figure 6.1. Mixed methods exploratory sequential design. Adapted from Creswell & Creswell, 2018. 127 Table 6.1. Sociodemographic and cancer characteristics of YACC (N=34) N Age at interview (years) 18-24 4 25-29 13 30-34 9 35-39 8 Gender Female 24 Male 10 Ethnicity Non-Hispanic White 31 Hispanic/Latino 3 1 Race White 31 Black/African American 1 Asian 0 AI/AN 2 NH/PI 0 Other 3 Employment status Employed 29 Unemployed 5 Educational status High school or less 3 Some college 13 College graduate or more 18 1 Caring for others besides patient Child under 18 21 Other (i.e., child over 18, parent, someone else) 4 Not caring for anyone besides the cancer patient 11 Cancer patient’s age at interview (years) 18-29 7 30-39 19 40-49 1 50-76 7 Cancer patient’s relationship with caregiver Spouse/partner 18 Mother 6 Father 2 Sibling 5 Other (e.g., cousin, child) 3 Time since cancer patient’s diagnosis 6 months to less than 1 year 12 1 years to less than 2 years 16 2 years to less than 5 years 6 1 Totals do not equal 100% because participants could select more than one option % 11.8 38.2 26.5 23.5 70.6 29.4 91.2 8.8 91.2 2.94 0.0 5.9 0.0 8.8 85.3 14.7 8.8 38.2 53.0 61.8 11.7 32.4 20.6 55.9 2.9 20.6 52.9 17.6 5.9 14.7 8.8 35.3 47.1 17.7 128 Table 6.2. Characteristics of responses from social media posts (N=1,527 responses from N=33 participants) Range Mean Median Standard deviation 1-343 47.7 21.5 72.6 Total number of social media responses per participant 0-659 204.4 141 207.7 Total number of likes per post 0-105 32.7 24 31.0 Total number of comments per post 0-1,500,000 2576.9 0 44,955.7 Total number of shares per post 0-848 10.7 6 26.8 Word count per response N % Cancer related post Yes 585 38.3 No 942 61.7 Visual content (video or photo) Yes 582 38.1 No 945 61.9 Platform Facebook 1074 70.3 Instagram 453 29.7 Table 6.3. Prevalence of five types of functional social support on social media during the first six months of caregiving (N=1,527 posts, N=33 YACC) N1 %1 Emotional 1,101 72.1 Information 312 20.4 Validation 265 17.3 Companionship 47 3.1 Instrumental 17 1.1 1 Totals to not equal N=1,527 or 100% because some posts contained more than one type of social support. 129 Table 6.4. Univariate linear mixed effects models comparing social support by platform type and cancer content (N=1,527 responses from N=33 participants) Platform: Instagram vs. Facebook (ref) Emotional Information Companionship Validation Instrumental β1 SE1 DF1 t-value p-value1 0.25 -0.04 -0.02 -0.14 -0.01 0.09 0.09 0.03 0.08 0.02 191 191 191 191 191 2.74 -0.49 -0.65 -1.71 -0.45 0.007 0.63 0.52 0.09 0.65 Cancer content: Cancer-related vs. Not cancer related (ref) Emotional -0.17 0.07 190 -2.29 0.02 Information 0.04 0.07 190 0.57 0.57 Companionship 0.01 0.03 190 0.53 0.59 Validation 0.20 0.07 190 3.10 0.002 Instrumental -0.06 0.02 190 -3.36 0.001 1 Unstandardized beta coefficient (β), Standard Error (SE), Degrees of freedom (DF), bolded values indicate significance at p<0.05. Figure 6.2: Differences in the mean percent of responses to YACC’s social media posts that contain emotional support by platform type 130 Figure 6.3: Differences in the mean percent of responses to YACC’s social media posts that contain information support by platform type Figure 6.4: Differences in the mean percent of responses to YACC’s social media posts that contain companionship support by platform type 131 Figure 6.5: Differences in the mean percent of responses to YACC’s social media posts that contain validation support by platform type Figure 6.6: Differences in the mean percent of responses to YACC’s social media posts that contain instrumental support by platform type 132 Figure 6.7: Differences in the mean percent of responses to YACC’s social media posts that contain emotional support by cancer-related content Figure 6.8: Differences in the mean percent of responses to YACC’s social media posts that contain information support by cancer-related content 133 Figure 6.9: Differences in the mean percent of responses to YACC’s social media posts that contain companionship support by cancer-related content Figure 6.10: Differences in the mean percent of responses to YACC’s social media posts that contain emotional support by cancer-related content 134 Figure 6.11: Differences in the mean percent of responses to YACC’s social media posts that contain instrumental support by cancer-related content 135 CHAPTER 7 SUMMARY Introduction YACC are a growing population in the U.S.,144 currently made up of Millennial and Z generations.45 YACC are digital natives and avid users of the internet and social media.99 Social media use persists across many aspects of YACC’s lives and there is substantial evidence that when young adults become cancer caregivers they turn to social media for support.21, 127, 129, 145-147 YACC have diverse and often demanding caregiving roles. Reasons for this include that YACC are often also caring for young children or aging parents, lack experience with severe illnesses, and have limited resources to draw from during the initial period following a loved one’s cancer diagnosis.2, 38 How YACC use social media to address their social support needs has not previously been studied. This dissertation describes the use of social media for social support among YACC using a convergent mixed-methods design.29 In this design, three aims were integrated to produce merged results.29 First, analysis of semistructured qualitative interviews described YACC’s perceptions of how they used social media after their family member was diagnosed with cancer (Aim 1). The goal of Aim 2 was to quantify the prevalence of five types of functional social support exchanges on social media. To 136 do this, textual social media data were manually extracted from YACC’s social media posts and transformed into five social support outcome variables that indicated the presence or absence of each type of social support. This analysis was supplemented with qualitative results to explain why social support exchanges declined or increased over the first six months of caregiving. Finally, in Aim 3 the social support outcome variables from Aim 2 were transformed to analyze the percentage of responses per post that contained each type of social support. Linear mixed effects models were used to assess the distribution of social media responses that contained each type of social support by platform (Facebook vs. Instagram) and cancer-related content (no vs. yes). This summary chapter begins with a review of important terms that were defined and used throughout the study to aid interpretation of these results. Following is a discussion of how the dissertation findings fit within the growing field of young adult cancer caregiving research and the existing theoretical models that informed this study, the Cancer Family Caregiving Experience Framework and the Stress and Coping Social Support Theory.35, 67, 68 Last, suggestions are made for how this work can guide future research, and potential applications of these findings in clinical settings for YACC are described. Review of Terms A challenge and opportunity of this work was defining terms to accurately describe how social media is used for social support among YACC. These definitions may be useful for future research and discussion about social media by YACC and cancer patients, and may also apply to the experiences of caregivers and patients with other 137 diseases. We defined terms in Table 2.1, and Figure 2.1 shows how the terms we defined appear on Facebook. These definitions and terms were applied similarly on Instagram. Original social media posts were defined as the posts made on YACC’s Facebook feed or Instagram profile, while any subsequent comments on those posts were classified as responses. The analyses on the exchange of social support in Aim 2 included all social media posts and responses made by YACC and their followers. Aim 3 analyses included only responses to posts made by YACC. Social Media Are Important Tools Used by Young Adults to Ease the Burden of Cancer Caregiving Previous caregiving research has documented the emotional burden caregivers face when communicating the cancer patient’s diagnosis, prognosis, and treatment plan to friends and family.148, 149 YACC used social media to ameliorate this burden. Social media helped them reach many people at once so they did not have to repeatedly share information and this was less emotionally burdensome than discussing the cancer patient in person. YACC also believed