A Sleep/Wake Protocol to Reduce Delirium in the CVICU and SICU

Critically ill and hemodynamically unstable patients are at risk for experiencing a decrease in quality and quantity of sleep. Sleep deprivation in critically ill patients has been associated with an increased incidence of ICU delirium. To decrease the incidence of delirium and improve perceived patient sleep quality through optimization of environmental factors (noise, light, nursing activities) as well as non-environmental factors (pain, mechanical support devices, procedures) in a Cardiovascular and Surgical ICU. This QI project was implemented at a large academic facility in the intermountain west within two intensive care units (N = 596). The SICU is a 12-bed unit and the CVICU is a 20-bed unit comprised of adult post-operative, trauma, and critically ill patients. Using a pretest-posttest study design, baseline data was collected prior to implementation of the Sleep/Wake protocol after which post-implementation data was collected. The Sleep/Wake protocol is comprised of environmental interventions to complete during day and night shift. Primary outcome of delirium was measured using the confusion assessment method-ICU. Secondary outcome of perceived sleep included a baseline sleep survey of sleep quality prior to hospitalization plus the Richards-Campbell Sleep Questionnaire. Nurse surveys were used to evaluate usability and satisfaction of the protocol. During the pre- and post-intervention there were 272 and 229 patients respectively who were screened for delirium and who participated in the RCSQ. Incidence of delirium decreased from 82/272 (30%) pre- vs 55/229 (24%) (p-value 0.13) post-intervention. Compliance with the Sleep/Protocol ranged from 20-98%, respectively. Perceived sleep quality showed similar mean (SD) between both groups with the greatest difference made in noise level (before 65(22), after 68(20), p=0.16). Satisfaction of the protocol was evaluated via nurse survey. Overall, 88% of those surveyed were satisfied with the protocol. The implementation of a Sleep/Wake Protocol is feasible within a surgical and cardiovascular ICU and showed a decrease in incidence of delirium pre- to post-intervention, but was not statistically different. There was neither improved nor reduced patient sleep quality, however staff were overall satisfied with the protocol.

Running head: IMPLEMENTATION OF SLEEP/WAKE PROTOCOL A Sleep/Wake Protocol to Reduce Delirium in the CVICU and SICU Anna T. Dalton Project Chair: Nancy A. Allen Content Expert: Joseph E. Tonna The University of Utah College of Nursing 1 IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 2 Abstract Background: Critically ill and hemodynamically unstable patients are at risk for experiencing a decrease in quality and quantity of sleep. Sleep deprivation in critically ill patients has been associated with an increased incidence of ICU delirium. Purpose: To decrease the incidence of delirium and improve perceived patient sleep quality through optimization of environmental factors (noise, light, nursing activities) as well as nonenvironmental factors (pain, mechanical support devices, procedures) in a Cardiovascular and Surgical ICU. Methods: This QI project was implemented at a large academic facility in the intermountain west within two intensive care units (N = 596). The SICU is a 12-bed unit and the CVICU is a 20-bed unit comprised of adult post-operative, trauma, and critically ill patients. Using a pretest-posttest study design, baseline data was collected prior to implementation of the Sleep/Wake protocol after which post-implementation data was collected. The Sleep/Wake protocol is comprised of environmental interventions to complete during day and night shift. Measures: Primary outcome of delirium was measured using the confusion assessment methodICU. Secondary outcome of perceived sleep included a baseline sleep survey of sleep quality prior to hospitalization plus the Richards-Campbell Sleep Questionnaire. Nurse surveys were used to evaluate usability and satisfaction of the protocol. Results: During the pre- and post-intervention there were 272 and 229 patients respectively who were screened for delirium and who participated in the RCSQ. Incidence of delirium decreased from 82/272 (30%) pre- vs 55/229 (24%) (p-value 0.13) post-intervention. Compliance with the Sleep/Protocol ranged from 20-98%, respectively. Perceived sleep quality showed similar mean (SD) between both groups with the greatest difference made in noise level (before 65(22), after IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 3 68(20), p=0.16). Satisfaction of the protocol was evaluated via nurse survey. Overall, 88% of those surveyed were satisfied with the protocol. Conclusions: The implementation of a Sleep/Wake Protocol is feasible within a surgical and cardiovascular ICU and showed a decrease in incidence of delirium pre- to post-intervention, but was not statistically different. There was neither improved nor reduced patient sleep quality, however staff were overall satisfied with the protocol. IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 4 Introduction Critically ill and hemodynamically unstable patients are at risk for experiencing a decrease in quality and quantity of sleep (Kamdar, Needham, & Collop, 2012). The intensity of the care provided during a patients' stay in the intensive care unit (ICU) often leads patients to experience a lack of sleep (Pulak & Jensen, 2016). Sleep deprivation in critically ill patients can lead to changes in patient's cognition, increased hospital