Collaboration and support in diabetes management across adulthood: the roles of of attachment and diabetes distress

Collaboration and support (C&S) from romantic partners is often, but not always, linked with better diabetes management. Attachment may explain when C&S is beneficial. More insecurely attached individuals may benefit more (e.g., higher anxious attachment) or less (e.g., higher avoidant attachment) from higher C&S and these differences may be amplified under higher distress both between and within persons. The objective of the study was to understand how attachment moderates associations between C&S, diabetes distress, and diabetes management (i.e., adherence, glycemic control, daily mean blood glucose). Individuals (N = 199, Mage = 46.35; 52.3% women) with type 1 diabetes in romantic partnerships completed cross-sectional surveys and a 14-day daily diary that assessed adherence, C&S, and distress to examine between- (surveys) and within-person (diary) associations. HbA1c was gathered from in-lab assays and blood glucose gathered from glucometers. Results indicate that higher anxious attachment associated with lower C&S and lower adherence between persons. C&S did not independently associate with diabetes management. Anxious attachment interacted with diabetes distress, such that higher distress was detrimental for glycemic control only for those with lower anxious attachment. Within persons, daily C&S was only beneficial for adherence when accounting for individuals' diabetes-related distress, and stressors and attachment insecurity did not moderate these associations. Results suggest that while attachment insecurity has meaningful implications for diabetes management, it does not moderate C&S and C&S itself does not significantly bolster diabetes management. These findings highlight the need to further examine how attachment insecurity operates to affect diabetes care.

COLLABORATION AND SUPPORT IN DIABETES MANAGEMENT ACROSS ADULTHOOD: THE ROLES OF ATTACHMENT AND DIABETES DISTRESS by Caitlin S. Kelly A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Psychology The University of Utah August 2018 Copyright © Caitlin S. Kelly 2018 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Caitlin S. Kelly has been approved by the following supervisory committee members: Cynthia A. Berg , Chair 05/31/2018 Date Approved Brian R. Baucom , Member 05/31/2018 Date Approved Lisa Diamond , Member 05/31/2018 Date Approved Bert N. Uchino , Member 05/31/2018 Date Approved Vicki S. Helgeson , Member 05/31/2018 Date Approved and by Lisa Aspinwall the Department/College/School of and by David B. Kieda, Dean of The Graduate School. , Chair/Dean of Psychology ABSTRACT Collaboration and support (C&S) from romantic partners is often, but not always, linked with better diabetes management. Attachment may explain when C&S is beneficial. More insecurely attached individuals may benefit more (e.g., higher anxious attachment) or less (e.g., higher avoidant attachment) from higher C&S and these differences may be amplified under higher distress both between and within persons. The objective of the study was to understand how attachment moderates associations between C&S, diabetes distress, and diabetes management (i.e., adherence, glycemic control, daily mean blood glucose). Individuals (N = 199, Mage = 46.35; 52.3% women) with type 1 diabetes in romantic partnerships completed cross-sectional surveys and a 14-day daily diary that assessed adherence, C&S, and distress to examine between- (surveys) and within-person (diary) associations. HbA1c was gathered from in-lab assays and blood glucose gathered from glucometers. Results indicate that higher anxious attachment associated with lower C&S and lower adherence between persons. C&S did not independently associate with diabetes management. Anxious attachment interacted with diabetes distress, such that higher distress was detrimental for glycemic control only for those with lower anxious attachment. Within persons, daily C&S was only beneficial for adherence when accounting for individuals' diabetes-related distress, and stressors and attachment insecurity did not moderate these associations. Results suggest that while attachment insecurity has meaningful implications for diabetes management, it does not moderate C&S and C&S itself does not significantly bolster diabetes management. These findings highlight the need to further examine how attachment insecurity operates to affect diabetes care. iv TABLE OF CONTENTS ABSTRACT ....................................................................................................................... iii LIST OF TABLES ............................................................................................................. vi LIST OF FIGURES .......................................................................................................... vii INTRODUCTION ...............................................................................................................1 Attachment Insecurity and Support .........................................................................2 Diabetes Management, Attachment, and Stress .......................................................3 Between- and Within-Person Effects Regarding Diabetes Support and Stress .......4 Current Study ...........................................................................................................6 METHOD ............................................................................................................................8 Participant Characteristics .......................................................................................8 Procedure .................................................................................................................9 Survey Measures ......................................................................................................9 Plan of Analysis .....................................................................................................14 RESULTS ..........................................................................................................................17 Descriptives and Correlations of Key Variables ....................................................17 Attachment Insecurity Predicting Collaboration and Support or Diabetes Management ...........................................................................................................17 Between-Persons Associations of C&S and the Moderating Roles of Attachment Insecurity and Diabetes Distress on Diabetes Management ..................................18 Within-Person Associations of Collaboration and Support and the Moderating Roles of Attachment Insecurity and Stress on Daily Diabetes Management ........19 DISCUSSION ....................................................................................................................31 The Roles of Attachment Insecurity and Diabetes Distress Between Persons ......31 The Roles of Attachment Insecurity and Diabetes Distress Within Persons .........33 Limitations, Conclusions, and Future Directions ..................................................34 REFERENCES ..................................................................................................................38 LIST OF TABLES Tables 1. Correlations and descriptives of key study variables .....................................................22 2. Associations of attachment on C&S and diabetes management ....................................23 3. Diabetes distress moderating the interaction between attachment and C&S and diabetes management .........................................................................................................24 4. Attachment moderating associations between daily C&S and diabetes management ...25 5. Attachment moderating associations between C&S*stressors and diabetes management .......................................................................................................................26 LIST OF FIGURES Figures 1. Three-way interaction between avoidant attachment, diabetes distress, with HbA1c ...27 2. Two-way interaction between anxious attachment and diabetes distress with HbA1c .28 3. Two-way interaction within-person C&S and anxious attachment on mean blood glucose ...............................................................................................................................29 4. Two-way interaction within-person stressors by anxious attachment on mean blood glucose ...............................................................................................................................30 5. Inverse plot of anxious attachment moderating diabetes distress with HbA1c .............37 INTRODUCTION