that social media extended the reach of their support requests to persons whom they would not otherwise have had access to. YACC felt social media extended their reach because they connected with members of their social network with whom they did not frequently speak to in person. These individuals would not have known about the cancer diagnosis and related caregiving needs without social media. Another reason YACC felt social media extended their reach was because they made new connections through social media as a result of posting about their cancer experience. These new connections often provided specialized support because they had also 138 experienced cancer and/or caregiving. Online support resources are becoming increasingly popular among young adults with cancer,150, 151 and may be more easily accessible for young adult caregivers.152 Recently Fletcher et al. emphasized the importance of incorporating new interventions that use new technology, like social media, to support cancer caregivers.35 As described above, YACC participants supported the use of social media for social support. This feedback, coupled with the increasing evidence of social media being used by cancer caregivers,129, 153 emphasizes the importance of increasing access for YACC to use technology-based interventions that address unmet social support needs. Indeed, a 2017 systematic review of web-based interventions for informal caregivers of adults with chronic conditions summarized that technology-based interventions can positively impact a caregiver’s mental health, depression, stress or distress, and anxiety.154 Prior research has demonstrated strengths and limitations of both in-person (e.g., personal connection, but inconvenient) and online support groups (e.g., flexibility, but less personal),152, 155, 156 so making both formats available to YACC may help them take advantage of the potential benefits of these services. All participants in this study used social media for cancer-related purposes regardless of whether they openly posted about their family member’s cancer on social media or not. This suggests that at least some YACC use social media discretely to gain information about cancer. However, we did not study these behaviors explicitly and cannot take a position on whether or how this discrete social media use influences YACC perceptions of social support. The fact that all of our sample used social media does not necessarily mean that all YACC use social media during a cancer experience. During our 139 recruitment on more than one occasion a YACC disclosed that they did not use social media. These individuals’ experiences as cancer caregivers, while important, were not captured in this study based on the inclusion/exclusion criteria. National data suggest that among young adults about 2-3% of the population in the U.S. does not use any social media regularly.99 This means that while social media is an important outlet for the vast majority of YACC, supportive resources must be available and advertised outside of these platforms to reach nonsocial media users. Young Adult Cancer Caregivers Experience Challenges and Benefits of Using Social Media During a Cancer Experience These results suggest there are unintended negative consequences of using social media for social support during young adult cancer caregiving. Social media posts that are optimistic tend to receive more attention than those that express negative emotions in online cancer communities.27 When YACC post about negative aspects of cancer or cancer caregiving their social networks, being largely comprised of other young and emerging adults, may be uncertain about how to respond because they are inexperienced in dealing with severe illness and are unfamiliar with caregiving demands.38 This uncertainty may lead them to respond in an unsupportive manner, provide an incongruent response (e.g., a mismatch of support, like offering emotional support when information support was sought), or they may not respond at all. YACC feedback supported this process (Table 4.3), where unintended consequences of using social media during a cancer experience are described. Unintended 140 consequences of social media use, like negative feedback or social comparison, can negatively affect health and wellbeing in general,157-159 and this may be the case for YACC when they compare the boundaries and limitations of their caregiving situation to the experiences of others who may have more liberty to pursue personal interests, financial freedom, and less emotional burden from having an ill loved one. While this study does not address the impact of unintended consequences on YACC behaviors, prior research regarding college students’ social media use suggests that when responses to their posts are not what they desired they may modify their posts to only focus on positive aspects of life.118 Therefore, if YACC share the negative aspects of caregiving online and do not receive the desired support they may stop expressing these experiences, or they may only discuss offline, both of which could have both positive and negative impacts on YACC wellbeing. Taken together, these findings coupled with extant literature suggest there is much still to learn about unintended and negative consequences of social media use by YACC. Another potential negative consequence of YACC’s online social interactions may be the sharing of clinically inaccurate information, sometimes called misinformation,160 on social media. As much as 70% of online health information may be inaccurate or of low quality.52, 53 When YACC seek information about cancer through social media they may be at risk for consumption of misinformation. Similarly, unsolicited advice from social media may be detrimental to YACC’s coping,54 yet as many as two-thirds of social media posts may include unsolicited advice about disease management.48 While quantifying the prevalence of misinformation and unsolicited advice was outside the scope of this study, YACC participants confirmed the presence of 141 these phenomena in their social media use (Table 4.3). Unintended negative consequences of seeking social support through social media like negative support, incongruence of support, absence of response, inaccurate information, and unsolicited advice should be examined in future research. Future studies in this area could seek to explain what effect negative consequences have on health outcomes of YACC and cancer patients, especially as it relates to implemented behavior and decision-making that is based on misinformation. A further consideration for future work related to information support seeking online is that the relationship between a cancer caregiver and patient likely influences a caregiver’s willingness to seek information and support on social media. YACC in our study described that they would use social media differently depending on their relationship with the cancer patient (Table 4.3). In contrast, parent cancer caregivers tend to avoid online information seeking related to their children’s cancer diagnosis,161 meaning that, when a caregiver is taking care of a spouse, parent, grandparent or other relation they may be less fearful of finding information and feedback about the cancer experience compared to parent caregivers of pediatric cancer patients and this likely influences the extent to which they use social media and other internet resources. Social Support on Social Media Changes Over the First Six Months of Caregiving High prevalence of emotional support has been documented in prior research on social media use related to young adult cancer.27, 146 Similarly, emotional support exchanges between YACC and their social media followers was markedly higher in 142 prevalence than other types of support across the six-month period following the cancer patient’s diagnosis. Furthermore, emotional support was particularly high in prevalence during the first month of caregiving compared to all subsequent months. Information support was the second most common type of social support, but was substantially less prevalent than emotional support across the entire six-month period. While the odds that posts contained emotional support declined after the cancer diagnosis, the odds that posts contained information or companionship support increased (Table 5.5). The change in social support on social media among YACC in this study can be understood through the constructs of the Cancer Family Caregiving Experience