and ICU length of stay, and physiologic consequences (Beltrami et al., 2015; Cavallazzi, Saad, & Marik, 2012; Pulak & Jensen, 2014). Moreover, sleep deprivation in critically ill patients has been associated with an increase risk in the development of ICU delirium, which can impact a patient's ability to recover as well as increase their morbidity and mortality (Kamdar, et al., 2013; Patel et al., 2014). Alteration of the ICU patients' environment "may lead to more and better-quality sleep, with an associated decrease in delirium" as well as an "associat[ion] with significant reductions in perceived nighttime noise levels and a substantial decrease in delirium/coma" (Patel et al., p. 548, 2014; Kamdar et al., p. 808, 2013). ICU delirium is defined as an acute brain dysfunction that causes a change in a patients' cognition, attention, and consciousness (Delirium in the ICU: An Overview, 2013). Cognitive change has been shown to increase length of ICU and hospital stay, increase cost, decrease survival, and cause a permanent impairment in cognition (Delirium in the ICU: An Overview, 2013). The physiologic consequences of ICU delirium are influenced by patient clinical status, modifiable risk factors, and whether they are delirious. Some barriers for optimizing patient sleep were identified as patient disease severity, mechanical ventilation, sedation, noise, light, and nurse-patient interaction (Kamdar, Needham, & Collop, 2012). Furthermore, there is a "possible association" identified between patients' lack of sleep and cognition with additional consequences of longer mechanical ventilator time, and cardiovascular, pulmonary, and immune IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 5 system dysfunctions (Kamdar, Needham, & Collop, p. 801, 2012; Patel et al., p. 541, 2014). It has been suggested that reducing noise, nursing activities, and light to improve sleep "can reduce the incidence of delirium" (Brummel & Girard, p. 6, 2013). Patient sleep in the ICU is often fragmented with 50 percent of patient sleep hours occurring in short bouts throughout the day (Weinhouse & Schwab, 2006). This leads to a decreased in the patient's overall sleep quality (Weinhouse & Schwab, 2006). Approximately 4587 percent of critically ill patients will experience at least one episode of delirium during their ICU stay. Both patients and staff can identify environmental factors, such as noise, light, and nursing activities, as well as non-environmental factors such as mechanical devices, procedures, and medications as contributing factors to lack of sleep (Beltrami et al, 2015; Kamdar et al., 2013). However, both nursing staff and providers underestimate the intensity of the perceived noise in the ICU. This includes those frequent night time interruptions which can cause fragmented sleep (Hopper, Fried, & Pisani, 2015; White & Zomorodi, 2016). Nursing staff and providers are aware of the factors contributing to sleep deprivation, but they may have a "lower prioritization of patient sleep [as a] tradeoff [for] current standard of care (Hopper, Fried, & Pisani, 2015). An association between patient lack of sleep and the incidence of delirium has previously been identified, however, the exact cause is unclear (Brummel & Girard, 2013). We know that ICU delirium and sleep deprivation can be prevented with modification of environmental factors and early mobilization of critically ill patients (Schweickert, et al., 2009). A few studies have focused on improving patient sleep quality through clustering nursing care activities, reducing noise and light level, and managing pain (Faraklas et al., 2013; Ritmala-Castren, Virtanen, & Leivo, S., 2015; Giusti, Tuteri, D., & Giontella, M., 2016; White & Zomorodi, 2016). There have IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 6 been two studies, conducted in surgical and medical ICUs, which implemented a protocol or bundle to optimize factors contributing to lack of sleep. In conjunction with evaluating sleep, they evaluated incidence and duration of delirium. With implementation of a multicomponent protocol, both studies showed a reduction in incidence and duration of ICU delirium and an association of improved sleep quality (Kamdar, et al., 2013; Patel, et al., 2014). Theoretical Framework Social Cognitive Theory (SCT) was used to understand how to approach nursing behavioral change regarding the sleep/wake environment necessary to improve patient sleep in the ICU (Croyle, 2005). SCT describes the interactions between an individual's personal, behavioral, and environmental factors that contribute to their ability to change. To create this behavioral change, the nurses needed to feel efficacious in their ability to master the tasks required of them. We used four self-efficacy enhancing strategies theorized to change behavior 1) performance accomplishment, 2) verbal persuasion, 3) role modeling, and 4) self-appraisal. Performance accomplishment occurred with implementation of the sleep/wake protocol through feedback and the nurse's performance. Verbal persuasion was used through education of the protocol. Role modeling was used to demonstrate how to change the patient's environment to enhance sleep by the principal investigator and the ICU nurse champions. SCT theory underscores the importance of goal setting. Clear goals were set for each nurse to follow the protocol to create environmental changes to improve patient sleep quality. Lastly, we used selfappraisal by giving the nurses feedback on how well their patient(s) were sleeping each night as well as how closely they were following