Protecting Sleep to Reduce Delirium in an Adult Intensive Care Unit

Practice Problem: Delirium is a common, yet often preventable complication in hospitalized patients. It is often caused by fragmented sleep, medications, environmental stimuli, and treatment therapies. PICOT: The PICOT question that guided this evidence-based practice change project was: For patients in an adult Intensive Care Unit (ICU), does using a nurse-initiated, non-pharmacological sleep-enhancement protocol, versus no sleep-enhancement protocol, reduce the incidence of ICU-delirium over a period of 2 months? Evidence: The reviewed literature supported the evidence for effective use of a nurse-initiated protocol in reducing delirium in the ICU. Sixteen articles met the inclusion criteria for the review of literature that supported the DNP project. Intervention: A nurse-initiated sleep-enhancement protocol was implemented, which reduced interruptions during the hours between midnight and 0400. Outcome: While there was a 50% reduction in delirious patients after the protocol was initiated, the data pool was small and was not proven to be statistically significant. Due to the Covid-19 outbreak, there were far fewer eligible patients than originally anticipated. Conclusion: Although a clinically significant project outcome was not realized, staff felt that the protocol improved patient care and advocated for its use on all patients as a standard of care. Unit-based shared governance councils on other acute care floors have also adopted the sleep enhancement protocol PROTECTING SLEEP TO REDUCE DELIRIUM 4 Protecting Sleep to Reduce Delirium in an Adult Intensive Care Unit Sleep, a critical factor in health recovery, has been noted to be fragmented in critical care settings, resulting in a correlative effect between sleep deprivation and delirium (Faraklas et al., 2013). Delirium is defined as a temporary syndrome of cognitive disability characterized by disorganized thinking and accompanied by a decreased ability to maintain attention (American Psychiatric Association, 2013). Sometimes even fatal, delirium is the most frequent complication affecting hospitalized patients over 65 years of age (Inouye et al., 2014). Sleep deprivation in the intensive care unit (ICU) can be attributed to loud noises, bright lights, monitors, frequent assessments, mechanical ventilation, and sedative or analgesic medications (Patel et al., 2014). Delirium, while often preventable, is prevalent in the ICU and is detrimental to restoration of health (Pessoa et al., 2019). Some negative outcomes caused by ICU delirium are increased mortality, increased hospital length of stay (LOS), increased financial cost of care, and long-term cognitive impairment (Barr et al., 2013; Inouye et al., 2014). Researchers have found a correlation between the length of time a patient suffered delirium and the level of functional disability for up to a year after discharge (Altman et al., 2018). Reducing the incidence of delirium has even been found to decrease the number of patient falls (Ferguson et al., 2018). Involving willing family members of ICU patients in delirium-prevention measures and decreasing the frequency of neurological assessments have been offered as feasible interventions to reduce delirium in the ICU (McLaughlin et al., 2018; Smithburger et al., 2017). Guidelines and protocols supporting protected rest periods in ICUs are ideal (Li et al., 2011). Each of the aforementioned studies looked at different interventions regarding their effect on delirium, and all acknowledged that lack of sleep is known to exacerbate it. This evidence-based practice change project implemented guidelines that promote noise and light control, and the PROTECTING SLEEP TO REDUCE DELIRIUM 5 reduction of non-emergent interventions during “quiet time” to maintain circadian rhythm and reduce sleep deprivation and fragmentation, thereby reducing ICU-associated delirium. Significance of the Practice Problem The effects of protected sleep in the ICU may positively impact patients’ outcomes by reducing delirium, LOS, and overall mortality (Irwin et al., 2016). The impact on the healthcare system may be profound because shortened LOS reduces the financial cost of recovery to the patient and the healthcare system. Families may also be positively impacted because they may be able to avoid the distress of seeing their loved ones in a state of delirium and confusion. The financial burden of treating delirium in the United States is a staggering $164 billion per year (Ferguson et al., 2018). In 2010 the average daily cost of an ICU stay was $4,300 (Society of Critical Care Medicine, n.d.). More recent data suggested an initial-day ICU expense of $10,794, which levels out to $3,968 by day 3, depending on whether the patient requires mechanical ventilation (Cunningham et al., 2020). Longer ICU LOS increases morbidity and mortality (Altman et al., 2018). There is a reliable association between sleep fragmentation and delirium (Irwin et al., 2016; Patel et al., 2014). One study, conducted in a neurological ICU, suggested the de-escalation of neuro-status checks after a patient has been there for more than 48 hours. The authors acknowledge that although the frequency of neurological assessment is initially done to avoid missing a key finding, after a certain time, it causes more harm than good (McLaughlin et al., 2018). The strength of this recommended practice change comes from the algorithm wherein sleep hygiene is a key recommendation in the diagnosis of ICU delirium, based on patient-oriented evidence and not opinion, consensus, or usual practice (Ebell et al., 2004). PROTECTING SLEEP TO REDUCE DELIRIUM 6 The hospital where this project was implemented has not historically collected data specifically related to ICU delirium, but they do closely monitor LOS. Compared with national benchmarks, this facility has higher than expected LOS for several diagnoses. The Strategic Analytics for Improvement and Learning (SAIL) Value Model is a system used within the organization for data collection and quality improvement project analysis (SAIL, n.d.).