Model: the stress process, cancer trajectory, and the contexts of caregiving experience.35 When their loved one is first diagnosed and a YACC becomes a caregiver, they take on the responsibility of communicating about the diagnosis with friends and family members. This is when YACC first turn to social media: to relieve the burden this primary stressor has created. Through the stress process, using social media as a resource for communicating about the cancer diagnosis leads YACC to appraise their situation compared to others, which is often termed social comparison in the social media research literature. This social comparison on social media likely has both positive and negative impacts on a YACC’s health and wellbeing, as past social media research has shown. YACC also appraise the availability of support based on responses to their posts. YACC with robust, congruent, kind support through social media likely experience stress relief and confidence in their ability to cope with being a cancer caregiver. YACC who do not receive minimal support or receive unsupportive responses to their social media posts may appraise their social networks as unsupportive and/or unavailable, and subsequently 143 feel less confident about their ability to cope with caregiving challenges. YACC’s continued use of social media over the first six months of caregiving is most likely highly influenced by the cancer trajectory. Declining social support is often interpreted as negative, but it may be that as YACC become more accustomed to their role they need less support over time as patient’s situation improves, and our qualitative results from Aim 2 suggested this was the case for at least some caregivers. While some cancer patients have acute caregiving needs over a short timeframe, other patients’ needs may require intense caregiving over many months and even years with worsening or improving prognosis. This highly variable path along the cancer trajectory suggests that YACC use social media to seek support during periods of high social need, perhaps during a recurrence, second cancer, poor prognosis, palliative care transition, and end-oflife. The influence of both the stress process and the cancer trajectory on YACC’s social media use is structured within a broader socioeconomic, personal, and healthcare context. Social media use was common during the first six months of cancer caregiving, and all but one participant actively posted on social media during this timeframe, ranging from 0-359 posts per participant (median=29 posts), with the most posts occurring during the first month after the cancer diagnosis (Figure 5.1). However, we took a 10% sample of the social media posts from this time, so it is plausible that YACC are seeking and receiving social support at a much higher prevalence than our findings show. As social media continues to play a role in the daily lives of YACC the relevance of online support should be more clearly recognized in caregiving research and clinical cancer settings. 144 Influence of Social Support on Young Adult Cancer Caregiver Wellbeing While we did not directly evaluate the impact of social support gained from social media use on YACC’s health and wellbeing, the Stress and Coping Social Support Theory suggests that when YACC use social media after a loved one is diagnosed with cancer they either receive support or perceive support is available to them.69 Helpful support on social media improves coping with caregiving stress directly, while simultaneously the perception that support is available through social media improves YACC’s appraisal of their situation and buffers the effect of caregiving stress on health.67, 69 Caregiving is a dynamic experience, and social support gained from social media use is hopefully only one of many sources of support during the caregiving experience. Limitations Due to the dynamic and challenging role of being a YACC, recruitment of participants was challenging. Our recruitment strategy was revised multiple times to reach the current sample size, which was smaller than the original goal of 40 participants. Purposive sampling of caregivers who use social media on a weekly basis limited the results because YACC who chose not to use social media for cancer-related purposes were systematically excluded, although based on our screening only 3% of otherwise eligible participants were excluded based on not their lack of social media use. Although the sample reflected the demographics of Utah’s population,162 there was limited racial/ethnic diversity and so minority caregivers’ experiences are likely 145 underrepresented in these data. YACC with a variety of relationships to the cancer patients (e.g., parents, children, spouses, siblings, cousins) were included, and the qualitative data suggested that the nature of the caregiving relationship influenced the YACC’s decision of whether or not to post about cancer on social media, potentially changing their social support on social media. We limited our data collection to Facebook and Instagram as these were the most common social media platforms used by our sample; however, there are many other social media platforms, including disease specific sites, which may yield more cancer-focused support and reach more diverse audiences. Diversifying the social media platforms that are included in research on YACC is a priority for future studies.118 We were unable to use technology-based textual mining software as originally proposed given the complexity of social media data and software limitations. Thus, manual coding of social media data, which was highly time and resource intensive, limited the scope of the social media data that were analyzed and may have led to Type I error. In other words, the significant findings related to changes in social support prevalence over time and differences in social support by platform type and cancer-content may have been erroneous. Implications for Research In a population without cancer, recent research suggests that online support networks are effective in providing information support to young adult caregivers, but do not fulfill other support needs.163 This, coupled with the fact that information support is the most commonly sought type of support online by caregivers,147 suggests that future interventions should consider multifaceted online and in-person formats to achieve the 146 most impact on broadly improving social support across different kinds of functional social support. Future research on how YACC can capitalize on the positive outcomes of social media, like sustaining emotional and information support, while being cautious of unintended negative consequences of social media, like the spread of misinformation and bad advice about cancer and caregiving is needed. Capitalizing on current infrastructure and personnel in clinical cancer settings may help establish practical solutions to supporting YACC in addressing their unmet social support needs. For example, future interventions that build upon existing patient navigation programs to offer training and skill development for YACC on how to identify and evaluate information support, and how to effectively respond to bad advice when it is shared may help them use social media in new ways that enhance their caregiving experience. Patient navigation may also support YACC by coaching them on making social media posts about cancer and caregiving experiences, including guidance about what kind of language to use, visuals to include, and privacy considerations. The sometimes unpredictable nature of a cancer diagnosis may make caregivers vulnerable to misinformation and bad advice about cancer treatment and causes, so future study of misinformation prevalence and the impact of misinformation on YACC behaviors and actions related to care provided to the cancer patient is needed. Our data beg the question of whether social support is sought less by YACC on social media over time following a cancer diagnosis because as time progresses, they are more accustomed to their caregiving role and need less support, or whether support needs stays the same or increase but social networks are less responsive on social media. If the latter is true, the need for socially supportive interventions is further established, and these alternate 147 sources of support may need to be available both online and in-person to make them most accessible for YACC. Interventions should consider that online health information needs to be targeted