the protocol. Using the sleep measurement tools, nurses could see how their environmental changes via the protocol were improving patient sleep quality and reducing incidence of ICU delirium. IMPLEMENTATION OF SLEEP/WAKE PROTOCOL Specific Aim The purpose of this quality initiative was to implement a multicomponent Sleep/Wake protocol for the optimization of environmental factors (noise, light, nursing activities) as well as non-environmental factors (pain, mechanical support devices, procedures) to improve patient sleep quality and to reduce the incidence of ICU delirium in a Cardiovascular and Surgical intensive care unit (ICU). Methods Context This quality initiative was completed at a large academic facility in the intermountain west in two intensive care units, surgical (SICU) and cardiovascular (CVICU). The SICU is a 12-bed unit and the CVICU is 20-bed unit comprised of adult post-operative, trauma, and critically ill patients. There was a total of 47 SICU nurses and 93 CVICU nurses who were involved in this project with six nurses, three from each unit, dedicated as nurse champions. They received extra training to help guide other nurses on completing the required surveys and checklist items for data collection. Study Design This quality initiative used a pretest-posttest study design. The initial baseline data collection occurred for 2 months in the CVICU and 3 months in the SICU. Following baseline data collection, an educational session was provided to each unit to educate nursing staff on the Sleep/Wake Protocol and adapted ICU checklist. Education was continued during the first week of protocol and checklist implementation. Data collection continued for 3 months in the CVICU and 2 months in the SICU for evaluation on the implemented protocol and checklist. Staggering 7 IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 8 implementation of the protocol was essential to prevent alterations in data if hospital policy were to change during the initiative period. Subjects All adult patients admitted to the CVICU and SICU during the initiative period were initially included, however exclusion criteria were identified for patients who were comatose, on paralytics, and had no brain reflexes. Measures Baseline survey data was collected one time per patient during their stay. The information was gathered by the patient or by a close family member who knew the patients home sleep habits. The survey consisted of four multiple choice questions that ask about (1) a history of sleep problems (sleep apnea, restless leg syndrome, insomnia, or narcolepsy), (2) home sleep quality, (3) any medications taken at home that help with the patients sleep, and (4) whether the patients likes to sleep with the television on at night (Kamdar et al., 2013). Confusion Assessment Method-ICU (CAM-ICU) was used to measure delirium in ICU patients, which follows current clinical practice guidelines (Barr, et al., 2013). It is "the most valid and reliable delirium monitoring tool in adult ICU patients" (Barr et al., p. 265, 2013). Retrospective data collection identified delirium using ‘yes' or ‘no'. This method was validated by Ely et al. (2001) in non-verbal patients with a reported interrater reliability of k=0.96, a sensitivity of 93-100%, and a specificity of 98-100%. Richards-Campbell Sleep Questionnaire (RCSQ) was used to evaluate patient perceived quality of sleep in the ICU. This visual analog scale, range 0-100, was validated by Richards, O'Sullivan, & Phillips (2000) against polysomnography and has an internal consistency reliability of 0.90. Further evaluation determined that nurse's perception of patient sleep was IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 9 higher than patient's perception of sleep on all 5 questions (Kamdar et al., 2012). Recognizing these findings, we determined that nurses would only fill out the survey for their patients if the patient was unable to fill it out for themselves (i.e. delirious and sedated patients). Questions include perceptions on sleep depth, sleep latency, awakenings, returning to sleep, sleep quality, and noise. The sixth question was added after evaluation of environmental factors that affect sleep (Kamdar et al., 2012). Nurse Surveys were created to evaluate the usability and satisfaction of the protocol and checklist from a nursing perspective. There were five questions with answer categories of ‘yes' or ‘no', followed by a comment box where nurses could explain their answer. The sixth question had an answer category on a range from very dissatisfied to very satisfied. Questions include (1) was the protocol easy to use, (2) do you feel that the protocol was beneficial to your patients, (3) are there aspects of the protocol that you would change, (4) would you like this protocol to become standard practice in the unit, (5) did you encounter issues while implementing the protocol, and (6) rate your overall satisfaction of the protocol. Device Monitoring was completed via a checklist created in RedCaps. A retrospective chart review on each patient in the study was used to evaluate use of mechanical circulatory support devices (venous-venous or venous-arterial extracorporeal membranous oxygenation, external right or left ventricular assist devices), continuous renal replacement therapy, intraaortic balloon pump, Impella, mechanical ventilator, or BiPAP/CPAP use. These devices are invasive and require frequent monitoring by nurses contributing to noise level and interrupted sleep. Interventions The initial step was to complete RCSQ education the week before implementation