Of the many relationships individuals have across adulthood, romantic partners are most frequently involved in chronic illness management and serve as sources of support (e.g., August & Sorkin, 2010; Trief, Sandberg, Dimmock, Forken, & Weinstock, 2013). Support is often, but not always, linked with better illness management (Gallant, 2003). How individuals perceive support may reflect longstanding differences in how they perceive close relationships, which can be captured by attachment (e.g., Ognibene & Collins, 1998). Individuals who are more securely attached report higher levels of support available to them compared to individuals with more insecure attachment (Collins & Feeney, 2004; Ognibene & Collins, 1998). Attachment has rarely been examined in the social context of chronic illness management. Yet, attachment insecurity may meaningfully change how perceived support is linked with self-management and illness outcomes, such that more insecurely attached individuals may not perceive or benefit from their partner's support - especially when under distress (e.g., Cohen et al., 2005; Vilchinsky et al., 2010). Utilizing attachment theory to understand who benefits (i.e., high or low attachment insecurity) and when (i.e., under high or low stress) from perceived illness-related support may provide greater insight as to how to best utilize close relationships to improve chronic illness management across adulthood. 2 Attachment Insecurity and Support Attachment theory is important for understanding how individuals in romantic relationships perceive and utilize their partners for support. Attachment styles reflect observable differences in individuals' underlying working models and are characterized as secure or insecure (e.g., Fraley, Waller, & Brennan, 2000). Individuals who are securely attached view attachment figures as reliable sources of support during times of distress (e.g., Reis, Collins, & Berscheid, 2000). Individuals who are insecurely attached are unable to effectively utilize attachment figures as a secure base. Individuals with more anxious attachment use a "maximization" strategy to elicit some of what they need from their attachment figure by increasing proximity and escalating outward signs of distress. Individuals with more avoidant attachment use a "minimization" strategy by distancing themselves and decreasing their outward signs of distress (e.g., Simpson & Rholes, 2012). Individuals who are higher in anxious or avoidant attachment also report lower availability of support and perceive their partner's support as less helpful (Collins & Feeney, 2004; Ognibene & Collins, 1998; Vogel & Wei, 2005). Further, these "working models" of attachment also may lead to differences in how individuals benefit from perceived support within romantic relationships (Simpson & Rholes, 2012). How attachment affects perceptions of support may be most observable during times of distress and may affect support differently for individuals high in anxious versus avoidant attachment. When they are distressed, more anxiously attached individuals may feel that increased support from partners is how they want their partner to respond. More avoidantly attached individuals may feel that increased support is antithetical to their more independent style of managing distress (Simpson & Rholes, 2012). Many studies 3 have relied on inducing stressful situations to examine attachment and supportive processes (e.g., Collins & Feeney, 2000, 2004; Simpson, Rholes, & Nelligan, 1992). Other studies have captured these processes in everyday life in healthy couples (e.g., Campbell et al., 2005; Pietromonaco & Barrett, 1997). In the management of chronic illnesses like type 1 diabetes, illness-related distress may be frequent and the need for support greater than in healthy individuals. Diabetes Management, Attachment, and Stress Type 1 diabetes provides an especially potent context to explore how attachment is linked with perceived support and stress in predicting illness management and outcomes. The management of type 1 diabetes requires adherence to a complex regimen that permeates daily life and is highly visible to romantic partners (e.g., counting carbohydrates, checking blood glucose, and administering insulin before eating dinner together; e.g., Chiang, Kirkman, Laffel, & Peters, 2014). Diabetes-related support involves not only traditional positive aspects of social support, like emotional and tangible support, but also collaboration. Collaboration - working with another person to help solve diabetes problems - is beneficial for bolstering self-management efforts across the lifespan (e.g., Wiebe, Helgeson, & Berg, 2016). The demanding nature of diabetes self-care is challenging. Some individuals experience diabetes-specific distress, a marked worsening of psychological well-being distinct from depressive symptoms (e.g., Polonsky et al., 2005). Encountering many diabetes-related stressful events throughout a day (e.g., underestimating insulin needs before dinner, which results in a high blood glucose reading before bed, and then overcorrecting the high blood glucose, which results 4 in a low blood glucose reading in the middle of the night) may further compound the experience of diabetes distress. As individuals' perceptions of support are affected by attachment (e.g., Campbell et al., 2005; Collins & Feeney, 2004), attachment insecurity also may affect how individuals benefit from perceived diabetes-related support. Cohen and colleagues (2005), in a sample of individuals with type 2 diabetes, found that those higher in avoidant attachment perceived lower levels of diabetes-related support from their partners. Attachment also has associated with relationship quality with diabetes care providers (Cienchanowski et al., 2004). Moreover, individuals with more avoidant-like attachment (i.e., "dismissive") were significantly more likely to experience problematic glycemic control, defined as HbA1c > 8% (Ciechanoswki et al., 2002). Despite the potential for attachment insecurity to affect how individuals perceive illness-related support and collaboration from their romantic partners, and lead to differences in diabetes management, we know little about how these processes play out in the context of type 1 diabetes across adulthood. Between- and Within-Person Effects Regarding Diabetes Support and Stress Perceptions of diabetes support vary not only between but also within individuals across days. Research on perceived support and attachment has largely focused on individual differences (e.g., Collins & Feeney, 2000; Ognibene & Collins, 1998). Diary designs have been used to understand how day-to-day relationship experiences and attachment affect individual well-being within persons in healthy samples (e.g., Campbell 5 et al., 2005; Iida, Stephens, Rook, Franks, & Salem, 2010; Pietromonaco & Barrett, 1997). For example, Campbell and colleagues (2005) found attachment insecurity created differences in individuals' experiences of support from their partners. Those higher in avoidance perceived daily support as less positive and those higher in anxiety saw greater long-term benefits of support, though there were no differences in the amount of support individuals perceived from their partners. Other research in type 2 diabetes management also shows that there is daily variability in how individuals perceive their partner's support (i.e., lower versus higher support days; Iida et al., 2010). Given the potential for daily fluctuations in support and connections to attachment (e.g., Campbell et al., 2005), extending research to include attachment in the context of daily perceived diabetes support and daily management is warranted. Similar to support, the experience of diabetes-related stress also may vary between and within individuals. Diabetes-related distress has generally been captured between-persons (Polonsky et al., 2005) and is associated with worse adherence and glycemic control in individuals with type 1 diabetes (Van Bastelaar et al., 2010). Individuals vary in their daily experiences of diabetes-related stressors (Berg et al., 2013), which may lead to disruptions to daily self-management even if their overall diabetes distress is low. Moreover, for individuals already experiencing high levels of diabetes distress, having days in which many diabetes-related stressful events occur may be especially detrimental for self-management. For individuals with more insecure attachment, experiencing distress may amplify how attachment interacts with perceived support in their associations with diabetes management. That is, more insecurely attached individuals may not experience the same kinds of benefits for diabetes-related 6 support on days in which when they are experiencing more stressful