Protecting Sleep to Reduce Delirium in an Adult Intensive Care Unit
Sleep, a critical factor in health recovery, has been noted to be fragmented in critical care settings, resulting in a correlative effect between sleep deprivation and delirium (Faraklas et al., 2013). Delirium is defined as a temporary syndrome of cognitive disability characterized by disorganized thinking and accompanied by a decreased ability to maintain attention (American Psychiatric Association, 2013). Sometimes even fatal, delirium is the most frequent complication affecting hospitalized patients over 65 years of age (Inouye et al., 2014).
Sleep deprivation in the intensive care unit (ICU) can be attributed to loud noises, bright lights, monitors, frequent assessments, mechanical ventilation, and sedative or analgesic medications (Patel et al., 2014). Delirium, while often preventable, is prevalent in the ICU and is detrimental to restoration of health (Pessoa et al., 2019). Some negative outcomes caused by ICU delirium are increased mortality, increased hospital length of stay (LOS), increased financial cost of care, and long-term cognitive impairment (Barr et al., 2013;Inouye et al., 2014). Researchers have found a correlation between the length of time a patient suffered delirium and the level of functional disability for up to a year after discharge (Altman et al., 2018). Reducing the incidence of delirium has even been found to decrease the number of patient falls (Ferguson et al., 2018). Involving willing family members of ICU patients in delirium-prevention measures and decreasing the frequency of neurological assessments have been offered as feasible interventions to reduce delirium in the ICU (McLaughlin et al., 2018;Smithburger et al., 2017).
Guidelines and protocols supporting protected rest periods in ICUs are ideal (Li et al., 2011).
Each of the aforementioned studies looked at different interventions regarding their effect on delirium, and all acknowledged that lack of sleep is known to exacerbate it. This evidence-based practice change project implemented guidelines that promote noise and light control, and the reduction of non-emergent interventions during "quiet time" to maintain circadian rhythm and reduce sleep deprivation and fragmentation, thereby reducing ICU-associated delirium.