of the data collection. No education was required for the CAM-ICU measurement as this was an IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 10 established practice within both ICU settings. Education on the RCSQ took approximately ten total hours. All nurses currently working on both units were trained and instructed on how to complete the RCSQ for or with their patients. To include all nurses in the education process, the principal investigator provided instructions each morning before night shift nurses went home and each afternoon for day shift nurses. At those times, nurses were given the chance to ask questions as well as give suggestions on how to improve the process of collecting data. The second step consisted of implementing the RCSQ. A clearly labeled ‘CVICU/SICU Sleep Study' binder was placed at each nursing station filled with blank surveys, dated tabs, and instructions on how to fill out the forms. Nurses were instructed to obtain blank surveys from the binder and file completed surveys under the date in which the survey was filled out. The third step involved continuing education to nurses during the first two weeks of RCSQ implementation, followed by planned, and as needed, weekly visits to improve compliance. During the first two weeks, it was noted that the task for handing out the RCSQ was difficult for the charge nurse to complete and difficult for the nurses to remember to fill out. To achieve a great number of compliance, the health unit coordinators (HUCs) were assigned to hand out the RCSQ each night. Each night the HUCs received a call from the principal investigator reminding them to hand out the survey. During the call, the charge nurse was instructed to educate any new or float nurse on how to properly fill out the RCSQ with their patients. This switch in the process helped to facilitate nursing compliance and provided a reminder to fill out the RCSQ. The fourth step was to conduct a retrospective chart review to obtain demographic data, CAM-ICU evaluation, and external devices. Chart review was completed by the principal IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 11 investigator and nurse champions who were educated on how to gather and input patient data into demographic forms. The fifth step included an educational session given to all nurses currently employed, on those units, on the Sleep/Wake Protocol and adapted ICU checklist (Kamdar et al., 2013). The principal investigator and nurse champions were available during the first week of protocol and checklist implementation to provide additional education if necessary. In addition to the educational session, an email was sent out with instructions on each component of the protocol. Nurses were encouraged to give suggestions on how best to adapt this new protocol into practice. Those suggestions were then integrated into the protocol to improve compliance. Further education was provided weekly and on an as needed basis until the end of the quality improvement initiative. The final step was to implement the Sleep/Wake Protocol and the adapted ICU checklist for 3 months in the CVICU and 2 months in the SICU. The Sleep/Wake Protocol consisted of three action steps. The first step was day shift actions occurring between 0700 and 2200. During this period, if applicable, nurses were encouraged to discontinue patient eye masks and/or ear plugs, turn on lights, open blinds and doors, make a reasonable effort for noise in the room via television, radio, family, or staff, apply hearing aids and eye glasses, encourage activities to prevent daytime napping, discourage caffeine intake after 3pm, and help the patient sit up or walk at least two times daily, unless contraindicated. The second step was night shift actions from 2200-0700. Prior to or at the start of this period the nurses were encouraged to have the following completed by 2200: offer ear plugs, eye mask, give ordered sleep medications, dim all lights and turn television off, close blinds, prevent unnecessary alarms by performing pump rounds, stay out of the room unless performing direct patient care, minimize nursing IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 12 interruptions, ask family and visitors to leave, partially or fully close the door, control patient pain, and optimize the room temperature. The third step was ICU-wide interventions which includes dimming the hallway lights by 2200, keeping the nurses station quiet, and placing the protocol stop sign on the patient's door. The adapted checklist included the components listed above and nurses were required to complete the list during their respective shifts. Sleep quality data collection continued using the RCSQ post-intervention of the protocol and checklist. Education occurred daily during the first week of implementation, followed by weekly and as needed until completion of the quality initiative. Study of the Interventions RCSQ completeness was evaluated daily during the first 2 weeks of implementation followed by a weekly evaluation. If information was missing, re-education on how to fill out the RCSQ completely was given to all nursing staff on shift. The CAM-ICU was an established part of ICU practice prior to implementation of this quality initiative. Education and further information was then given to each charge nurse, who was then instructed to pass on the information to the night shift team. Monthly emails were sent out to clarify, re-educate, and remind nurses about the importance of survey completeness. Analysis Descriptive statistics, including mean (standard deviation) was used to assess demographics and outcome variables on delirium and perceived sleep. Descriptive statistics were also used to described outcome variables of nurse usability and satisfaction of the Sleep/Wake Protocol. Categorical characteristics were compared using the Wilcoxon-Mann-Whitney, Chisquare, and fisher exact test to evaluate the changes between pre- and post-intervention groups. 