for immigrant, racial/ethnic minorities, and non-ELS individuals who have lower confidence in seeking health information online,147 as well as other marginalized groups like sexual minorities and rural or inner-city communities with limited internet access. Implications for Clinical Practice Oncology care teams have the intense responsibility of caring for cancer patients, and because of concerns about patient autonomy, time, financial, and resource constraints of these demanding clinical settings cancer caregiver’s needs are often not formally recognized.49 Social media may be a platform to deliver more complete care for the cancer patient and their family caregivers. These results support the conclusions of others who have suggested that clinicians recognize the important role social media plays in cancer caregivers’ communication about their experience and the cancer patient’s health.129 The high prevalence of social media use suggests YACC are highly capable of using technology to enhance their coping with a cancer experience. This is particularly relevant for supportive care personnel who interact with YACC on social media. This is important because typical supportive services are underutilized by YACC given time and resource constraints. They lack time to go to support groups or to be away from the cancer patient and their families, which often includes young children. Making support services more readily available through social media and other online technologies could enhance use of these services and address YACCs unmet support needs. 148 Conclusion YACC who are often avid users of social media before their loved one is diagnosed with cancer continue to use social media purposefully for social support during the first six months of caregiving. This study addresses critical objectives in young adult cancer caregiving research by establishing the feasibility and limitations of studying social support through social media in this population and by describing social support on social media including the change in social support over time as a caregiver. These findings can inform future interventions that are designed to coach YACC on leveraging social media for sustained functional social support. This research has implications for supporting young adult caregivers of patients with other diseases in social media research and clinical supportive services. APPENDIX A SEMISTRUCTURED INTERVIEW AND SURVEY SCRIPT Date: ___/____/________ □ Consent completed □ Recorder turned on Before we get started, I wanted to thank you for taking time to talk with me today. Please know that I only ask you to share what you are comfortable discussing. This information you share will help us learn how to support young adults who are taking care of a cancer patient. To thank you for talking with me today, you will receive a $5 gift card at the end of the interview. Do you have any questions before we begin? First I have some questions about how your use of social media might intersect with taking care of someone with cancer. 1. Some people don’t really think of themselves as cancer caregivers even though they are taking care of someone with cancer. What do you think about this? a. Probe: What has your experience been like taking care of someone with cancer? 2. Who are the people in your life who give you support? These people might be near to you or far away. a. Probe: Who gives you support when you are taking care of your loved one with cancer? You might get support from someone you know personally, someone you know online or even your close coworkers. b. Probe: What kind of help do these people provide to you? c. Probe: How has this help changed since your loved one was diagnosed with cancer until now? 3. I am interested in learning how using social media might connect with how you take care of someone with cancer. Do you use social media? Probe 1: Has the way you use social media changed at all since you started taking care of someone with cancer? 150 4. What kind of things do you usually post about on social media? 5. How would you describe the time you spend on social media when you are not making a post? For example, you might spend this time doing things like reading/responding to other’s posts, searching for information, or interacting with support groups. 6. Since you started taking care of someone with cancer have you had any negative or uncomfortable experiences with social media? (If needed: For example, you may have received unsolicited advice or bad information?) If yes, please tell me more about that? How did that experience influence your decisions? Probe: How was this experience different than any negative or uncomfortable social media experiences you might have had before you started taking care of someone with cancer? 7. How do you talk about taking care of someone with cancer on social media? Probe: How does this differ from the way you talk about taking care of someone with cancer in general? 8. How do you think social media could be used to improve the cancer experience for patients and families? 9. Are there times when you might prefer to talk about taking care of someone with cancer on social media rather than face-to-face? Why? 10. When it comes to taking care of someone with cancer, are there certain topics you would not feel comfortable discussing on social media? For example, discussing finances. 11. Do you have any other comments related to using social media since you started taking care of someone with cancer? Demographics Before we finish the interview, I would like to ask some questions about you. 12. What is your date of birth? Month: _________ Date: _______ Year:______________ 151 13. What is your current zip code? __________________ 14. Do you consider yourself to be Hispanic/Latino? a. Yes b. No 15. Which of the following describes your race? You may select all that apply. a. White b. Black or African American c. Asian d. American Indian or Alaska Native e. Native Hawaiian/Other Pacific Islander f. Other (please describe):_______________ 16. What is your current living situation? 17. Since your loved one was diagnosed with cancer, where do you get money for living/financial support? a. Support myself- I work b. Support myself- through savings c. Support from government programs, i.e. unemployment d. Spouse supports me e. Parents/family f. Other: _____________ 18. Are you currently employed? a. Yes i. If yes, how many hours per week do you work? b. No 19. In total, what would you say your net income is per year given all of those financial supports? a. Less than $24,999 b. $25,000 - $34,999 c. $35,000 - $49,999 d. $50,000 - $74,999 e. $75,000 or more 20. Do you currently have health insurance coverage? a. Yes  How is your insurance provided? (read prompts if necessary, mark all that apply) i. Employer-sponsored (through a policy offered by employment, from parent or spouse) ii. Individual insurance (through a policy purchased by you/your policy holder) iii. Medicare iv. Medicaid v. Obamacare/Affordable Care Act/State Exchange vi. Military healthcare (TRICARE,/VA/Champ-VA) vii. Don’t know viii. Other (ask them to specify, includes insurance through school) b. No 152 21. What is the highest level of education you have completed? a. Less than high school/GED b. High school graduate/GED c. Some college or technical school d. College graduate e. Master’s degree f. Doctorate/Professional degree 22. What language do you speak at home? 23. What is your current relationship status? a. Single (never married) b. Separated or divorced c. Married d. Registered domestic partnership or civil union e. Widow or widower f. Other (please specify): __________________ 22.a. If you are married or partnered, how many years have you been married or in your present relationship? ______________ 24. What is your gender? a. Female b. Male c. Prefer to self-describe (OR something else): ____________ d. Prefer not to say 25. Do you identify as transgender? a. Yes b. No 26. Do you consider yourself to be? a. Heterosexual/straight b. Lesbian or gay c. Bisexual d. Other (please specify):____________e. Don’t know/not sure f. Prefer not to answer Now I have some questions related to [name of care recipient] and their cancer diagnosis. 27. How old is [name of care recipient]? Age in years: ____________ 28. How long ago was [name of care recipient] diagnosed? 