events. Current Study The current study examined how attachment insecurity may be involved in perceptions of diabetes-related collaboration and support (C&S) in a sample of adults with type 1 diabetes in a set of complementary aims. The first set of aims utilized crosssectional data to understand more generally how attachment insecurity, C&S, and diabetes distress associate with diabetes management between persons using survey measures. First, I examined how attachment insecurity associated with perceptions of diabetes-related C&S and diabetes management (Aim 1). I expected that individuals higher in avoidant and anxious attachment would perceive lower C&S from their partners, and individuals higher in avoidant and anxious attachment also would report worse diabetes management (i.e., lower adherence; higher HbA1c). Next, I examined how patients' attachment insecurity moderated associations between perceived C&S and diabetes management (Aim 2a), and whether higher diabetes distress strengthened the moderating effect of attachment insecurity (Aim 2b). I expected attachment insecurity would moderate associations between perceived C&S and diabetes management. That is, individuals higher in attachment insecurity (e.g., higher avoidant attachment; higher anxious attachment) would benefit differently from perceptions of C&S and these differences would be most pronounced when insecurely attached individuals also experienced higher diabetes distress. For individuals higher in anxious attachment, I expected that higher diabetes distress would change the associations between C&S and diabetes management to create a stronger positive association (i.e., indicating better 7 management with higher C&S and distress). For individuals higher in avoidant attachment, I expected that higher diabetes distress would change the associations between higher C&S and diabetes management to create a stronger negative association (i.e., worse management with higher C&S and distress). The second set of aims utilized daily diary measures to understand how differences in attachment extended to withinperson C&S, stress, and daily diabetes management. I examined whether attachment insecurity moderated associations between daily perceptions of C&S and daily diabetes management (Aim 3a) and whether the moderating effect of attachment insecurity was most prominent on days in which individuals experienced more diabetes-related stressors, above and beyond diabetes distress (Aim 3b). I expected daily C&S would associate with better daily diabetes management, but individuals with higher attachment insecurity would benefit less on perceived higher C&S days. However, I expected that on days in which individuals experienced more stressful events, attachment insecurity would strengthen beneficial associations between C&S and diabetes management for individuals higher in anxious attachment, and weaken associations with C&S and become detrimental for daily management for individuals higher in avoidant attachment. METHOD Participant Characteristics Participants were recruited from university-affiliated endocrinology clinics across two sites. Study procedures were approved by the Institutional Review Boards at both sites. All participants provided informed consent. Participants were eligible to participate if patients were 25 years of age or older, had a diagnosis of type 1 diabetes for at least 1 year and were taking insulin within 1 year of diagnosis, spoke English as their primary language1, and were married or in a cohabitating relationship for at least 1 year. Eligible couples (398 individuals, 199 patients) were enrolled and completed study measures. Patients were on average 46.35 (SD = 13.93) years old, 52.3% were women, and were generally well educated. Patients indicated their highest educational attainment as: 12.1% GED or high school diploma; 16.1% some college coursework; 12.1% an associate's or vocational degree; 32.2% a bachelor's degree; and 27.7% a master's degree or higher. Patients were largely non-Hispanic White (89.9%). Patients reported having lived with diabetes for an average of 26.97 (SD = 13.88) years, 68.3% reported using a pump for insulin delivery, and average glycemic control was slightly above current ADA guidelines of an HbA1c level < 7.0% (M = 7.57, SD = 1.06). Participants were mostly married (91.5%) and the average length of relationship was 19.36 years (SD = 14.56). 1 Speaking English as a primary language was necessary for the cognitive testing that was part of the larger study protocol. 9 Procedure The study measures included a self-report online survey completed at home prior to an in-lab assessment, an online survey completed in the lab on the day of assessment, and a 14-day nightly survey completed after the assessment. Additionally, patients' glycemic control was measured during the in-lab assessment and blood glucose was measured across the 14-day nightly survey period, using a study-provided glucometer. Data for the present study utilized patient responses on measures of interest from the online surveys, daily diary, and measures of glycemic control. Survey Measures Attachment insecurity Attachment insecurity was measured using a shortened version of the Experiences in Close Relationships-Revised scale (ECR-R; Fraley et al., 2000). This shortened 26item version was reduced from the original 36 item measure to reduce participant burden in the larger protocol, and assessed attachment insecurity on two subscales: anxiety and avoidance. Participants were instructed to "think about how you generally feel in important relationships" in responding to each statement. The anxiety subscale (α = .89) consisted of 13 items relating to an anxious attachment (e.g., "I worry about being abandoned"). The avoidance subscale (α = .88) consisted of 13 items that related to an avoidant attachment (e.g., "I prefer not to show people how I feel deep down"). Response options were rated on a scale from 1 (disagree strongly) to 7 (agree strongly). 10 Diabetes-related collaboration and support Participants answered questions about their general perceptions of how their partners provided diabetes-related collaboration and support (C&S) as part of a larger measure of diabetes-related support and persuasion (Berg, Schindler, Smith, Skinner, & Beveridge, 2011; Helgeson, Jakubiak, Seltman, Hausmann, & Korytkowski, 2016). A principal components analysis of the full measure revealed six factors that overlapped with the theoretically derived subscales. Two subscales that assessed collaboration and support were included in this study (i.e., collaboration/tangible support; emotional support). The emotional support subscale (α = .80) consisted of five items to assess diabetes-specific emotional support, in which patients were asked to indicate how often their partners engaged in supportive behaviors over the past month (e.g., "listened to me talk about my feelings"). The collaboration/tangible support subscale (α = .94) consisted of seven items to assess diabetes-specific collaboration/tangible support over the past month (e.g., "My partner and I worked together to manage diabetes"). Participant responses were rated on a 5-point scale from 1 (not at all) to 5 (a lot). These two subscales were highly correlated (r = .63, p < .001) and, because there were no a priori hypotheses expecting differences between the types of positive support included in these scales and outcomes, items were collapsed into a single measure of collaboration and support to more closely mirror the daily diary measure (α = .93). Diabetes-related distress Participants completed the Diabetes Distress Scale (Polonsky et al., 2005). The 17-item Diabetes Distress Scale (DDS; Polonsky et al., 2005) assessed the emotional 11 distress individuals living with diabetes experience and has been used in samples of individuals with type 1 diabetes (Polonsky et al., 2005). Items on this measure comprise four empirically-derived subscales (emotional burden, regimen-related distress, physician-related distress, diabetes-related interpersonal distress), which were combined to give an overall indication of diabetes-related distress. Participants rated items on a scale from 1 (not a problem) to 6 (a very serious problem). In our sample, reliability of the total scale was excellent (α = .90). Overall adherence Self-reported adherence was assessed using the widely used Self-Care Inventory (SCI; Lewin et al., 2009). Participants rated how often they regularly engaged in 13 adherence behaviors (e.g., glucose testing, administering correct insulin dose, exercising regularly) as recommended by their healthcare provider in the past month from 1 (did not do) to 5 (always did without fail). Reliability was acceptable (α = .76). Glycemic control Glycemic control was measured using HbA1c, a gold standard measure that estimates how well blood glucose was controlled over the previous ~2 months. Participants gave a capillary blood sample to measure their HbA1c levels using Siemens DCA Vantage Analyzer during the in-lab assessment. HbA1c from the in-lab correlated highly with clinic-visit HbA1c values from medical record (r = .84). 