95% confidence intervals and p-values were reported from all models. Statistical analyses were IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 13 conducted in R v.3.4. Significance was assessed at the 0.05 level and all tests are two-tailed. Patients were divided and analyzed in two groups, pre-and post-intervention. The primary outcome measured was delirium as categorical, yes verse no. Change statistics were used for analysis of delirium and perceived sleep from pre- to post-intervention. Ethical Considerations This study was determined to be a non-human subjects research by the University of Utah Institutional Review Board (IRB) and considered a quality initiative. Results Pre-intervention data were collected from 272 patients and post-intervention data were collected from 229 patients. Most participants were non-Hispanic/Latino, white, males with a mean age of 58(SD=17.2) years pre-intervention and 59(SD=17.8) years post-intervention (Table 1). Those with completed baseline surveys showed a history of pre-existing sleep problems with pre-intervention 66(24%) and post-intervention 83(36%), p=0.003. History of sleep apnea was shown among those patients as 27(10%) in the pre-intervention period vs 40(17%) in the postintervention period, p=0.013. Of those patients surveyed, 35(26%) pre-intervention and 22(14%) post-intervention stated that they take medication to help them sleep more than four nights per week. Baseline surveys also showed statistically significant differences in sleep quality in the pre- vs post-intervention group. There was a statistically significant difference in those patients who described their sleep as ‘somewhat bad' showing pre-intervention 19(14%) vs postintervention 38(24%), p=0.033. There was no statistically significant difference in the pre- vs post-intervention group among patients taking sleep medications prior to hospitalization (p=0.11) and sleeping with the television on at night (p=0.10). The results did show a statistically IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 14 significant different in those patients who used BiPAP/CPAP overnight between the preintervention, 25(9%) vs post-intervention group, 40(17%), p=0.006 (Table 1). Environmental Interventions The Sleep/Wake protocol day shift actions ranged from 42-94% compliance over the post-intervention collection period. The night shift actions ranged from 20-98% compliance (Table 2). The four lowest areas of compliance were chair position/mobility, warm bath before 2200, and ear plugs and eye masks offered and used. Warm bath compliance was low due to confusion among nursing staff on time of day the patient received a warm bath. The bath was only marked as completed when the patient received a warm bath between 1900 and 2200. However, most baths occur during the day shift hours. As shown in Table 2, ear plugs and eye masks were measured as offered and used, yet, those numbers do not reflect not applicable or offered and refused. Delirium and Sleep Outcomes ICU delirium was measured at the patient level and showed a reduced incidence from 82(30%) pre-intervention vs 55(24%) post-intervention, however the results showed no statistically significant difference, p=0.13 (Table 3). Sleep quality remained constant from pre- to post-intervention group with the greatest difference in noise level from pre-intervention 65(22%) to post-intervention 68(20%), p=0.16. Again, this data did not show a significant difference from pre- to post-intervention. Protocol Satisfaction A total of 146 nurses participated in this quality improvement initiative. Satisfaction surveys were emailed twice to each nurse with a response rate of 45 nurses (30%). There were six questions asked regarding the usability and feasibility of the Sleep/Wake protocol (Table 5). IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 15 Of the nurses surveyed 97% (44/45) agreed that the protocol was easy to use with one person stating there were "communication issues" on how to fill out the daily forms. Eighty-six percent of nurses (39/45) thought the protocol was beneficial to their patient population. Of those who did not think the protocol was beneficial, comments made were, the protocol did not apply to their patient population and they did not see how the protocol benefited their patients. Suggestions were made by nurses on what to change with the protocol to make it more applicable to the patient population (30/45, 66%). Some of those suggestions included incorporating the protocol into the electronic health record, afternoon quiet hours, and changing the start time from 2200 to 2300. There was an overall positive response to having the protocol become standard practice in the ICUs studied (39/45, 86%). Thirty-three percent (15/45) of those surveyed identified issues while implementing the protocol. Respondents noted that it was difficult to understand and use the RCSQ, difficult to minimize interruptions given patient hemodynamic status, and staff resistance to implementation. Overall, 40/45, 88% of nurses rated the Sleep/Wake protocol in the satisfied categories. Discussion Implementation of a Sleep/Wake protocol in two intensive care units at a large academic