Significance of the Practice Problem
The effects of protected sleep in the ICU may positively impact patients' outcomes by reducing delirium, LOS, and overall mortality (Irwin et al., 2016). The impact on the healthcare system may be profound because shortened LOS reduces the financial cost of recovery to the patient and the healthcare system. Families may also be positively impacted because they may be able to avoid the distress of seeing their loved ones in a state of delirium and confusion.
The financial burden of treating delirium in the United States is a staggering $164 billion per year (Ferguson et al., 2018). In 2010 the average daily cost of an ICU stay was $4,300 (Society of Critical Care Medicine, n.d.). More recent data suggested an initial-day ICU expense of $10,794, which levels out to $3,968 by day 3, depending on whether the patient requires mechanical ventilation (Cunningham et al., 2020). Longer ICU LOS increases morbidity and mortality (Altman et al., 2018). There is a reliable association between sleep fragmentation and delirium (Irwin et al., 2016;Patel et al., 2014). One study, conducted in a neurological ICU, suggested the de-escalation of neuro-status checks after a patient has been there for more than 48 hours. The authors acknowledge that although the frequency of neurological assessment is initially done to avoid missing a key finding, after a certain time, it causes more harm than good (McLaughlin et al., 2018). The strength of this recommended practice change comes from the algorithm wherein sleep hygiene is a key recommendation in the diagnosis of ICU delirium, based on patient-oriented evidence and not opinion, consensus, or usual practice .
The hospital where this project was implemented has not historically collected data specifically related to ICU delirium, but they do closely monitor LOS. Compared with national benchmarks, this facility has higher than expected LOS for several diagnoses. The Strategic Analytics for Improvement and Learning (SAIL) Value Model is a system used within the organization for data collection and quality improvement project analysis (SAIL, n.d.).

PICOT Question
This project sought to answer a practice question using the PICOT format, which means that for a certain population (P), the implementation of an intervention (I), compared with not doing the intervention (C), leads to an outcome (O) over time (T). For this project, the following PICOT question was addressed: For patients in an adult ICU, does using a nurse-initiated, nonpharmacological sleep-enhancement protocol, versus no sleep-enhancement protocol, reduce the incidence of ICU-delirium over 2 months?
This hospital is unique in that there are two divisions, Uptown and Downtown, separated by approximately 4 miles. Uptown houses nursing homes, rehabilitation units, and many outpatient clinics. Downtown houses the Spinal Cord Injury Unit (SCIU), three acute care units, and a Critical Care Unit (CCU), which comprises the Intensive Care Unit (ICU) and the Stepdown Unit (SDU).
A single team governs both facilities and reports to the organization's Integrated Service Network (VISN). Local leadership is led by a facility director who has a chief of medicine and a chief of nursing as direct reports, supported by many service line chiefs, then the unit managers and assistant managers.
In order to ensure consistent application of the protocol, bedside nurses and physicians were trained by PowerPoint-supported inservices regarding the detriment of delirium and the importance of protected rest periods. Staff members were post-tested on the purpose of the project and rationale in the hopes of gaining buy-in as to the importance of it. Once all were inserviced on the sleep enhancement protocol a firm start date was determined. Ancillary departments were notified of the practice changes, so they were able to adapt their workflow as needed. All CCU patients received the intervention, but those that were admitted for fewer than 24 hours or were comatose at the time of admission have been excluded from data collection.
Because chemical paralysis would preclude an accurate assessment of delirium presence, those patients have also been excluded from the data.