12 Demographics and medical regimen covariates Demographic and medical demographic variables were included as potential covariates a priori. Patients provided information on how they received their insulin to indicate whether they used an insulin pump (referred to in analyses as pump-status), which was recoded to indicate pump use (1) compared to non-pump use (0). Use of an insulin pump has been consistently linked with better glycemic control (e.g., Miller et al., 2015). Patients provided their date of birth during an online survey and age was computed from the date of the lab assessment. Older age has been linked, in this sample, with both less diabetes distress and better adherence (Kelly, Berg, & Helgeson, 2016). Patients reported their gender (0 = male, 1 = female, 2 = other). No participant identified their gender as "other," thus gender was recoded to a dichotomous variable (0 = men; 1 = women). Gender was included to address potential differences in C&S between men and women (e.g., Neff & Karney, 2005). Daily diary measures Daily collaboration and support. Daily perceptions of collaboration and support (daily C&S) were measured using study-created items to examine perceptions of collaboration and support at the daily level. Participants were asked to "describe the ways in which your partner might have been involved in your diabetes today." The initial eight-items of the full scale were comprised of items to capture collaboration (e.g., "Made decisions together with me for diabetes care") and positive social support (e.g., "Tried to understand my situation"). The eight items were subjected to an exploratory factor analysis procedure suitable for diary items in MPlus (Version 8), which revealed 13 that these items loaded onto a single collaboration and support factor. Within-person reliability was λ00 = .96. Daily diabetes stressors. Participants completed a brief checklist (0 = no, 1 = yes) of their experience of 6 common daily diabetes problems (e.g., problem with a high blood glucose, problem with a low blood glucose, taking the wrong amount of insulin). These problems were based on empirically-derived coding of patients' descriptions of diabetes stressors (Beveridge et al., 2005) and represent the most common diabetes problems reported in previous daily diary studies (Berg et al., 2013). A daily total was created by using a count of the number of diabetes problems experienced each day. Daily adherence. Daily adherence was measured using six items from a shortened version of the Self-Care Inventory created for use in daily diaries (Berg et al., 2014). Participants rated how well they followed recommendations from their healthcare provider to engage in six adherence behaviors in the past 24 hours on a 1 (did not do it) to 5 (did it exactly as recommended) scale for various self-management behaviors. Withinperson reliability was λ00 = .97. Daily mean blood glucose. Daily blood glucose (BG) was measured using OneTouch Verio IQ glucometers across participants. Participants were instructed to use this meter as they would their regular glucometer across the 14-day daily diary portion of the study. BG values were uploaded from the meters for each patient. A daily mean was computed across multiple values to estimate average BG level across a 1-day period. 14 Plan of Analysis Missing data was minimal in the cross-sectional portion of the study. Only 1 individual (< 0.01%) did not complete all key measures, thus list-wise deletion was deemed appropriate. All analyses included gender and age as covariates. Analyses with HbA1c as an outcome also included pump-status as a covariate. All nonbinary predictor variables were centered at the grand-mean prior to being included in analyses. Multiple linear regressions were used to address Aims 1 and 2. To address Aim 1, covariates (e.g., gender, age, pump-status) were entered on Step 1 and anxious attachment and avoidant attachment were entered on Step 2 with C&S or diabetes management (i.e., HbA1c; adherence) as outcome variables. To address Aims 2a and 2b, covariates (e.g., gender, age, pump-status) were entered on Step 1, anxious attachment and avoidant attachment, C&S, and 2-way interaction terms (either anxious attachment*C&S or avoidant attachment*C&S) were entered on Step 2, and diabetesrelated distress and all remaining interaction terms to account for a three-way interaction between attachment*C&S*diabetes distress were entered on Step 3. Significant interactions were graphed and simple slopes tested using Dawson's utility for two- and three-way interactions (Dawson, n.d.). Missingness in the diary data also was minimal. Number of completed days ranged from a minimum of 10 to a maximum of 14 days for the total 14-day diary period. On average, participants completed 13.82 days. Missingness for the diary items of interest varied across measures: daily C&S = 2.62% missingness; daily stressors = 1.33% missingness; daily adherence = 0.01%; daily mean BG = 5.81%. For Aim 3, to examine whether attachment insecurity moderated daily associations of C&S, stress, and diabetes 15 outcomes, I utilized a series of multilevel models. All models included both withinperson (daily) and between-person (across 14 days) effects by person-centering daily predictors (Level 1 below) and including centered between-person means across the 14 days as between-person predictors (Level 2 below) to separate out within- and betweenperson effects (Hoffman & Stawksi, 2009). Age and gender were included as covariates in all models. Number of BG checks was person centered and included as a covariate for models with daily mean BG as an outcome. Time was controlled for in all models using grand-mean centered day. First (Aim 3a), two models were utilized to address whether attachment insecurity (i.e., anxious or avoidant) moderated daily C&S, in a cross-level interaction, to predict daily diabetes outcomes (e.g., Daily Adherence or Daily Mean BG) as represented below: Level 1: Daily_Adherenceij = β0j + β1j(Daily_C&Sij) + β2j(Dayij) + eij Level 2: β0j = γ00 + γ01(Attachmentj) + γ02(Average_Daily_C&Sj) + γ03(Agej) + γ04(Genderj) u0j β1j = γ10 + γ11(Attachmentj) β2j = γ20 Random effects were allowed on the intercept of diabetes outcomes (u0j), but grand-mean centered day was expected to remain constant (β2j). Next, (Aim 3b) I examined whether the moderating effect of attachment on perceived C&S is most predictive of daily diabetes outcomes on days in which patients experience more diabetes-related stressful events, above and beyond diabetes distress. First, I created an interaction term between the person-centered daily C&S and personcentered daily diabetes stressors. Grand-mean centered C&S across the two weeks was 16 included in the model. Grand-mean centered diabetes distress from the survey also was included in the model to control for between-person distress in lieu of grand-mean centered diabetes stressors across the two weeks to capture the impact of stressful events above and perceptions of distress. Attachment insecurity (i.e., anxious, avoidant) and age were grand-centered and included in the models. A cross-level three-way interaction between the person-centered C&S*stressors and attachment insecurity (i.e., avoidance or anxiety), along with relevant 2-way interactions, also were included. For models with mean BG as an outcome, person-centered number of BG checks was included. The final model is represented in the sample equation below: Level 1: Daily_Adherenceij = β0j + β1j(Daily_C&Sij) + β2j(Stressorsij) + β3j(Dayij) + β4j (Daily_C&Sij * Stressorsij) + eij Level 2: β0j = γ00 + γ01(Attachmentj) + γ02(Average_C&Sj) + γ03(Diabetes_Distressj) + γ04(Agej) + γ04(Genderj)+ u0j β1j = γ10 + y11(Attachmentj) β2j = γ20 + y21(Attachmentj) β3j = γ30 β4j = γ40 + γ41(Attachmentj) Random effects were allowed on the intercept of diabetes outcomes (u0j), but day was expected to remain constant (β3j). All models were run in SPSS Version 25 using the MIXED command (IBM Corp., 2017). Significant interactions were plotted and simple slopes tested using Preacher and colleagues' Hierarchical Modeling utility (Preacher, Curran, & Bauer, 2006). Statistical significance for all analyses was tested at p ≤ .05. RESULTS Descriptives and Correlations of Key Variables Descriptive statistics and correlations of all key variables are presented in Table 1. Higher anxious attachment associated with greater diabetes distress (r = .41) and more frequent experiences of stressors aggregated across the 2-week diary (r = .24). Higher anxious attachment also associated with lower adherence (r = -.31), but this was limited to the survey measure. Higher avoidant attachment only associated with higher diabetes distress (r = .24) and did not associate with any aspect of diabetes management. Attachment Insecurity Predicting Collaboration and Support or Diabetes Management Multiple regressions were used to understand whether attachment insecurity associated with C&S and diabetes outcomes (Table 2). Anxious attachment associated with perceived C&S and adherence, such that higher anxious attachment associated with lower perceived C&S and lower adherence. However, there were no significant main effects of anxious attachment on HbA1c. Moreover, avoidant attachment, when controlling for anxious attachment, did not emerge as a significant predictor of C&S or diabetes management. 