facility was feasible and nurse satisfaction with the protocol was high. ICU delirium rates decreased from the pre-intervention group (30%) to the post-intervention group (24%), although no statistically significant difference was noted. Our results show that perceived sleep quality remained unchanged from pre- to post-intervention group which indicates that we cannot show an association between altering patient environment to improve sleep and incidence of delirium. There was a large range of compliance with the Sleep/Wake protocol. It is possible that the low compliance rate for chair position/mobility is due to a higher rate of patients with orders for full IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 16 spine precautions and those who have devices or lines in their femoral artery/vein. A further analysis using a multivariate regression analysis controlling for variables such as home sleep habits, medical conditions, and mechanical devices may reveal more meaningful results. The negative results of this QI project will be further discussed to inform the next steps in the QI process to elucidate delirium and sleep quality in the ICU. Summary The components of the Sleep/Wake protocol were replicated based on the Kamdar et al. (2013) study on sleep in the medical ICU setting. Each aspect of the protocol is science-based as shown in the most current literature. Creating an environment that distinguishes between daytime and nighttime, such as lights on, blinds open, doors open, helps to keep the patient in their normal circadian rhythm (Engwall et al., 2015). Offering ear plugs and eye masks has been shown to improve patient sleep quality in the ICU (Dave, Qureshi, & Gopichandran, 2015). Minimizing nursing interactions is important as patients are frequently interrupted during the night for various cares which can impact the quality of sleep they receive (Giusti, Tuteri, & Giontella, 2016; Ritmala-Castren et al., 2015). Current clinical practice guidelines for management of delirium in the ICU setting is to incorporate early mobilization, manage pain, and optimize sleep by altering the patients' environment. That includes light and noise management, grouping nursing cares, and minimizing nursing interruptions throughout the night (Barr et al., 2013). The strengths of this project are the large sample size that was collected and the two different types of ICU environments studied. All patients were originally included as aspects of the Sleep/Wake Protocol are considered best practice. Interpretation IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 17 Two similar studies were conducted by Kamdar et al. (2013) and Patel et al. (2014), that studied sleep quality and delirium in the ICU. Both Kamdar et al. (2013) and Patel et al. (2014) found a statistically significant difference in sleep quality and ICU delirium from their preintervention group to their post-intervention group. Compliance for the interventions were higher in both projects comparatively to this QI project (Kamdar et al., 2013, >86% compliance and Patel et al., 2014, >90%). It is important to note that Kamdar et al. (2013) included a protocol phase that incorporated sleep medications such as Zolpidem and Haloperidol, into their nightly routine. Another difference between Kamdar et al. (2013) and Patel et al. (2014) was the analysis design. This QI projected used a univariate approach that did not control for variables, whereas Kamdar et al. (2013) and Patel et al. (2014) used a multivariate approach to control for variables such as ICU admission diagnosis, home sleep habits, number of nursing interruptions, mechanical ventilation status, APACHE-2 score, and length of stay. Both Kamdar et al. (2013) and Patel et al. (2014) had no significant difference between pre-intervention and postintervention demographic data. When comparing Kamdar et al. (2013) and Patel et al. (2014) demographic data to this QI project, there are significant differences in home sleep habits, home sleep quality, and use of BiPAP/CPAP from pre- to post-intervention groups (Table 1). The populations studied in Kamdar et al. (2013) and Patel et al. (2014) focused on patients in a medical and surgical ICU respectively. However, this QI project included patients from both a basic surgical ICU and a specialized cardiovascular ICU. This QI project also had a larger sample size (N=501) vs Kamdar et al. (2013) (N=300) vs Patel et al. (2014) (N=338). Exclusion criteria was different among those studies and this QI project which excluded patients who were comatose, on paralytics, or had no brain reflexes. Whereas Kamdar et al. (2013) only IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 18 excluded comatose patients and Patel et al. (2014) has more extensive exclusion criteria including, but not limited to, patients who had pre-existing sleep problems or those or developed delirium during the study. Our next step is to conduct a multivariate regression analysis to account for variables such as home sleep habits, medical conditions, and mechanical devices that can possibly influence patient sleep quality and rates delirium. Limitations There were limitations identified during this project which include the use of a subjective measurement of sleep, RCSQ, rather than an objective measurement, polysomnography. Using polysomnography would not have been feasible in this quality improvement project as all patients admitted during the study period were initially included. Compliance to the Sleep/Wake protocol created some limitations indicating that portions of the protocol were more