Evidence-Based Practice Framework and Change Theory
Using a three-step process called PET (practice question, evidence, and translation), the Johns Hopkins Nursing Evidence-Based Practice (JHNEBP) model framed the quality improvement initiative for this project. Each of the PET steps are supported by sub-steps intended to facilitate project success. For instance, in the practice question phase, a problem was identified, and the team recruited. In the evidence phase, literature supporting the practice change was synthesized and a recommendation for change was made. In the translation phase, the action plan was created and implemented, then evaluated and reported. The goal of this model was to expeditiously apply research into bedside practice (Dang & Dearholt, 2017). The JHNEBP model was the best fit for this project because it is an uncomplicated model for implementation of best practice (Schaffer et al., 2013). The PET process (see Appendix A) details the steps of implementation. See Appendix B for permission letter to use the PET tool.
The theory that best fit this practice change is Lewin's Change Theory, which has three phases: unfreezing, movement, and refreezing (Hussain, 2018;Lewin, 1947). This process is prevalent in practice literature (Stouten et al., 2018). The initial phase, unfreezing, is intended to allow time to develop a vision and a plan for change. Unfreezing will identify the problem and possible cause. Garnering support from stakeholders in this phase, or earlier, bolsters the chance of success for the project (Hussain, 2018). The second phase involves the movement of applying the change. In this case, movement is where the protocol implementation happens, and measurement begins and is evaluated in terms of improvement (Lewin, 1947, as cited in Hussain, 2018. Refreezing, the final phase, entails making the change or new processes permanent by solidifying it in the organization (Lewin, 1947, as cited in Hussain, 2018Stouten et al., 2018).
Lewin's Change Model applied to this project because it spoke to the simplicity of the process of movement while also acknowledging the difficulty of it. Unless nurses understood the negative impact of the status-quo, and were included in the process of change, they may have abandoned the project or, worse, never supported it at all.

Evidence Search Strategy
A review of literature was conducted using the following databases: The University of St.
Augustine for Health Sciences' Library, EBSCO, Cumulative Index to Nursing and Allied Health (CINAHL), MEDLINE, Google Scholar, PubMed, and Ovid. The basic search began, unfiltered, with "non-pharmacologic sleep protection in the ICU." This search yielded 1605 results over several decades. Initial exclusions included articles that were not written in English, not immediately available as a full document, or that were pediatric based. Subsequent exclusions included articles focused on delirium in patients with head injuries, brain tumors, or other pre-existing neurological disorders. Applying filters using the Boolean operators, including "AND" and "OR" to form relevant statements that included prevention of delirium, reduction of delirium, non-pharmacologic, nursing protocols, and ICU reduced the results to 201 citations and articles. The following Medical Subject Headings ("delirium," "delirium AND nurses," "nurses AND initiated," "initiated AND nonpharmacologic") were added to narrow the selection of articles Comparative words ("reduction" and "intervention") were also included. This reduced the number of articles to 68. The time frame was limited to 2013-2020, which brought the results to 16 articles. See Table 1 for complete inclusion and exclusion criteria.

Evidence Search Results
The Strength of Recommendation Taxonomy (SORT) grading tool was used to categorize recommendations in the clinical review articles . Articles represented a variety of methods and designs including prospective-retrospective, quasiexperimental, randomized, systematic review, and descriptive. The qualifying articles were included based on the JHNEBP grading from levels I to V for project use. Figure 1, the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA), illustrates the summary of results.

Themes with Practice Recommendations
The literature was assessed and synthesized for common themes (see Appendices C and D). Reliable evidence was found supporting nurse-initiated, nonpharmacologic promotion of sleep for the purpose of reducing or preventing delirium. The Society of Critical Care Medicine (SCCM) stated in its latest revision of the Clinical Practice Guidelines for the Management of Pain, Agitation, Delirium and Immobility (PADI): "We recommend promoting sleep in adult ICU patients by optimizing patients' environments, using strategies to control light and noise, clustering patient care activities, and decreasing stimuli at night to protect patients' sleep cycles" (Barr et al., 2013, p. 266). To that end, each of the following themes supports a quality change initiative.

Intentional Quiet Time
Intentional quiet time includes several simple, but often overlooked interventions like noise reduction by turning off televisions, closing doors when safe and appropriate to do so, eliminating overhead pages, and turning down nonessential monitor volume (Faraklas, et al., 2013;Kram et al., 2015;Patel et al., 2014). McLaughlin et al. (2018) suggested that for the nonneurologic patient, serial neurologic checks may be unnecessary once baseline is established.

Clustered Care
Avoiding fragmentation of sleep included care clustering, where disruptions for the resting patient are limited (Kram et al., 2015;Rivosecchi, et al., 2016). An example would be doing med pass, lab work, and chest x-ray all in the same interaction instead of three separate interactions. This required intentional coordination with ancillary departments.