153 29. What age were you when [name of care recipient] was first diagnosed with cancer? Age in years: ____________ 30. What type of cancer was [name of care recipient] diagnosed with? Type of cancer: _____________________ 31. What is your relationship to [name of care recipient]? a. Spouse/partner b. Mother/step-mother c. Father/step-father d. Grandmother/step-grandmother e. Grandfather/step-grandfather f. Sibling g. Friend h. Other: _________________ 32. How long have you been in this relationship? _____________________ months _____________________ years 33. Are you currently taking care of any other persons? a. Yes If Yes, who are you also caring for? (Select all that apply) b. No A child under 18 years A child over 18 years A parent A spouse Someone else: _______________________ The following questions ask about your relationship with [the care recipient]. 34. How important is [the care recipient] to you? a. Not at all important b. A little important c. Somewhat important d. Moderately important e. Very important f. Extremely important 35. Ignoring any upsetting aspects of the relationship, how positive or supportive is [the care recipient] to you? a. Not at all b. A little c. Somewhat d. Moderately e. Very 154 f. Extremely 36. Ignoring any helpful or positive aspects of the relationship, how upsetting is [the care recipient] to you? a. Not at all b. A little c. Somewhat d. Moderately e. Very f. Extremely The next question asks about people who give you personal support. A supportive person is one who is helpful, who will listen to you, or who will back you up when you are in trouble. 37. INSTRUCTIONS: Please look at the following list and decide how much each person (or group of persons is supportive of you at this time in your life. Check (√) your answer. How supportive are these people now: a) Your wife, husband, or significant other person……….. b) Your children or grandchildren…………………………… c) Your parents or grandparents……………………………. d) Your brothers or sisters…………………………………… e) Your other blood relatives………………………………… f) Your relatives by marriage (for example: in-laws, exwife, exhusband)………………………………………….. g) Your neighbors…………………………………………….. h) Your coworkers……………………………………………. i) Your church members…………………………………….. j) Your other friends………………………………………….. None Some A Lot There is no such person ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ k) Do you have a particular person whom you trust and to whom you can go with personal difficulties? Please circle Yes or No. Yes No 155 l) If you answered “yes”, which of the above types of person is he or she? (for example: child, parent, neighbor) ___________________________________________________________ The next question asks about people who cause personal stress. A person who stresses you is one who causes problems for you or makes your life more difficult. 38. INSTRUCTIONS: Please look at the following list and decide how much each person (or group of persons is a stress for you at this time in your life. Check (√) your answer. How stressed do you feel by these people now: a) Your wife, husband, or significant other person………… b) Your children or grandchildren……………………………. c) Your parents or grandparents…………………………….. d) Your brothers or sisters……………………………………. e) Your other blood relatives…………………………………. f) Your relatives by marriage (for example: in-laws, exwife, exhusband)…………………………………………… g) Your neighbors……………………………………………… h) Your coworkers…………………………………………….. i) Your church members……………………………………… j) Your other friends…………………………………………... None Some A Lot There is no such person ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ k) Do you have a particular person who is causing you the most personal stress now? Please circle Yes or No. Yes No 156 l) If you answered “yes”, which of the above types of person is he or she? (for example: child, parent, neighbor) ___________________________________________________________ In this last section, I want to ask you about your social media use. For this section, please think about how often you use social media on a weekly basis. 39. Please indicate which social media platforms you currently use, and how often you use them: Platform More than once per day Once per day A few times per week Once per week A few times per month Do not use Facebook Twitter Instagram Pinterest Vine SnapChat LinkedIn YouTube Google+ Other 40. How much time per day do you spend on social media? _____________minutes ___________hours 41. If you need help, which of the following do you go to for the following types of support? Information support (Info): Someone who shares knowledge, information, advice, suggest alternative actions. Emotional support (Emo): Someone who provides sympathy, caring, acceptance. 157 Instrumental support (Inst): Someone who provides/assists with transportation, household chores, child care, finances. Companionship support (Comp): Someone who spends time with you. Validation (Val): Someone who provides feedback and makes social comparisons. Source Family, in person Family, online/social media Family, phone/video call Friend, in person Friend, online/social media Friend, phone/video call Healthcare professional Community leader (priest, bishop, teacher) Other Internet (Google, newspaper, health website) Other Information Emotion Instrumental Companionship Validation If you marked “Other” for any type of support, please list below the source(s) you use: ______________________________________________________________________ 42. Before we finish the interview, is there anything else you’d like to tell me about your caregiving experience? 43. Do you have any other questions or comments? 44. May we contact you for future research studies? a. Yes i. What is the best way for us to contact you? ii. Phone:____-______-______________ iii. Email: 158 b. No iv. Other: This is the end of our interview. Thank you very much for your time and feedback. □ Obtain address for mailing/emailing gift card. Mailing Address: _________________________________________ Email Address: __________________________________________ Notes: APPENDIX B SOCIAL MEDIA SCREENING FORM To complete the study, your social media posts from the previous six months will be sampled and maintained for the research. In the research, you will not be identified by your personal information such as your name, age, location, and username. Your personal information will not be shared with anyone outside the research team unless required by law. We will be vigilant to ensure that your identity is not linked in any way to your social media data. We are only requesting temporary access to your social media accounts, so the researchers will not remain linked to you in anyway on social media after data collection is finished. 1. Please indicate your username for Facebook: __________________________ a. Is your Facebook account a private account? Yes No If yes, you will be contacted by the research team for temporary access to your social media profile in order for data collection to occur. 2. Please indicate your username for Twitter: ____________________________ a. Is your Twitter account a private account? Yes No If yes, you will be contacted by the research team for temporary access to your social media profile in order for data collection to occur. 160 3. Please indicate your username for Instagram: ____________________________ a. Is your Instagram account a private account? Yes No If yes, you will be contacted by the research team for temporary access to your social media profile in order for data collection to occur. 161 REFERENCES 1. Caregiving in the U.S.: AARP Public Policy Instiute and the National Alliance for Caregiving, 2015. 2. Levine C, Hunt GG, Halper D, Hart AY, Lautz J, Gould DA. Young Adult Caregivers: A First Look at an Unstudied Population. Am J Public Health. 2005;95: 2071-2075. 3. Do EK, Cohen SA, Brown MJ. Socioeconomic and demographic factors modify the association between informal caregiving and health in the Sandwich Generation. BMC Public Health. 2014;14: 362. 4. Kim H. Enacted Social Support on Social Media and Subjective Well-being. 2014. 5. Ownsworth T, Henderson L, Chambers SK. Social support buffers the impact of functional impairments on caregiver psychological well-being in the context of brain tumor and other cancers. Psychooncology. 2010;19: 1116-1122. 6. Kahriman F, Zaybak A. Caregiver Burden and Perceived Social Support among Caregivers of Patients with Cancer. Asian Pac J Cancer Prev. 2015;16: 3313-3317. 7. Shieh SC, Tung HS, Liang SY. Social support as influencing primary family caregiver burden in Taiwanese patients with colorectal cancer. J Nurs Scholarsh. 