difficult to implement. Conclusion In conclusion, this is the first large scale quality improvement project in this area involving altering patient environment to improve sleep quality. It was feasible to implement a sleep/wake protocol in two intensive care units and usability/satisfaction was high among nursing staff. This study did not show a statistically significant difference in the reduction of ICU delirium or perceived sleep quality pre- to post-intervention. The next step is to conduct an analysis to control for variables such as home sleep habits, medical conditions, and mechanical devices that have been known to influence findings in similar studies. Although this study does not show an association between sleep quality and delirium through this intervention, the science supporting the importance of maintaining an environment conducive for adequate sleep to IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 19 continue in preventing ICU delirium is persuasive. Further study needs to be conducted to make a greater impact on patient environment to enhance their sleep during critical illness. Acknowledgements We would like to thank the University of Utah CVICU and SICU nurses, aides, and health unit coordinators whom without we would not have been able to implement the Sleep/Wake protocol and collect important data regarding sleep and delirium in the ICU. We would also like to thank our nurse champions who spent extra time providing education to nurses and assisting where needed. IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 20 References Barr, J., Fraser, G.L., Puntillo, K., Ely, E.W., Gelinas, C., Dasta, J.F., … Jaeschke, R. (2013). Clinical Practice Guidelines for the Management of Pain, Agitation, and Delirium in Adult Patients in the Intensive Care Unit. Critical Care Medicine, 41(1), 263-306. https://doi.org/10.1097/CCM.0b013e3182783b72 Beltrami, F.G., Nguyen, X-L., Pichereau, C., Maury, E., Fleury, B., & Fagondes, S. (2015). Sleep in the intensive care unit. J Bras Pneumol, 41(6), 539-546. https://doi.org/dx.doi.org/10.1590/s1806-37562015000000056 Brummel, N.E., & Girard, T.D. (2013). Preventing delirium in the intensive care unit. Crit Care Clin, 29(1), 51-65. https://doi.org/doi:10.1016/j.ccc.2012.10.007. Cavallazzi, R., Saad, M., & Marik, P.E. (2012). Delirium in the ICU: an overview. Annals of Intensive Care, 2(49), 1-11. https://doi.org/https://doi.org/10.1186/2110-5820-2-49 Croyle, R. T. (2005). Theory at a Glance: A Guide for Health Promotion Practice. National Institute of Health. Retrieved from http://www.sbccimplementationkits.org/demandrmnch/wpcontent/uploads/2014/02/Theory-at-a-Glance-A-Guide-For-Health-PromotionPractice.pdf Dave, K., Qureshi, A., & Gopichandran, L. (2015). Effects of Earplugs and Eye Masks on Perceived Quality of Sleep during Night among Patients in Intensive Care Units. Asian Journal of Nursing Education and Research, 5(3), 319-322. https://doi.org/10.5958/2349-2996.2015.00065.8 Delirium in the ICU: An Overview. (2013). Vanderbilt University. Retrieved from http://icudelirium.org/delirium.html IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 21 Ely, E.W., Inouye, S.K., Bernard, G.R., Gordon, S., Francis, J., May, L., … Dittus, R. (2001). Delirium in Mechanically Ventilated Patients: Validity and Reliability of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU). JAMA, 286(21), 2703- 2710. Engwall, M., Fridh, I., Johansson, L., Bergbom, I., & Lindahl, B. (2015). Lighting, sleep and circadian rhythm: An intervention study in the intensive care unit. Intensive and Critical Care Nursing, 31, 325-335. https://doi.org/http://dx.doi.org/10.1016/j.iccn.2015.07.001 Faraklas, I., Holt, B., Tran, S., Lin, H., Saffle, J., & Cochran, A. (2013). Impact of a NursingDriven Sleep Hygiene Protocol on Sleep Quality. Journal of Burn Care & Research, 34(2), 249-254. https://doi.org/10.1097/BCR.0b013e318283d175 Giusti, G.D., Tuteri, D., & Giontella, M. (2016). Nursing Interactions With Intensive Care Unit Patients Affected by Sleep Deprivation. Dimensions of Critical Care Nursing, 35(3), 154-159. https://doi.org/10.1097/DCC.0000000000000177 Hopper, K., Fried, T.R., & Pisani, M.A. (2015). Health care worker attitudes and identified barriers to patient sleep in the medical intensive care unit. Heart & Lung, 44, 95-99. https://doi.org/http://dx.doi.org/10.1016/j.hrtlng.2015.01.011 Kamdar, B.B., King, L.M., Collop, N.A., Sakamuri, S., Colantuoni, E., Neufeld, K.J., … Needham, D.M. (2013). The Effect of a Quality Improvement Intervention on Perceived Sleep Quality and Cognition in a Medical ICU. Critical Care Medicine Journal, 41(3), 800-809. https://doi.org/DOI: 10.1097/CCM.0b013e3182746442 Kamdar, B.B., Needham, D.M., & Collop, N.A. (2012). Sleep Deprivation in Critical Illness: Its Role in Physical and Psychological Recovery. J Intensive Care Med, 27(2), 97-111. https://doi.org/doi:10.1177/0885066610394322 IMPLEMENTATION OF SLEEP/WAKE PROTOCOL 22 Kamdar, B.B., Shah, P.A., King, L.M., Kho, M.E., Zhou, X., Colantuoni, E., … Needham, D.M. (2012). Patient-Nurse Interrater Reliability and Agreement of the Richards-Campbell Sleep Questionnaire. Am J Crit Care, 21(4), 261-269. https://doi.org/doi:10.4037/ajcc2012111 Patel, J., Baldwin, J., Bunting, P., & Laha, S. (2014). The effect of a multicomponent multidisciplinary bundle of interventions on sleep and delirium in medical and surgical intensive care patients. Anaesthesia, 69, 540-549. https://doi.org/doi:10.1111/anae.12638 Pulak, L.M., & Jensen, L. (2014). Sleep in the Intensive Care Unit: A Review. Journal of Intensive Care Medicine, 31(1), 14-23. https://doi.org/10.1177/0885066614538749 Richards, K.C., O'Sullivan, P.S., & Phillips, R.L. (2000). Measurement of sleep in critically ill patients. J Nurs Meas, 8(2), 131-144. https://doi.org/https://www.ncbi.nlm.nih.gov/pubmed/11227580 Ritmala-Castren, M., Virtanen, I., & Leivo, S. (2015). Sleep and nursing care activities in an intensive care unit. Nursing and Health Science, 17, 354-361. https://doi.org/doi: 10.1111/nhs.12195 Schweickert, W.D., Pohlman, M.C., Pohlman, A.S., Nigos, C., Pawlik, A.J., Esbrook, C.L., … Kress, J.P. (2009). Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial. Lancet, 30(373), 1874-1882. https://doi.org/doi: 10.1016/S0140-6736(09)60658-9 Weinhouse, G.L., & Schwab, R.J. (2006). Sleep in the Critically Ill Patient. SLEEP, 29(5), 707- 716. IMPLEMENTATION OF SLEEP/WAKE PROTOCOL White, B.L., & Zomorodi, M. (2016). Perceived and actual noise levels in critical care units. Intensive and Critical Care Nursing, 38, 18-23. https://doi.org/dx.doi.org/10.1016/j.iccn.2016.06.004 23 IMPLEMENTATION OF SLEEP/WAKE PROTOCOL TABLE 1. Participant Characteristics stratified by pre/post intervention PrePostCharacteristic Intervention Intervention (n=272) (n=229) 58.3 (17.2) 58.6 (17.8) Age, mean(SD) 173 (64) 145 (63) Male, n(%) Race 221 (81) 186 (81) White or Caucasian, n(%) 4 (1) 4 (2) Black or African American, n(%) 8 (3) 4 (2) American Indian and Alaska Native, n(%) 6 (2) 2 (1) Asian, n(%) 2 (1) 7 (3) Native Hawaiian and other Pacific Islander 26 (10) 15 (7) Other Ethnicity 21 (8) 18 (8) Hispanic/Latino, n(%) 242 (89) 198 (86) Not Hispanic/Latino, n(%) Baseline Home Sleep Questionnaire 66 (24) 83 (36) Pre-existing history of sleep problems, n(%) 27 (10) 40 (17) History of Sleep Apnea 34 (12) 29 (13) History of Insomnia Describes home sleep quality as, n(%) 41 (30) 43 (28) Very good 59 (44) 61 (39) Somewhat good 19 (14) 38 (24) Somewhat bad 16 (12) 10 (6) Very bad 0 (0) 4 (3) Unknown Taking Sleep Medications at Home, n(%) 82 (60) 103 (66) None 6 (4) 12 (8) Less than once a week 7 (5) 10 (6) 1~2 times a week 5 (4) 9 (6) 3~4 times a week 35 (26) 22 (14) >4 times a week 1 (1) 0 (0) Unknown 33 (24) 51 (33) Sleeps with the TV on at night at home, n(%) Ever used a mechanical device overnight, n(%) 17 (6) 12 (5) Continuous Renal Replacement Therapy 11 (4) 3 (1) Mechanical Circulatory Support Devices (ECMO, RVAD, External LVAD) 2 (1) 4 (2) Impella 4 (1) 6 (3) Intraaortic Balloon Pump 113 (42) 77 (34) Ventilator 24 PValue 0.95 0.12 0.56 0.003 0.013 0.96 0.033 0.11 0.10 0.63 0.06 0.42 0.52 0.07 IMPLEMENTATION OF SLEEP/WAKE PROTOCOL BIPAP/CPAP 25 25 (9) 40 (17) 0.006 TABLE 2. Compliance with Sleep/Wake Protocol - Post-Intervention group only Intervention Compliance (n=229, n(%)) Day Shift Actions 118 (64) No caffeine After 3pm 123 (70) % of day shift spent Napping 194 (93) Lights on, Blinds open, Doors open 194 (94) Reasonable effort for noise in room 113 (73) Eye glasses, hearing aids applied 84 (42) Chair position/mobility at least 2x30 minutes daily Night Shift Actions 201 (91) Appropriate Pain Control 224 (98) Optimize Room Temperature 80 (35) Warm Bath before 2200 145 (64) TV off by 2200 208 (91) Prevent unnecessary alarms after 2200 191 (83) Close curtains by 2200 184 (80) Dim room light by 2200 179 (78) Family out of room by 2200 197 (86) Door partially or fully closed by 2200 164 (73) Minimized RN Interruptions after 2200 7.1 (4.6) # of Interruptions after 2200, mean(SD) 42 (20) Eye Masks Offered and Used 43 (21) Ear Plugs Offered and Used ICU Wide-Interventions 220 (97) Dim Hallway lights by 2200 209 (92) Nurses station quiet 183 (81) Stop sign placed on Patient's Door TABLE 3. ICU Delirium and Perceived Sleep with Protocol Intervention by Pre/Post-Intervention ICU Outcome PrePostPost-Pre PIntervention Intervention (95% CI) Value (n=272) (n=229) 82 (30) 55 (24) -6(-14 ~ 2) 0.13 Any delirium, n(%) RCSQ 51.1 (26.4) 48.7 (24.3) -2.36(-6.87~2.15) 0.30 Sleep Depth, Mean(SD) 56 (23.9) 56.6 (24.1) 0.59(-3.71~4.90) 0.79 Sleep Latency, Mean (SD) 52.1 (24.3) 52.8 (23.4) 0.69(-3.58~4.95) 0.75 # of Awakenings, Mean (SD) 57.5 (25.2) 58.7 (23.4) 1.17(-3.16~5.51) 0.59 Ability to Return to Sleep, Mean (SD) 49.5 (25.9) 48.9 (24.7) -0.60(-5.12~3.92) 0.79 Sleep Quality, Mean (SD) 65.1 (22.2) 67.8 (19.6) 2.67(-1.05~6.39) 0.16 Noise Level, Mean (SD) IMPLEMENTATION OF SLEEP/WAKE PROTOCOL TABLE 4. Nurse Usability and Satisfaction N = 45 YES 44/45 (97%) Was the protocol easy to use? 39/45 (86%) Do you feel that the protocol was beneficial to your patients 26 NO 1/45 (0.02%) 6/45 (13%) Are there aspects of the protocol that you would change? 15/45 (33%) 30/45 (66%) Would you like this protocol to become standard practice in the unit? Did you encounter any issues while implementing the protocol? 39/45 (86%) 6/45 (13%) 15/45 (33%) 30/45 (66%) Rate your overall satisfaction with this protocol RN Comments -Communication issues -Unsure if the protocol worked -Unsure of what protocol is -Doesn't apply to our patient population -Good reminder on thinking about patient sleep -Incorporate into EMR -No daily checklists or surveys -Need better initial education -Afternoon quiet hours -Some patients too sick for protocol -Change start time to 2300 -Too much paperwork -Concept seems reasonable without the daily sheets -Difficulty using the RCSQ -Minimal interruptions hard with ICU patients -Difficult time finding sleep masks and ear plugs -Staff resistance (nursing and MD/PA/NP) Very Satisfied Moderately Satisfied Satisfied Moderately dissatisfied Very Dissatisfied 12/45 (26%) 19/45 (42%) 9/45 (20%) 5/45 (11%) 0