18 Between-Persons Associations of C&S and the Moderating Roles of Attachment Insecurity and Diabetes Distress on Diabetes Management To understand how attachment insecurity moderated associations between C&S and diabetes management, and whether diabetes distress moderated associations between attachment insecurity, C&S, and diabetes management, a set of multiple regressions was conducted (Table 3). Attachment insecurity, C&S, and attachment*C&S explained 9% of the variance in adherence. A significant main effect only emerged for anxious attachment with adherence, such that higher anxious attachment associated with lower self-reported adherence. Diabetes distress and interaction terms explained an additional 8% of the variance in adherence for the model with anxious attachment and 12% of the variance for the model with avoidant attachment. Significant main effects emerged for diabetes distress with adherence, such that higher diabetes distress associated with lower adherence. No significant interactions emerged for the model with anxious attachment. A significant three-way interaction emerged between avoidant attachment*diabetes distress*C&S (see Figure 1). Simple slopes tests showed higher support was only significantly associated with lower adherence for individuals high in avoidance and low in distress (t = -2.08, p = .04). Attachment insecurity, C&S, and attachment*C&S explained only 1% of the variance in HbA1c. Diabetes distress and interaction terms explained an additional 8% of the variance in HbA1c for the model with anxious attachment and 5% of the variance for the model with avoidant attachment. Significant main effects emerged for diabetes distress with HbA1c, such that higher diabetes distress associated with higher (or worse) 19 HbA1c. A significant two-way interaction emerged between anxious attachment and diabetes distress for HbA1c (see Figure 2). Simple slopes tests indicated that for individuals with higher diabetes distress (1 SD above the mean), higher anxious attachment associated with lower HbA1c (t = -2.39, p = .02). For individuals with lower diabetes distress (1 SD below the mean), higher anxious attachment associated with higher HbA1c (t = 2.09, p = .04). Within-Person Associations of Collaboration and Support and the Moderating Roles of Attachment Insecurity and Stress on Daily Diabetes Management Multilevel models examined whether attachment insecurity moderated daily associations of C&S and diabetes management (see Table 4). For daily adherence, no main effects emerged for within-person or between-person C&S, anxious attachment, avoidant attachment, or interactions between daily C&S and attachment insecurity. For daily mean BG, only within-person C&S emerged as a significant predictor, such that on days in which participants reported higher support, they also experienced higher mean BG. A significant main effect also emerged for day, suggesting that BG was higher across the 14 days of the diary. The effect of day on mean BG was consistent but small (est. = .48-.49). A significant cross-level interaction was found for within-person C&S and anxious attachment - but not avoidant attachment - on mean BG (see Figure 3). Testing of slopes indicated that at low levels of anxious attachment, the slope of the line was not different from zero (p = .40). For individuals higher in anxious attachment, 20 higher daily support associated with higher mean BG (simple slope = 4.60, p = .01). Multilevel models then examined whether the addition of distress and stressful events further changed associations between attachment insecurity, daily C&S, and daily diabetes management (Table 5). For models with adherence as an outcome, anxious attachment, within-person C&S, within-person stressors, and between-person diabetes distress significantly associated with reports of daily adherence. That is, on days in which participants reported higher C&S, adherence was higher. On days in which participants reported experiencing more diabetes-related stressors and those who experienced higher overall diabetes distress also reported lower adherence. Attachment insecurity did not emerge as a significant predictor of daily adherence and neither anxious attachment nor avoidant attachment moderated associations between within-person C&S*stressors and adherence. When taken with the results presented above, where support did not significantly associate with adherence, controlling for diabetes-related stress is important for determining whether or not C&S is beneficial for adherence. That is, as higher perceived C&S may be in response to experiencing more diabetes problems, it is only beneficial for adherence when controlling for the number of diabetes-related stressful events. For daily mean BG as the outcome, within-person stressors and between-person diabetes distress, but not within-person C&S, associated with increased daily mean BG. That is, on days in which participants reported more diabetes stressors and for participants who reported higher overall diabetes distress, mean BG was higher. Attachment insecurity did not emerge as a significant predictor of daily mean BG. A statistically significant association emerged for anxious attachment * daily C&S (p = 21 .05), but simple slopes for low (p =.33) and high (p = .10) anxious attachment were not statistically different from zero, thus findings were not interpreted. A significant association also emerged for anxious attachment * daily stressors (see Figure 4). For both individuals with lower and higher anxious attachment, experiencing more stressors associated with higher mean BG.2 However, the slope of the line was steeper for individuals lower in anxious attachment (simple slope = 11.82, p < .001) than for individuals higher in anxious attachment (simple slope = 8.47, p < .001). These results suggest that individuals lower in anxious attachment may be more impacted by the experience of daily stressors compared to individuals higher in anxious attachment. No significant two-way interaction emerged for avoidant attachment on mean BG. Neither anxious attachment nor avoidant attachment moderated associations between withinperson C&S*stressors and mean BG. 2 The stressor checklist included two items to capture problems with low or high blood glucose. Though participants did not report having a problem with a high blood glucose or low blood glucose every day, to be prudent this analysis was re-run without these two items. Using the 4-item composite, this interaction failed to reach significance (p = .13). Though at the aggregate level there is not much overlap with daily stressors and mean BG (r = .23), these two items appear to be driving the daily associations. Table 1. Correlations and descriptives of key study variables 1 2 3 4 5 6 7 8 9 10 11 M (SD) 1. Anxious Attachment --- --- --- --- --- --- --- --- --- --- --- 2.82 (1.17) 2. Avoidant Attachment .28 --- --- --- --- --- --- --- --- --- --- 3.14 (1.08) 3. Collaboration & Support -.27 -.08 --- --- --- --- --- --- --- --- --- 3.01 (.93) 4. Diabetes Distress Scale .41 .24 -.07 --- --- --- --- --- --- --- --- 1.98 (.74) 5. Gender .08 -.27 -.08 .11 --- --- --- --- --- --- --- .52 (.50) 6. Age -.14+ .08 -.05 -.21 -.08 --- --- --- --- --- --- 46.35 (13.93) 7. Adherence (Survey) -.31 -.073 .04 -.39 .14 .25 --- --- --- --- --- 3.67 (.56) 8. HbA1c .03 -.00 .04 .20 -.00 .12 -.32 --- --- --- --- 7.57 (1.06) 9. Average Diary C&S -.20 -.14+ .70 -.06 -.12 -.04 .06 -.03 --- --- --- 2.14 (.92) 10. Diary Stressors .24 .04 -.02 .43 .15 -.28 -.32 .18 .05 --- --- 1.71 (.94) 11. Diary Adherence -.10 -.08 -.05 -.33 .22 .28 .58 -.33 -.06 -.43 --- 4.38 (.59) 12. Diary Mean BG .05 .07 .02 .17 -.03 -.03 -.34 .59 .05 .23 -.36 175.52 (35.30) Note. Diary items were averaged across the 2-week diary. Gender coded 0 (men) and 1 (women). Bold indicates significant at p ≤ .05. + p = .06. 