Reorientation
Constant reorientation measures included updating whiteboards, clocks, and calendars (Birge & Aydin, 2017;Faraklas, 2013). Consistent presence of family members may help mitigate the incidence and duration of delirium in the ICU (Smithburger et al., 2017).

Practice Recommendations
The specific aim of this project was to reduce the incidence of delirium in the ICU.
Evidence from the literature suggested a direct link between improving sleep and reducing delirium (Irwin et al., 2016;Li et al., 2011;Patel et al., 2014). Some suggestions for improved sleep were utilizing protocolized care bundles that include clustering of care to provide intentional rest periods, avoiding benzodiazepines, manipulating the physical environment, and reorienting the patient frequently (Farklas et al., 2013;Irwin et al., 2016;Li et al., 2011).
Delirium causes increased LOS and mortality and reduced functional ability for up to a year after discharge (Altman et al., 2018;Faraklas et al., 2013;Ferguson et al., 2018). Evidence further suggested that utilizing a nurse-initiated sleep enhancement protocol in the ICU may have a positive financial impact to the organization, since the average increase in cost of ICU care once delirium develops is 39% (Vasilevskis et al., 2018). This project sought to follow best practices to reduce delirium by nonpharmacologically reducing fragmentation of sleep; the cost to implement was negligible. The greatest investment was that of time: on the front-end training nurses about the project and reinforcing the standard assessment tools; on the back end the analysis and dissemination of data. The organization already had a comprehensive training manual for using the CAM-ICU which was embedded into one of the computer-based training platforms (Elsevier) and tracked for compliance (Boehm et al., 2016). The practice change recommendation was to check each shift for ICU-associated delirium using an evidence-based assessment tool, the CAM-ICU (Boesen et al., 2015;Li et al., 2011;Patel, et al., 2014). Nurses were uniquely positioned to assess for and recognize delirium, since they spend much of their work time at the bedside (Pessoa et al., 2019). Expected outcomes included the implementation of a facility-approved, nurse-initiated protocol (NIP) which resulted in a slightly reduced rate of delirium in this ICU population.

Project Setting
This project took place in the 20-bed Critical Care Unit (CCU) of a hospital in eastern Georgia. Ten CCU beds are designated as Intensive Care Unit (ICU) beds and ten are Stepdown Unit (SDU) beds. The pre-pandemic average daily hospital census was 78-85 % total capacity, with occasional community diversions when the unit was at capacity (Hospital website, n.d.). On the ICU side, nurse to patient ratio was 1:2 except for high-acuity procedures (e.g., continuous dialysis, targeted temperature management, or balloon pump). On the SDU side the nurse-topatient ratio was 1:3. Ever-present in the CCU is tele-ICU, which represents 24/7 nursing coverage, and board-certified physician coverage during off-tour hours when attending physicians were not in-house. Resident physicians were always in the facility, but tele-ICU serves as a second set of eyes for nurses and providers. Other support staff necessary for the project's success included pharmacists, respiratory therapists, nurses' aides, students, and unit clerks.

Organizational Structure
The hospital where this project was implemented is a two-division facility with three affiliated community-based outreach centers. The downtown division houses the CCU, as well as 156 acute care and Spinal Cord Injury Unit (SCIU) beds. It also contains an invasive cardiology suite with two cardiac catheterization laboratories and a fully operational 16-bed Emergency Department (ED).

Organizational Culture
Organizational core values of this hospital system are Integrity, Commitment, Advocacy, Respect, and Excellence (I-CARE) which encapsulates the vision and supports the organization's mission to fulfill President Lincoln's (1865) promise "to care for him who shall have borne the battle, and for his widow, and his orphan" (p. 4). The organization is led by the director, the associate director, the assistant director, associate director of patient care services, and the medical director. This team is known locally as "The Pentad." The next tier of leadership includes service line chiefs, followed by managers, and then the front-line care staff.