2012;44: 223-231. 8. Nijboer C, Tempelaar R, Triemstra M, van den Bos GA, Sanderman R. The role of social and psychologic resources in caregiving of cancer patients. Cancer. 2001;91: 10291039. 9. Reblin M, Uchino BN. Social and Emotional Support and its Implication for Health. Curr Opin Psychiatry. 2008;21: 201-205. 10. Determinants of Health. Available from URL: https://www.healthypeople.gov/2020/about/foundation-health-measures/Determinants-ofHealth#social [accessed 10/08/2016]. 11. Well-Being Concepts. Available from URL: http://www.cdc.gov/hrqol/wellbeing.htm [accessed 10/08/2016]. 162 12. Litzelman K, Kent EE, Rowland JH. Social factors in informal cancer caregivers: The interrelationships among social stressors, relationship quality, and family functioning in the CanCORS data set. Cancer. 2016;122: 278-286. 13. Luong G, Charles ST, Fingerman KL. Better With Age: Social Relationships Across Adulthood. J Soc Pers Relatsh. 2011;28: 9-23. 14. Kent EE, Rowland JH, Northouse L, et al. Caring for caregivers and patients: Research and clinical priorities for informal cancer caregiving. Cancer. 2016;122: 19871995. 15. Adolescent and Young Adult Oncology Progress Review Group. Closing the Gap: Research and Care Imperatives for Adolescents and Young Adults with Cancer: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, LIVESTRONG Young Adult Alliance, 2006. 16. Caregiving in the U.S. 2015-A Focused Look at Caregivers of Younger Adults: National Alliance for Caregiving and the AARP Public Policy Institute, 2015. 17. Warner EL, Kent EE, Trevino KM, Parsons HM, Zebrack BJ, Kirchhoff AC. Social well-being among adolescents and young adults with cancer: A systematic review. Cancer. 2016;122: 1029-1037. 18. Kim Y, Kashy DA, Evans TV. Age and attachment style impact stress and depressive symptoms among caregivers: a prospective investigation. J Cancer Surviv. 2007;1: 35-43. 19. Bakas T, Pressler SJ, Johnson EA, Nauser JA, Shaneyfelt T. Family caregiving in heart failure. Nurs Res. 2006;55: 180-188. 20. Heckel L, Fennell KM, Reynolds J, et al. Unmet needs and depression among carers of people newly diagnosed with cancer. Eur J Cancer. 2015;51: 2049-2057. 21. Dellmann‐Jenkins M, Blankemeyer M, Pinkard O. Young adult children and grandchildren in primary caregiver roles to older relatives and their service needs. Fam Relat. 2000;49: 177-186. 22. Arnett JJ. Learning to Stand Alone: The Contemporary American Transition to Adulthood in Cultural and Historical Context. Human Development. 1998;41: 295-315. 23. Shifren K. Early caregiving and adult depression: good news for young caregivers. Gerontologist. 2001;41: 188-190. 24. Hirst M. Carer distress: a prospective, population-based study. Soc Sci Med. 2005;61: 697-708. 25. Social Networking Fact Sheet. Available from URL: http://www.pewinternet.org/fact-sheets/social-networking-fact-sheet/ [accessed 06/13/2016]. 163 26. Social networking sites and our lives. Available from URL: http://www.pewinternet.org/2011/06/16/social-networking-sites-and-our-lives/ [accessed 06/28/2016]. 27. Myrick JG, Holton AE, Himelboim I, Love B. #Stupidcancer: Exploring a Typology of Social Support and the Role of Emotional Expression in a Social Media Community. Health Commun. 2016;31: 596-605. 28. Moorhead AS, Hazlett ED, Harrison L, Carroll KJ, Irwin A, Hoving C. A New Dimension of Health Care: Systematic Review of the Uses, Benefits, and Limitations of Social Media for Health Communication. J Med Internet Res. 2013;15: e85. 29. Creswell JW, Creswell JD. Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Los Angeles: SAGE, 2018. 30. Rains SA, Brunner SR, Akers C, Pavlich CA, Goktas S. Computer-mediated communication (CMC) and social support Testing the effects of using CMC on support outcomes. J Soc Pers Relatsh. 2016: 0265407516670533. 31. Rains SA, Young V. A Meta-Analysis of Research on Formal Computer-Mediated Support Groups: Examining Group Characteristics and Health Outcomes. Hum Commun Res. 2009;35: 309-336. 32. Wang YC, Kraut RE, Levine JM. Eliciting and Receiving Online Support: Using Computer-Aided Content Analysis to Examine the Dynamics of Online Social Support. J Med Internet Res. 2015;17. 33. Duggan M. The Demographics of Social Media Users: Internet, Science & Tech PEW Research Center, 2015. 34. Adelman RD, Tmanova LL, Delgado D, Dion S, Lachs MS. Caregiver burden: a clinical review. JAMA. 2014;311: 1052-1060. 35. Fletcher BS, Miaskowski C, Given B, Schumacher K. The cancer family caregiving experience: an updated and expanded conceptual model. Eur J Oncol Nurs. 2012;16: 387398. 36. Intervening with Cancer Caregivers to Improve Patient Health Outcomes and Optimize Health Care Utilization (R21). Available from URL: http://grants.nih.gov/grants/guide/pa-files/PAR-16-318.html [accessed 10/08/2016]. 37. Arnett JJ. Conceptions of the transition to adulthood among emerging adults in American ethnic groups. New Dir Child Adolesc Dev. 2003: 63-75. 38. Arnett JJ. Emerging adulthood. A theory of development from the late teens through the twenties. Am Psychol. 2000;55: 469-480. 164 39. Ridings CM, Gefen D. Virtual Community Attraction: Why People Hang Out Online. J Comput Mediat Commun. 2004;10. 40. Badr H, Carmack CL, Diefenbach MA. Psychosocial interventions for patients and caregivers in the age of new communication technologies: opportunities and challenges in cancer care. J Health Commun. 2015;20: 328-342. 41. Goode WJ. A Theory of Role Strain. American Sociological Review. 1960;25: 483496. 42. Mazur E, Kozarian L. Self-presentation and interaction in blogs of adolescents and young emerging adults. J Adolesc Res. 2010;25: 124-144. 43. Swanberg JE. Making it work: informal caregiving, cancer, and employment. J Psychosoc Oncol. 2006;24: 1-18. 44. Suh J. Measuring the "Sandwich": Care for Children and Adults in the American Time Use Survey 2003-2012. J Fam Econ Issues. 2016;37: 197-211. 45. Dimock M. Defining generations: Where Millenials end and Generation Z begins. Available from URL: https://www.pewresearch.org/fact-tank/2019/01/17/wheremillennials-end-and-generation-z-begins/ [accessed 07/23/2019]. 46. Fry R. Millenials overtake Baby Boomers as America's largest generation. Available from URL: http://www.pewresearch.org/fact-tank/2016/04/25/millennials-overtake-babyboomers/ [accessed 09/21/2019]. 47. Duggan M, Ellison N, Lampe C, Lenhart A, Madden M. Social Media Update 2014. PewResearchCenter, 2015. 48. Greene JA, Choudhry NK, Kilabuk E, Shrank WH. Online social networking by patients with diabetes: a qualitative evaluation of communication with Facebook. J Gen Intern Med. 2011;26: 287-292. 49. Chen X, Yang L, Toso-Salman J, CHang Y, Schear R, McGoldrick D. Social support within online communities: Internet reach and content analysis of a cancer anti-stigma Facebook page in Mexico. Global Media J. 2016;S1. 50. Rains SA. Language style matching as a predictor of perceived social support in computer-mediated interaction among individuals coping with illness. Communic Res. 2016;43: 694-712. 51. Rains SA, Wright KB. Social Support and Computer-Mediated Communication: A State-of-the-Art Review and Agenda for Future Research. Annals of the International Communication Association. 2015;40: 175-211. 165 52. Eysenbach G, Powell J, Kuss O, Sa ER. Empirical studies assessing the quality of health information for consumers on the world wide web: a systematic review. JAMA. 2002;287: 2691-2700. 53. Young SD. Recommendations for Using Online Social Networking Technologies to Reduce Inaccurate Online Health Information. Online J Health Allied Sci. 2011;10. 54. Lakey B. Social Support Theory. Social support measurement and intervention: A guide for health and social scientists. 2000: 29. 55. Kim Y, Baker F, Spillers RL, Wellisch DK. Psychological adjustment of cancer caregivers with multiple roles. Psychooncology. 2006;15: 795-804. 56. Weitzner MA, Haley WE, Chen H. The family caregiver of the older cancer patient. Hematol Oncol Clin North Am. 2000;14: 269-281. 57. Schumacher KL, Beidler SM, Beeber AS, Gambino P. A transactional model of cancer family caregiving skill. ANS Adv Nurs Sci. 2006;29: 271-286. 58. Matthews B, Baker F, Spillers R. Family caregivers and indicators of cancer-related distress. Psychol Health Med. 2003;8: 46-56. 59. Nijboer C, Triemstra M, Tempelaar R, Mulder M, Sanderman R, van den Bos GA. Patterns of caregiver experiences among partners of cancer patients. Gerontologist. 2000;40: 738-746. 