22 Table 2. Associations of attachment on C&S and diabetes management Collaboration & Support Adherence Βin βin 𝑅2∆ 𝑅2∆ Step 1 .01 .09 Intercept 3.09 3.57 Gender -.15 .19 Age -.00 .01 1 Pump-Status ----Step 2 .08 .08 Anxious Attachment -.22 -.15 Avoidant Attachment -.01 .03 1 Note. Insulin pump-status only entered into model with HbA1c. Bold is significant at p ≤ .05. HbA1c 𝑅2∆ .03 Βin 7.69 .07 .01 -.23 .01 .07 -.03 23 Table 3. Diabetes distress moderating the interaction between attachment and C&S and diabetes management Adherence Anxious Attachment Step 1 Intercept Gender Age Pump-Status1 Step 2 Anxious Attachment Avoidant Attachment C&S Anxious Attachment*C&S Step 3 Diabetes Distress (DDS) C&S*DDS Anxious Attachment*DDS Anxious Attachment*DDS*C&S Avoidant Attachment Step 1 Intercept Gender Age Pump-Status1 Step 2 Anxious Attachment Avoidant Attachment C&S Avoidant Attachment*C&S Step 3 Diabetes Distress (DDS) C&S*DDS Avoidant Attachment*DDS Avoidant Attachment*DDS*C&S 𝑅2 ∆ .08 HbA1c Βin 𝑅2 ∆ .03 3.57 .19 .01 --.09 7.70 .01 .01 -.22 .01 -.15 .02 -.01 .05 .08 𝑅2 ∆ .08 Βin .09 -.02 .05 .04 .08 -.26 .03 -.01 -.05 Βin 𝑅2 ∆ .03 3.57 .19 .01 --.09 .39 -.09 -.24 .04 Βin 7.70 .01 .01 -.22 .01 -.15 .02 -.01 .01 .12 .09 -.02 .05 01 .05 -.22 .04 -.05 .16 .35 .05 -.01 -.03 24 Note. 1Insulin pump-status only entered into models with HbA1c. Gender coded 0 (men) and 1 (women). Bold is significant at p ≤ .05. Table 4. Attachment moderating associations between daily C&S and diabetes management Adherence Mean Blood Glucose Anxious Attachment Estimate SE 95% CI Estimate SE 95% CI Intercept 4.24 .06 [4.13, 4.35] 175.94 3.76 [168.52, 183.36] Day .00 .00 [-.00, .00] .49 .24 [.02, .95] 1 BG Checks (WP) ------.01 .72 [-1.41, 1.43] C&S (WP) .01 .01 [-.01, .04] 3.01 1.51 [.05, 5.98] C&S (BP) -.02 .04 [-.11, .06] 2.86 2.91 [-2.56, 8.39] Anxious Attachment -.04 .04 [-.11, .03] 1.42 2.29 [-3.11, 5.94] Avoidant Attachment -.00 .04 [-.08, .07] 1.96 2.57 [-3.10, 7.02] Age .01 .00 [.01, .02] -.05 .18 [-.41, .31] Gender .28 .08 [.12, .44] -1.07 5.32 [-11.56, 9.41] WP C&S*Anxious Attachment .00 .01 [-.02, .02] 2.51 1.21 [.15, 4.88] Avoidant Attachment Estimate SE 95% CI Estimate SE 95% CI Intercept 4.24 .06 [4.13, 4.35] 175.93 3.78 [168.51, 183.35] Day .00 .00 [-.00, .00] .48 .24 [.01, .95] BG Checks (WP)1 -------.04 .72 [-1.46, 1.39] C&S (WP) .01 .01 [-.01, .04] 3.01 1.51 [.04, 5.98] C&S (BP) -.02 .04 [-.11, .06] 2.92 2.77 [-2.55, 8.40] Anxious Attachment -.04 .04 [-.11, .03] 1.42 2.29 [-3.09, 5.94] Avoidant Attachment -.00 .04 [-.08, .07] 1.96 2.57 [-3.10, 7.03] Age .01 .01 [.01, .02] -.05 .18 [-.41, .31] Gender .28 .08 [.12, .44] -1.05 5.31 [-11.54, 9.43] WP C&S*Avoidant Attachment .02 .01 [-.00, .04] -.64 1.29 [-3.16, 1.88] 1 Note. Number of person-centered BG checks only entered into models with mean BG. Gender coded as men (0) and women (1). Bold is significant at p ≤ .05. 25 Table 5. Attachment moderating associations between C&S*stressors and diabetes management Anxious Attachment Intercept Day BG Checks (WP)1 C&S (WP) Diabetes Stressors (WP) C&S*Stressors (WP) C&S (BP) Diabetes Distress (BP) Anxious Attachment Avoidant Attachment Age Gender C&S (WP)*Anxious Attachment Stressors (WP)*Anxious Attachment C&S*Stressors (WP)*Anxious Attachment Avoidant Attachment Intercept Day BG Checks1 (WP) C&S (WP) Diabetes Stressors (WP) C&S*Stressors (WP) Diabetes Distress (BP) Anxious Attachment Avoidant Attachment Age Gender C&S (WP) * Avoidant Attachment Stressors (WP) * Avoidant Attachment C&S*Stressors (WP) *Avoidant Attachment Estimate 4.21 -.00 --.04 -.11 -.01 -.01 -.26 .01 .03 .01 .33 .01 .00 .01 Estimate 4.22 -.00 --.04 -.11 -.01 -.26 .01 .03 .01 .33 .02 -.01 .01 Adherence SE 95% CI .05 [4.11, 4.32] .00 [-.00, .00] ----.01 [.02, .06] .01 [-.12, -.10] .01 [-.03, .01] .04 [-.09, .07] .06 [-.37, -.15] .04 [-.06, .08] .04 [-.04, .11] .00 [.00, .02] .08 [.18, .48] .01 [-.01, .02] .01 [-.01, .01] .01 [-.01, 02] SE 95% CI .05 [4.11, 4.32] .00 [-.00, .00] ----.01 [.02, .06] .01 [-.12, -.10] .01 [-.03, .01] .06 [-.37, -.15] .04 [-.06, .08] .04 [-.04, .11] .00 [.00, .02] .08 [.18, .48] .01 [-.00, .04] .01 [-.02, .00] .01 [-.01, .02] Estimate 176.71 .59 -.43 .69 10.14 .53 2.65 8.39 -.60 .72 .03 -2.90 2.32 -1.43 1.76 Estimate 176.67 .57 -.48 .63 9.89 .79 8.46 -.44 .78 .03 -2.82 .02 -.65 -.88 Mean Blood Glucose SE 95% CI 3.76 [169.30, 184.13] .23 [.14, 1.05] .70 [-1.81, .95] 1.51 [-2.26, 3.65] .82 [8.52, 11.77] 1.17 [-1.75, 2.83] 2.76 [-2.80, 8.10] 3.78 [.93, 15.86] 2.42 [-5.38, 4.18] 2.61 [-4.42, 5.87] .19 [-.33, .40] 5.35 [-13.48, 7.65] 1.19 [-.01, 4.65] .73 [-2.86, -.00+] 1.01 [-.21, 3.73] SE 95% CI 3.75 [169.26, 184.07] .23 [.11, 1.03] .71 [-1.87, .90] 1.51 [-2.33, 3.60] .82 [8.28, 11.50] 1.17 [-1.50, 3.08] 3.78 [1.01, 15.90] 2.42 [-5.20, 4.33] 2.60 [-4.36, 5.91] .18 [-.33, .40] 5.34 [-13.36, 7.71] 1.29 [-2.55, 2.50] .78 [-2.18, .87] 1.00 [-2.85, 1.08] 26 Note. 1Number of person-centered BG checks only entered into models with mean BG. Gender coded 0 (men) and 1 (women). Bold = p ≤ .05. +Upper Bound 95% CI = -.003 . 27 5 4.5 Adherence 4 3.5 (1) High Avoidant Attachment, High Diabetes Distress (2) High Avoidant Attachment, Low Diabetes Distress (3) Low Avoidant Attachment, High Diabetes Distress (4) Low Avoidant Attachment, Low Diabetes Distress 3 2.5 2 1.5 1 Low C&S High C&S Figure 1. Three-way interaction between avoidant attachment, diabetes distress, with HbA1c 28 Figure 2. Two-way interaction between anxious attachment and diabetes distress with HbA1c 29 Figure 3. Two-way interaction within-person C&S and anxious attachment on mean blood glucose 30 Figure 4. Two-way interaction within-person stressors by anxious attachment on mean blood glucose. DISCUSSION The Roles of Attachment Insecurity and Diabetes Distress Between Persons The results indicated that anxious attachment associated with lower perceived C&S and worse self-reported, but not objective, diabetes management. Higher anxious attachment consistently associated with worse adherence in both correlations and regression analyses. These results align with research that suggests higher anxious attachment correlates with higher reports of physical health symptoms (Wearden, Cook, & Vaughan-Jones, 2003) and more negative self-views (Mikulincer, 1995) in young, healthy samples. Given that this finding does not appear with HbA1c, similar to other negative self-evaluations, individuals higher in anxious attachment may evaluate their self-care efforts more negatively. Avoidant attachment itself did not significantly associate with C&S or diabetes management, though its association with C&S trended in the expected direction (i.e., lower perceived C&S). Attachment insecurity did not moderate associations between C&S and diabetes management. Moreover, C&S itself did not associate cross-sectionally with better diabetes management. This is in contrast to research in type 2 diabetes, which suggests that positive aspects of support associate with better diabetes management (e.g., Mayberry & Osborn, 2014). A recent meta-analysis of social support in type 1 diabetes 32 suggests there are differences in how much individuals benefit from social support compared to individuals with type 2 diabetes. That is, although social support was beneficial for those with type 1 diabetes, compared to type 2 diabetes, studies of individuals with type 1 diabetes had smaller effect sizes (Song, Nam, Park, Shin, & Ku, 2017). It may be for adults with type 1 diabetes expectations of partner involvement are more nuanced than our measure of C&S captured. One qualitative study found individual differences in how much support patients expected and wanted from their partners (Trief et al., 2013). Our measure of C&S only assessed perceptions of what partners provided, but not how much C&S patients desired. Further, although attachment insecurity captures differences in how individuals perceive support (e.g., Collins & Feeney, 2004), it may not adequately capture how much individuals desire support. Contrary to the hypothesized associations, diabetes distress did not amplify associations between C&S and attachment with diabetes management. However, higher diabetes distress was consistently linked with reports of lower adherence and worse HbA1c. Though the addition of diabetes distress did produce two significant interactions, the results were not as expected. Instead of diabetes distress making individuals with anxious attachment more prone to worse outcomes for HbA1c, the results suggest that distress matters less for individuals higher in anxious attachment (see Figure 2). To further understand this interaction, a plot of anxious attachment as the moderator of diabetes distress also was explored (see Figure 5). When plotted this way, the slope of the line only indicated worse