Organizational Need
At the outset of this project, there were no standard operating procedures or protocols for CCU staff to follow or fall back on when delirium was suspected. The same was true at the end of the project. Delirium occurrence in this ICU follows national trends with as much as 80% of the patients experiencing some level of delirium during their stay (M. Stipsits, Nurse Manager, personal communication, December 2018). Therefore, the recommendation for improvement remains the implementation of a nurse-driven, non-pharmacologic sleep protocol for the purpose of avoiding the development of delirium whenever possible.

Stakeholders
Project stakeholders included the aforementioned pentad, nursing and physician leadership, tele-ICU, ancillary department members, and patients and their family members in the CCU. The medical director, nurse manager, and assistant nurse manager are key stakeholders, alongside front-line nursing staff. Given the interdisciplinary team approach of CCU from physician providers, nursing staff, and supporting ancillary departments (including but not limited to the laboratory, chaplaincy, radiology, physical therapy, occupational therapy, speech therapy, nutrition, hospital education, and wound care consultants) this project was set to become a successful practice improvement initiative. And then a pandemic struck.

Strengths, Weaknesses, Opportunities, and Threat Analysis
A strengths, weaknesses, opportunities, and threats (SWOT) analysis was done to assess and mitigate potential barriers to project success. See Figure 2 for a visual representation of the SWOT. Strengths included the support of upper leadership, CCU teamwork, and the potential for increased patient and family satisfaction. Weaknesses included the newness of the CCU leadership team. The medical director had been on board about 1 year at the time the project gained facility approval, and the nurse manager had been there for a shorter period of time. He is the 11th nurse manager in 8 years. There was a new assistant nurse manager (hired September 2020), and no clinical nurse leader. The CCU educator transferred from a related facility in May 2020. Opportunities included the standardization of delirium care and prevention. Since this protocol was nurse-driven, nursing staff experienced autonomy in recognition and implementation of best practices. Threats included resistance to change, non-compliance, and the incredible ingenuity of nurses to invent "workarounds" as a result. In-servicing was planned to help lessen these weaknesses and threats. The biggest threat was one that could not have been anticipated, which was the COVID-19 global pandemic. This worldwide threat changed nearly everything about critical care delivery.

Figure 2
SWOT Analysis

Mission, Vision, and Objectives
Mission statements convey the purpose of businesses while vision statements describe a project's anticipated goals (Meyer, 2019). The mission statement of this facility is to serve our patients with care, compassion, and commitment (Facility Volunteer Handbook, n.d.). The vision of this project was to use nurses' intervention via a standardized protocol to affect a reduction in incidence of delirium in the CCU over 2 months. Anticipated anecdotal benefits were shortened overall LOS and earlier liberation from mechanical ventilation for patients on life support. At implementation, all CCU nurses understood the sleep enhancement protocol, were able to articulate its benefits, and properly apply it.

Short-and Long-term Objectives of the Project
Short-term goals for this project included obtaining stakeholder support and implementation through training and allowing time for questions and answers, which included reference to similar successful projects. Within 5 days of initial training, there was > 90% use of the protocol. This was verified by a 100% audit per the project manager.
The long-term goal of this project was to reduce delirium in the CCU. Sustained use of the sleep enhancement protocol was evidenced by the adoption of it into standard practice in the CCU. Ongoing clinical training will be accomplished with annual assessment of competence using the CAM-ICU tool. To sustain stakeholder buy-in, quarterly status reports of delirium rates will be provided to the staff so they can continue see the direct results of their efforts.