60. Gaugler JE, Hanna N, Linder J, et al. Cancer caregiving and subjective stress: a multisite, multi-dimensional analysis. Psychooncology. 2005;14: 771-785. 61. Gaugler JE, Linder J, Given CW, Kataria R, Tucker G, Regine WF. The proliferation of primary cancer caregiving stress to secondary stress. Cancer Nurs. 2008;31: 116-123; quiz 124-115. 62. Doumit MA, Huijer HA, Kelley JH, Nassar N. The lived experience of Lebanese family caregivers of cancer patients. Cancer Nurs. 2008;31: E36-42. 63. Northouse LL, Mood DW, Montie JE, et al. Living with prostate cancer: patients' and spouses' psychosocial status and quality of life. J Clin Oncol. 2007;25: 4171-4177. 64. Schumacher KL, Stewart BJ, Archbold PG. Mutuality and preparedness moderate the effects of caregiving demand on cancer family caregiver outcomes. Nurs Res. 2007;56: 425-433. 65. Verdonck-de Leeuw IM, Eerenstein SE, Van der Linden MH, Kuik DJ, de Bree R, Leemans CR. Distress in spouses and patients after treatment for head and neck cancer. Laryngoscope. 2007;117: 238-241. 166 66. Redinbaugh EM, Baum A, Tarbell S, Arnold R. End-of-life caregiving: what helps family caregivers cope? J Palliat Med. 2003;6: 901-909. 67. Cohen S, Underwood L, Gottlieb B. Social support measurement and intervention: A guide for health and social scientists. New York, NY. Oxford University Press, 2000. 68. Lazarus R, Folkman S. Stress, Appraisal, and Coping. New York, NY: Springer Publishing Company, Inc., 1984. 69. Lazarus RS, Folkman S. Stress, appraisal, and coping. Springer publishing company, 1984. 70. Meng J, Martinez L, Holmstrom A, Chung M, Cox J. Research on Social Networking Sites and Social Support from 2004 to 2015: A Narrative Review and Directions for Future Research. Cyberpsychol Behav Soc Netw. 2017;20: 44-51. 71. Lowe W. Yoshikoder: An Open Source Multilingual Content Analysis Tool for Social Scientists. University of Nottingham: Methods and Data Institute, 2006. 72. Morse J, Niehaus L. Mixed Method Design: Principles and Procedures (Developing Qualitative Inquiry). Walnut Creek, CA. Routledge, 2009. 73. Zhang W, Creswell J. The use of "mixing" procedure of mixed methods in health services research. Med Care. 2013;51: e51-57. 74. Sandelowski M, Voils CI, Knafl G. On Quantitizing. J Mix Methods Res. 2009;3: 208-222. 75. Sandelowski M. Combining qualitative and quantitative sampling, data collection, and analysis techniques in mixed-method studies. Res Nurs Health. 2000;23: 246-255. 76. Burton LC, Zdaniuk B, Schulz R, Jackson S, Hirsch C. Transitions in spousal caregiving. Gerontologist. 2003;43: 230-241. 77. Warner EL, Parnell K, Cohen AL, Colman H, Reblin M. Feasibility testing of ecomapping in neuro-oncology caregivers to identify social support networks. Boston, MA: Poster Preseantation at the American Society of Clinical Oncology Quality Care Symposium, 2014. 78. Nicholls W, Patterson P, McDonald FE, Hulbert-Williams NJ. Unmet needs in young adults with a parent with a chronic condition: a mixed-method investigation and measure development study. Scand J Caring Sci. 2016. 79. Guest G, Bunce A, Johnson L. How many interviews are enough? An experiment with data saturation and variability. Field Meth. 2006: 59-82. 80. Nam JM. A simple approximation for calculating sample sizes for detecting linear trend in proportions. Biometrics. 1987;43: 701-705. 167 81. Campo RA, Uchino BN, Holt-Lunstad J, Vaughn A, Reblin M, Smith TW. The assessment of positivity and negativity in social networks: the reliability and validity of the social relationships index. J Community Psychol. 2009;37: 471-486. 82. Parkerson GR, Broadhead WE, Tse CK. Validation of the Duke Social Support and Stress Scale. Family Medicine. 1991;23: 357-360. 83. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42: 377-381. 84. Birks M, Chapman Y, Francis K. Memoing in qualitative research probing data and processes. J Res Nurs. 2008;13: 68-75. 85. Mays N, Pope C. Assessing quality in qualitative research. BMJ. 2000;320: 50. 86. Glaser BG. The Constant Comparative Method of Qualitative Analysis. Social Problems. 1965;12: 436-445. 87. Saldana J. The Coding Manual for Qualitative Researchers. 2nd ed: London, UK. SAGE Publications Ltd, 2012. 88. Hruschka DJ, Schwartz D, John DCS, Picone-Decaro E, Jenkins RA, Carey JW. Reliability in coding open-ended data: Lessons learned from HIV behavioral research. Field Methods. 2004;16: 307-331. 89. Graham T, Ackland R. SocialMediaLab: Tools for Collecting Social Media Data and Generating Networks for Analysis. Available from URL: https://cran.rproject.org/web/packages/SocialMediaLab/index.html [accessed 11/03/2016]. 90. Agichtein E, Castillo C, Donato D, Gionis A, Mishne G. Finding high-quality content in social media. Proceedings of the 2008 international conference on web search and data mining: ACM, 2008:183-194. 91. De Choudhury M, Lin Y-R, Sundaram H, Candan KS, Xie L, Kelliher A. How does the data sampling strategy impact the discovery of information diffusion in social media? ICWSM. 2010;10: 34-41. 92. Singer J, Willet J. Longitudinal Research: Present Status & Future Prospects: Harvard Graduate School of Education, 2005. 93. Fetters MD, Curry LA, Creswell JW. Achieving Integration in Mixed Methods Designs—Principles and Practices. Health Serv Res. 2013;48: 2134-2156. 94. Bazeley P. Editorial: Integrating Data Analyses in Mixed Methods Research. J Mix Methods Res. 2009;3: 203-207. 168 95. Creswell J, Plano Clark V. Designing and conducting mixed methods research. 2nd ed. Washington, DC: Sage, 2011. 96. Luarn P, Kuo HC, Chiu YP, Chang SC. Social support on Facebook: The influence of tie strength and gender differences. Int J Electron Comm. 2015;6: 37-50. 97. Pilisuk M, Parks SH. The Place of Network Analysis in the Study of Supportive Social Associations. Basic and Applied Social Psychology. 1981;2: 121-135. 98. Kim Y, Schulz R. Family caregivers' strains: comparative analysis of cancer caregiving with dementia, diabetes, and frail elderly caregiving. J Aging Health. 2008;20: 483-503. 99. Villanti AC, Johnson AL, Ilakkuvan V, Jacobs MA, Graham AL, Rath JM. Social Media Use and Access to Digital Technology in US Young Adults in 2016. J Med Internet Res. 2017;19: e196. 100. Smith A, Anderson M. Social Media Use in 2018: Pew Research Center, 2018. 101. A. BD, Perez AG. The Relationship Between Online Social Networking and Depression: A Systematic Review of Quantitative Studies. Cyberpsychology, Behavior, and Social Networking. 2016;19: 638-648. 102. Tromholt M. The Facebook Experiment: Quitting Facebook Leads to Higher Levels of Well-Being. Cyberpsychol Behav Soc Netw. 2016;19: 661-666. 103. Fallon EA, Driscoll D, Smith TS, Richardson K, Portier K. Description, characterization, and evaluation of an online social networking community: the American Cancer Society’s Cancer Survivors Network®. Journal of Cancer Survivorship. 2018;12: 691-701. 104. McHugh ML. Interrater reliability: the kappa statistic. Biochemia Medica. 2012;22: 276-282. 105. Soklaridis S. The process of conducting qualitative grounded theory research for a doctoral thesis: experiences and reflections. . The Qualitative Report. 2009;14: 718-733. 106. Caregiving in the U.S. A Focused Look at Caregivers of Younger Adults: National Alliance for Caregiving and the AARP Public Policy Institute, 2015. 107. Aneshensel CS, Pearlin LI, Schuler RH. Stress, role captivity, and the cessation of caregiving. J Health Soc Behav. 1993;34: 54-70. 108. Nijboer C, Tempelaar R, Triemstra M, van den Bos GAM, Sanderman R. The role of social and psychologic resources in caregiving of cancer patients. Cancer. 2001;91: 1029-1039. 169 109. McCarthy MC, McNeil R, Drew S, et al. Psychological Distress and Posttraumatic Stress Symptoms in Adolescents and Young Adults with Cancer and Their Parents. J Adolesc Young Adult Oncol. 2016. 110. Reblin M, Uchino BN. Social and emotional support and its implication for health. Current opinion in psychiatry. 2008;21: 201-205. 111. Prensky M. Digital natives, digital immigrants. Part 1. On the horizon. 2001;9: 1-6. 112. Turner A. Generation Z: Technology and social interest. The Journal of Individual Psychology. 2015;71: 103-113. 113. Falisi AL, Wiseman KP, Gaysynsky A, Scheideler JK, Ramin DA, Chou W-yS. Social media for breast cancer survivors: a literature review. Journal of Cancer Survivorship. 2017;11: 808-821. 114. Warner EL, Fowler B, Pannier ST, et al. Patient Navigation Preferences for Adolescent and Young Adult Cancer Services by Distance to Treatment Location. J Adolesc Young Adult Oncol. 2018;7: 438-444. 