glycemic control for individuals lower in anxious attachment. It may be that, as individuals higher in anxious attachment are always in a state of higher activation (e.g., Simpson & Rholes, 2012), distress is only derailing for individuals lower 33 in anxious attachment. The significant three-way interaction that emerged between avoidant attachment, diabetes distress, and C&S (Figure 1) suggests it is only during times of low distress that individuals higher in avoidant attachment are negatively affected by increased C&S. It may be that individuals higher in avoidant attachment, when they are not distressed, perceive C&S as intrusive in their typical independent diabetes management. Yet, when experiencing high distress, they understand a need for greater C&S, even if they do not expect or significantly benefit from it. This interpretation is consistent with the independent orientation of individuals with more avoidant attachment and the high desire for control and autonomy (e.g., Simpson & Rholes, 2012). Consistent with this idea, other work in this sample suggests that the benefits of C&S may partially lie in how patients "appraise" their illness. Individuals who report a more individual illness appraisal (compared to shared) do not benefit from C&S (Berg et al., 2018; Helgeson et al., 2018). Moreover, as our sample on average reported low diabetes distress (Fisher, Hessler, Polonsky, & Mullan, 2012), it also is possible that individuals did not reach the higher levels of distress needed to activate the attachment system. The Roles of Attachment Insecurity and Diabetes Distress Within Persons The results of the within-person analyses failed to support a distinct, moderating role of attachment insecurity between daily C&S and diabetes management. One interaction (see Figure 3) suggested that individuals higher in anxious attachment showed a steeper increase in mean BG values on days in which they perceived higher C&S. 34 Though the direction of causality cannot be determined, it is unlikely - given other studies of support (e.g., Song et al., 2017) - that C&S itself is worse for daily mean BG for individuals higher in anxious attachment. Instead, it is plausible that more anxiously attached individuals may more effectively pull in their partners (i.e., elicit and perceive higher support) when they are experiencing problems with diabetes management because of their tendency to "maximize" their cues to gain their attachment figure's support (Pietromonaco et al., 2013). The other interaction (see Figure 4) suggested the experience of more daily stressors associated with higher daily mean BG for individuals both lower and higher in anxious attachment. However, the association appeared stronger for individuals who were lower in anxious attachment. This finding was fairly consistent with the cross-sectional results, which showed that individuals lower in anxious attachment experienced worse glycemic control when under higher distress. It may be that, while individuals higher in anxious attachment are affected by experiencing more diabetes-related stressful events, their state of ‘higher activation' buffers the impact of stressors on mean BG (e.g., Simpson & Rholes, 2012). Limitations, Conclusions, and Future Directions Though the current study had several strengths, including examining both between- and within-person associations, it was not without limitations. First, the sample was largely of higher socioeconomic status, non-Hispanic White, and experiencing good self-management. Results may not generalize to individuals struggling with their diabetes care, whether because of lack of access, financial concerns about cost of care, or other issues that are not as common in this sample. Second, the presented analyses were limited 35 to patients' attachment, perceptions of diabetes-related C&S, and diabetes management. Individuals in romantic partnerships and their social interactions are dyadic (e.g., Reis et al., 2000). As the outcomes of interest for this study were at the level of the patient and there are few studies that have applied attachment theory to diabetes management in the context of romantic relationships, the presented results were an important first step. Future studies should directly examine the dyadic nature of attachment in the context of diabetes management. Finally, the lack of findings for attachment may be due, in part, to asking participants to think broadly of "important relationships in general." The reason for this wording choice was to capture a more developmental definition of attachment rather than attachment to one person. Yet, individuals have different attachments to different individuals (Collins & Feeney, 2000). A romantic-relationship specific wording assessing attachment to participants' current partner may yield different results. Further, attachment was only measured cross-sectionally, and attachment insecurity may fluctuate within-person (La Guardia, Ryan, Couchman, & Deci, 2000). It is possible that a daily measure of attachment would be more closely related to daily perceptions of C&S and diabetes management than a one-time assessment. Despite the theoretical justification for exploring attachment in the processes of collaboration and support, it does not appear that attachment insecurity moderates associations between C&S and diabetes management. Though it is expected that attachment is a fundamental developmental process that should affect perceptions of relationships and health (e.g., Pietromonaco et al., 2013), these processes may be too far removed from the management of diabetes on a day-to-day basis. Interestingly, unlike findings from research in type 2 diabetes (e.g., Khan et al., 2013) and type 1 diabetes 36 (Song et al., 2017), C&S itself was not consistently associated with better diabetes management. It was only when controlling for individuals' stressors and experiences of distress that a small positive main effect of daily C&S was found for adherence. Other work in this sample indicates for C&S to be positive, individual differences in how diabetes is appraised may be especially important (Berg et al., 2018; Helgeson et al., 2018). Future studies should continue to work toward disentangling issues surrounding attachment and chronic illness management. These findings alone do not to suggest that attachment is not important for diabetes management. The experience of diabetes distress and stressors consistently associated with diabetes management both between and within persons. Given that attachment insecurity (both anxious and avoidant attachment) associated with higher diabetes distress, it may be that attachment is most important for more psychosocial outcomes of diabetes rather than self-management. Moreover, findings from this study highlight the need for expanding psychosocial research in type 1 diabetes across adulthood to find more definitive answers as to when and why collaboration and support are beneficial to adults with type 1 diabetes. 37 9 HbA1c 8.5 Low Anxious Attachment 8 7.5 High Anxious Attachment 7 6.5 6 5.5 5 Low Diabetes Distress High Diabetes Distress Figure 5. Inverse plot of anxious attachment moderating diabetes distress with HbA1c REFERENCES August, K. J., & Sorkin, D. H. (2010). Marital status and gender differences in managing a chronic illness: The function of health-related social control. Social Science & Medicine, 71(10), 1831-1838. doi:10.1016/j.socscimed.2010.08.022 Berg, C. A., Helgeson, V. S., Kelly, C. S., Lee, E., van Vleet, M, Butner, J., & Litchman, M.L. (2018, March). Benefit of support for daily diabetes management depends on shared illness appraisal in adults with type 1 diabetes. Paper presented at the 39th Annual Meeting & Scientific Sessions of the Society of Behavioral Medicine, Orlando, FL. Berg, C. A., Wiebe, D. J., Suchy, Y., Hughes, A. E., Anderson, J. H., Godbey, E. I., … White, P. C. (2014). Individual differences and day-to-day fluctuations in perceived self-regulation associated with daily adherence in late adolescents with type 1 diabetes. Journal of Pediatric Psychology, 39(9), 1038-1048. doi:10.1093/jpepsy/jsu051 Berg, C. A., Butner, J. E., Butler, J. M., King, P. S., Hughes, A. E., & Wiebe, D. J. (2013). Parental persuasive strategies in the face of daily problems in adolescent type 1 diabetes management. Health Psychology, 32(7), 719-728. doi:10.1037/a0029427 Berg, C. A., Schindler, I., Smith, T. W., Skinner, M., & Beveridge, R. M. (2011). Perceptions of the cognitive compensation and interpersonal enjoyment functions of collaboration among middle-aged and older married couples. Psychology and Aging, 26(1), 167-173. doi:10.1037/a0021124 Beveridge, R. M., Berg, C. A, Wiebe, D .J. , & Palmer, D. L. (2005). Mother and adolescent representations of illness ownership and stressful events surrounding diabetes. Journal of Pediatric Psychology, 31(8), 818-827. doi:10.1093/jpepsy/jsj094 Campbell, L., Simpson, J. A., Boldry, J., & Kashy, D. A. (2005). Perceptions of conflict and support in romantic relationships: The role of attachment anxiety. Journal of Personality and Social Psychology, 88(3), 510-531. doi:10.1037/0022-3514.88.3.510 Chiang, J. L., Kirkman, M. S., Laffel, L. M. B., & Peters, A. L. (2014). Type 1 diabetes 39 through the life span: A position statement of the American Diabetes Association. Diabetes Care, 37(7), 2034-2054. doi:10.2337/dc14-1140 Ciechanowski, P., Russo, J., Katon, W., Von Korff, M., Ludman, E., Lin, E., ... Bush T. (2004). Influence of patient attachment style on self-care and outcomes in diabetes. Psychosomatic Medicine, 66(5), 720-728. Ciechanowski, P. S., Hirsch, I. B., & Katon, W. J. (2002). Interpersonal predictors of HbA1c in patients with type 1 diabetes. Diabetes care, 25(4), 731-736. Cohen, O., Birnbaum, G. E., Meyuchas, R., Levinger, Z., Florian, V., & Mikulincer, M. (2005). Attachment orientations and spouse support in adults with type 2 diabetes. Psychology, Health & Medicine, 10(2), 161-165. doi: 10.1080/1354850042000326575 Collins, N. L., & Feeney, B. C. (2000). A safe haven: An attachment theory perspective on support seeking and caregiving in intimate relationships. Journal of Personality and Social Psychology, 78(6), 1053. doi: 10.1O37//OO22-3514.78.6.1053 Collins, N. L., & Feeney, B. C. (2004). Working models of attachment shape perceptions of social support: evidence from experimental and observational studies. Journal of Personality and Social Psychology, 87(3), 363. doi: 10.1037/0022-3514.87.3.363 Dawson, J. F. Interpreting interaction effects. Retrieved from http://www.jeremydawson.co.uk/slopes.htm Fisher, L., Hessler, D. M., Polonsky, W. H., & Mullan, J. (2012). When is diabetes distress clinically meaningful?: Establishing cut points for the Diabetes Distress Scale. Diabetes Care, 35(2), 259-264. doi:10.2337/dc11-1572 Fraley, R. C., Waller, N. G., & Brennan, K. A. (2000). An item response theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78(2), 350-365. doi:10.1037/0022-3514.78.2.350 Gallant, M. P., Spitze, G. D., & Prohaska, T. R. (2007). Help or hindrance? How family and friends influence chronic illness self-management among older adults. Research on Aging, 29(5), 375-409. doi:10.1177/0164027507303169 Helgeson, V. S., Jakubiak, B., Seltman, H., Hausmann, L., & Korytkowski, M. (2016). Implicit and explicit communal coping in couples with recently diagnosed type 2 diabetes. Journal of Social and Personal Relationships, 34(7), 1099-1121. doi:10.1177/0265407516669604 Helgeson, V. S., Van Vleet, M., Zajdel, M., Lee, E., Litchman, M., Kortykowski, M., & 40 Berg, C. A. (2018, April). Communal coping: An examination of shared illness appraisal and collaboration in the context of adults with type 1 diabetes. Paper presented at Society for Behavioral Medicine Meetings, New Orleans, LA. Hoffman, L., & Stawski, R. S. (2009). Persons as contexts: Evaluating between-person and within-person effects in longitudinal analysis. Research in Human Development, 6(2-3), 97-120. doi:10.1080/15427600902911189 IBM Corp. (2017). IBM SPSS statistics for Windows, Version 25.0. Armonk, NY: IBM Corp. Iida, M., Parris Stephens, M. A., Rook, K. S., Franks, M. M., & Salem, J. K. (2010). When the going gets tough, does support get going? Determinants of spousal support provision to type 2 diabetic patients. Personality and Social Psychology Bulletin, 36(6), 780-791. doi:10.1177/0146167210369897 Kelly, C. S., Berg, C. A., & Helgeson, V. S. (2016, April). Diabetes distress in type 1 diabetes: A lifespan approach. Poster presented at the 37th Annual Meeting & Scientific Sessions of the Society of Behavioral Medicine. Washington, D.C. Khan, C. M., Stephens, M. A. P., Franks, M. M., Rook, K. S., & Salem, J. K. (2013). Influences of spousal support and control on diabetes management through physical activity. Health Psychology, 32(7), 739-747. doi:10.1037/a0028609 La Guardia, J. G., Ryan, R. M., Couchman, C. E., & Deci, E. L. (2000). Within-person variation in security of attachment: A self-determination theory perspective on attachment, need fulfillment, and well-being. Journal of Personality and Social Psychology, 79(3), 367-384. doi:10.1037/0022-3514.79.3.367 Lewin, A. B., LaGreca, A. M., Geffken, G. R., Williams, L. B., Duke, D. C., Storch, E. A., & Silverstein, J. H. (2009). Validity and reliability of an adolescent and parent rating scale of type 1 diabetes adherence behaviors: The Self-Care Inventory (SCI). Journal of Pediatric Psychology, 34(9), 999-1007. doi:10.1093/jpepsy/jsp032 Mayberry, L. S., & Osborn, C. Y. (2014). Family involvement is helpful and harmful to patients' self-care and glycemic control. Patient Education and Counseling, 97(3), 418-425. doi:10.1016/j.pec.2014.09.011 Miller, K. M., Foster, N. C., Beck, R. W., Bergenstal, R. M., DuBose, S. N., DiMeglio, L. A., … Tamborlane, W. V. (2015). Current state of type 1 diabetes treatment in the U.S.: Updated data From the T1D Exchange Clinic Registry. Diabetes Care, 38(6), 971-978. doi:10.2337/dc15-0078 Mikulincer, M. (1995). Attachment style and the mental representation of the self. Journal of Personality and Social Psychology, 69(6), 1203-1215. 41 doi:10.1037/0022-3514.69.6.1203 Neff, L. A., & Karney, B. R. (2005). Gender differences in social support: A question of skill or responsiveness? Journal of Personality and Social Psychology, 88(1), 79-90. doi:10.1037/0022-3514.88.1.79 Ognibene, T. C., & Collins, N. L. (1998). Adult attachment styles, perceived social support and coping strategies. Journal of Social and Personal Relationships, 15(3), 323-345. doi:10.1177/0265407598153002 Pietromonaco, P. R., & Barrett, L. F. (1997). Working models of attachment and daily social interactions. Journal of Personality and Social Psychology, 73(6), 1409- 1423. doi:10.1037/0022-3514.73.6.1409 Pietromonaco, P. R., Uchino, B., & Dunkel Schetter, C. (2013). Close relationship processes and health: Implications of attachment theory for health and disease. Health Psychology, 32(5), 499-513. doi: 10.1037/a0029349 Polonsky, W. H., Fisher, L., Earles, J., Dudl, R. J., Lees, J., Mullan, J., & Jackson, R. A. (2005). Assessing psychosocial distress in diabetes. Diabetes Care, 28(3), 626631. Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31(4), 437-448. doi:10.3102/10769986031004437 Reis, H. T., Collins, W. A., & Berscheid, E. (2000). The relationship context of human behavior and development. Psychological Bulletin, 126(6), 844-872. doi: I0.1037//0033-2909.126.6.844 Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving within couples in an anxiety-provoking situation: The role of attachment styles. Journal of Personality and Social Psychology, 62(3), 434-446. Song, Y., Nam, S., Park, S., Shin, I., & Ku, B. J. (2017). The impact of social support on self-care of patients with diabetes: What is the effect of diabetes type? The Diabetes Educator, 43(4), 396-412. doi:10.1177/0145721717712457 Stephens, M. A. P., Franks, M. M., Rook, K. S., Iida, M., Hemphill, R. C., & Salem, J. K. (2013). Spouses' attempts to regulate day-to-day dietary adherence among patients with type 2 diabetes. Health Psychology, 32(10), 1029-1037. doi:10.1037/a0030018 Trief, P. M., Sandberg, J. G., Dimmock, J. A., Forken, P. J., & Weinstock, R. S. (2013). Personal and relationship challenges of adults with type 1 diabetes. Diabetes 42 Care, 36(9), 2483-2488. doi:10.2337/dc12-1718 Van Bastelaar, K. M. P., Pouwer, F., Geelhoed-Duijvestijn, P. H. L. M., Tack, C. J., Bazelmans, E., Beekman, A. T., … Snoek, F. J. (2010). Diabetes-specific emotional distress mediates the association between depressive symptoms and glycaemic control in type 1 and type 2 diabetes. Diabetic Medicine, 27(7), 798- 803. doi:10.1111/j.1464-5491.2010.03025.x Vilchinsky, N., Haze-Filderman, L., Leibowitz, M., Reges, O., Khaskia, A., & Mosseri, M. (2010). Spousal support and cardiac patients' distress: The moderating role of attachment orientation. Journal of Family Psychology, 24(4), 508-512. doi: 10.1037/a0020009 Vogel, D. L., & Wei, M. (2005). Adult attachment and help-seeking intent: The mediating roles of psychological distress and perceived social support. Journal of Counseling Psychology, 52(3), 347-357. doi: 10.1037/0022-0167.52.3.34 Wiebe, D. J., Helgeson, V., & Berg, C. A. (2016). The social context of managing diabetes across the life span. American Psychologist, 71(7), 526-538. doi:10.1037/a0040355 Wearden, A., Cook, L., & Vaughan-Jones, J. (2003). Adult attachment, alexithymia, symptom reporting, and health-related coping. Journal of Psychosomatic Research, 55(4), 341-347. doi:10.1016/s0022-3999(02)006