115. Saldana J. The Coding Manual for Qualitative Researchers. London, UK: SAGE, 2013. 116. Mays N, Pope C. Assessing quality in qualitative research. British Medical Journal. 2000;320: 50. 117. glm - Generalized linear models. Available from URL: https://www.stata.com/manuals13/rglm.pdf [accessed 07/31/2019]. 118. Hayes RA, Carr CT, Wohn DY. It’s the Audience: Differences in Social Support Across Social Media. Social Media + Society. 2016;2: 2056305116678894. 119. Chou W-yS, Moskowitz M. Social media use in adolescent and young adult (AYA) cancer survivors. Current Opinion in Psychology. 2016;9: 88-91. 120. Cancer caregiving in the U.S.: National Alliance for Caregiving in partnership with the National Cancer Institute and the Cancer Support Community, 2016. 121. A Narrative Review of Models of Care for Adolescents and Young Adults with Cancer: Barriers and Recommendations. J Adolesc Young Adult Oncol. 2018;7: 148-152. 122. HINTS brief: Trust in health information sources among american adults: Health Information National Trends Survey, 2019. 123. Warner EL, Ellington L, Kirchhoff AC, Cloyes KG. Acquisition of Social Support and Linguistic Characteristics of Social Media Posts About Young Adult Cancer. J Adolesc Young Adult Oncol. 2018;7: 196-203. 170 124. Chou WS, Oh A, Klein WMP. Addressing Health-Related Misinformation on Social Media. JAMA. 2018. 125. Bo L, Scammon DL. We learn from each other: exploring interpersonal communications in online communities. International Journal of Web Based Communities. 2018;14: 289-302. 126. Flinn B. Millenials: The emerging generation of family caregivers: AARP Public Policy Institute, 2018. 127. Social media "likes" healthcare: From marketing to social business: PWC Health Research Institute, 2012. 128. Jones EC, Storksdieck M, Rangel ML. How Social Networks May Influence Cancer Patients' Situated Identity and Illness-Related Behaviors. Front Public Health. 2018;6: 240. 129. Gage-Bouchard EA, LaValley S, Mollica M, Beaupin LK. Cancer Communication on Social Media: Examining How Cancer Caregivers Use Facebook for Cancer-Related Communication. Cancer Nurs. 2017;40: 332-338. 130. Hayes RA, Carr CT, Wohn DY. One Click, Many Meanings: Interpreting Paralinguistic Digital Affordances in Social Media. Journal of Broadcasting & Electronic Media. 2016;60: 171-187. 131. Rains SA, Brunner SR. What can we learn about social network sites by studying Facebook? A call and recommendations for research on social network sites. New Media & Society. 2015;17: 114-131. 132. Crook B, Glowacki EM, Love B, Jones BL, Macpherson CF, Johnson RH. Hanging by a thread: exploring the features of nonresponse in an online young adult cancer survivorship support community. J Cancer Surviv. 2016;10: 185-193. 133. Adolescent and young adult. Available from URL: https://healthcare.utah.edu/huntsmancancerinstitute/wellness-support/patientnavigators/adolescent-young-adult.php [accessed 07/31/2019]. 134. Huntsman-Intermountain Adolescent and Young Adult Cancer Care Program. Available from URL: https://intermountainhealthcare.org/services/cancer-care/huntsmanintermountain/hi-aya/ [accessed 07/31/2019]. 135. Bates D. Fitting linear mixed models in R. R news. 2005;5: 27-30. 136. Rozzell B, Piercy CW, Carr CT, et al. Notification pending: Online social support from close and nonclose relational ties via Facebook. Computers in Human Behavior. 2014;38: 272-280. 171 137. Koltai K, Walsh C, Jones B, Berkelaar BL. Applying Social Network Analysis to Identify the Social Support Needs of Adolescent and Young Adult Cancer Patients and Survivors. J Adolesc Young Adult Oncol. 2018;7: 181-186. 138. Chou W-YS, Oh A, Klein WMP. Addressing Health-Related Misinformation on Social MediaAddressing Health-Related Misinformation on Social MediaAddressing Health-Related Misinformation on Social Media. JAMA. 2018;320: 2417-2418. 139. Gage-Bouchard E, LaValley S, Ms MA, et al. Cancer Communication on Social Media: Examining How Cancer Caregivers Use Facebook for Cancer-Related Communication. Cancer Nurs. 2017;40: 332-338. 140. Escoffery C, Miner KR, Adame DD, Butler S, McCormick L, Mendell E. Internet Use for Health Information Among College Students. Journal of American College Health. 2005;53: 183-188. 141. Lee K, Hoti K, Hughes JD, Emmerton L. Dr Google and the Consumer: A Qualitative Study Exploring the Navigational Needs and Online Health InformationSeeking Behaviors of Consumers With Chronic Health Conditions. J Med Internet Res. 2014;16: e262. 142. Ghaddar SF, Valerio MA, Garcia CM, Hansen L. Adolescent Health Literacy: The Importance of Credible Sources for Online Health Information. Journal of School Health. 2012;82: 28-36. 143. Morahan-Martin DJM. How Internet Users Find, Evaluate, and Use Online Health Information: A Cross-Cultural Review. CyberPsychology & Behavior. 2004;7: 497-510. 144. Durant TJ, Christian OG: Caregiving to Aging Parents. Forum on Public Policy, Available: http://forumonpublicpolicy.com/archive07/durant.pdf. [Accessed 12/09/2018]. 145. Gage-Bouchard EA, LaValley S, Mollica M, Beaupin LK. Cancer Communication on Social Media: Examining How Cancer Caregivers Use Facebook for Cancer-Related Communication. Cancer Nurs. 2016. 146. Warner EL, Ellington L, Kirchhoff AC, Cloyes K. Acquisition of social support and linguistic characteristics of social media posts about young adult cancer. J Adolesc Young Adult Oncol. 2017. 147. HintsBriefs: Health information seeking among caregivers: Health Information National Trends Survey, 2019. 148. Wittenberg E, Borneman T, Koczywas M, Del Ferraro C, Ferrell B. Cancer Communication and Family Caregiver Quality of Life. Behav Sci (Basel). 2017;7. 149. Ewing G, Ngwenya N, Benson J, et al. Sharing news of a lung cancer diagnosis with adult family members and friends: a qualitative study to inform a supportive intervention. Patient Educ Couns. 2016;99: 378-385. 172 150. Kaal SE, Husson O, van Dartel F, et al. Online support community for adolescents and young adults (AYAs) with cancer: user statistics, evaluation, and content analysis. Patient Prefer Adherence. 2018;12: 2615-2622. 151. Keegan THM, Lichtensztajn DY, Kato I, et al. Unmet adolescent and young adult cancer survivors information and service needs: a population-based cancer registry study. J Cancer Surviv. 2012;6: 239-250. 152. Matar JL, Maybery DJ, McLean LA, Reupert A. Web-Based Health Intervention for Young People Who Have a Parent with a Mental Illness: Delphi Study Among Potential Future Users. J Med Internet Res. 2018;20: e10158. 153. Gage-Bouchard EA, LaValley S, Mollica M, Beaupin LK. Communication and Exchange of Specialized Health-Related Support Among People With Experiential Similarity on Facebook. Health Commun. 2017;32: 1233-1240. 154. Ploeg J, Markle-Reid M, Valaitis R, et al. Web-Based Interventions to Improve Mental Health, General Caregiving Outcomes, and General Health for Informal Caregivers of Adults With Chronic Conditions Living in the Community: Rapid Evidence Review. J Med Internet Res. 2017;19: e263. 155. Høybye MT, Dalton SO, Deltour I, Bidstrup PE, Frederiksen K, Johansen C. Effect of Internet peer-support groups on psychosocial adjustment to cancer: a randomised study. British Journal Of Cancer. 2010;102: 1348. 156. Huber J, Muck T, Maatz P, et al. Face-to-face vs. online peer support groups for prostate cancer: a cross-sectional comparison study. J Cancer Surviv. 2018;12: 1-9. 157. Dibb B. Social media use and perceptions of physical health. Heliyon. 2019;5: e00989. 158. Pujazon-Zazik M, Park MJ. To tweet, or not to tweet: gender differences and potential positive and negative health outcomes of adolescents' social internet use. Am J Mens Health. 2010;4: 77-85. 159. Vogel EA, Rose JP, Okdie BM, Eckles K, Franz B. Who compares and despairs? The effect of social comparison orientation on social media use and its outcomes. Personality and Individual Differences. 2015;86: 249-256. 160. Chou WS, Oh A, Klein WMP. Addressing Health-Related Misinformation on Social Media. JAMA. 2018;320: 2417-2418. 161. Gage EA, Panagakis C. The devil you know: parents seeking information online for paediatric cancer. Sociology of Health & Illness. 2012;34: 444-458. 162. QuickFacts: Utah. Available from URL: https://www.census.gov/quickfacts/UT [accessed 08/30/2019]. 173 163. Vaughan C, Trail TE, Mahmud A, Dellva S, Tanielian T, Friedman E. Informal Caregivers’ Experiences and Perceptions of a Web-Based Peer Support Network: MixedMethods Study. J Med Internet Res. 2018;20: e257.