Risks and Unintended Consequences of the Project
Mitigating pitfalls and barriers is crucial to the success of this or any future project to be implemented in the CCU. The primary barrier to this project was the COVID-19 pandemic and its effect on patient flow, treatment, and outcomes. This hospital experienced a second wave of the pandemic in November 2020 which resulted in the majority of patients being intubated and sedated, with many being placed in the prone position for maximum oxygenation. Proning requires chemical paralyzation, which means that those patients were unable to participate in delirium assessment. This factor alone reduced the sample size markedly.
Nursing staff who completed the pre-implementation training early and demonstrated competence in the use of the assessment tool (CAM-ICU) served as subject matter experts.
These experts helped encourage their peers who either had not done the training or did not initially buy in to the protocol. Data, essential to understanding intervention efficacy, was collected twice weekly, and analyzed weekly. Unintended negative effects were mitigated via regular communication with the CCU staff throughout the length of the project. While it was known that a large data pool would not be realized because of the pandemic, the project manager proceeded with the project and communicated as originally planned.

Project Plan (Method)
For this project, a pre-and post-assessment was done to compare CAM-ICU scores in CCU patients before and after implementing the nurse-initiated sleep enhancement protocol. A chart review using the Computerized Patient Record System (CPRS) was done to compare CAM-ICU scores at baseline (pre-implementation) and post-implementation of the sleep enhancement protocol. Total number of patients with positive CAM-ICU was assessed in retrospective patient populations during the 30 days before implementation and compared to those receiving the nurse-driven protocols over a period of 30 days. The goal was to reduce the incidence of ICU-delirium in that time frame. As evidenced by assessing the number of delirious patients divided by the total number of patients in the CCU on a random day, the initial rate of delirium was 60-70%. Lewin's Change Theory was the framework for this project alongside the JHNEBP model for implementing change.

Project Steps
Leadership in CCU and the Hospital Education Department supported this project to decrease delirium in the unit and see it as an important evidence-based practice change initiative for the organization overall. Implementation of this project began with the planning phase followed by training by inservice for the nursing and provider staff regarding delirium as an effect of sleep fragmentation. Use of the CAM-ICU tool was reinforced using an alreadypublished, system-wide how-to manual embedded in, and tracked by, one of the computer-based training systems of the facility. Staff members were able to download it to their cloud drive for reference. All training was done using the Donna Wright method, which emphasized lateral training with three main components of understanding: a) competencies are collaborative, b) the learner is the central focus of competency evaluation, and c) leaders create a dual-focused (employee, organization) culture of success (Wright, 2005).
Implementation occurred in the following order: 1. Planning for the implementation of an evidence-based, nurse-initiated sleep enhancement protocol was formally approved by the CCU medical director, nurse manager, and DNP preceptor.
2. Staff training was done (utilizing Donna Wright method) on the detriment of delirium, use of the CAM-ICU tool, and sleep enhancement protocol.
3. Dry-run evaluations were completed prior to go-live.
4. Implemented the sleep enhancement protocol in November 2020.
5. Periodic data collection and PDSA cycling were done.
6. Data analysis was conducted using Intellectus, and MS-Excel spreadsheets.
7. Dissemination of project results.
A detailed timeline of the project is included (see Appendix E) and represents the schedule for this quality improvement process.

Budget and Resource Needs
Training with protocolized measures to avoid delirium b). empowerment of nursing staff to use their assessment findings to apply a new protocol c). pride of the nursing staff at trying to reduce delirium and its associated sequelae of negative events that often accompany it.

Team Specifics
The team involved in the work of this project consisted of the project manager, CCU nursing leadership, and the bedside nursing staff. Support staff included the resident and attending physician group, the pharmacy and other ancillary departments (for medication and therapy scheduling purposes). Family members and other surrogates would have been recruited to help minimize the incidence of ICU delirium by engaging them in protecting the sleep of their loved ones, but at the time of this project all visitations were on hold as a result of COVID-19 safety recommendations. Families are largely willing to be involved in helping to prevent delirium (Smithburger et al., 2017) so the fact that they could not be included was disappointing.
For nurses, the non-COVID care they delivered did not change, only the timing of it did. The effort of the team was supported by stakeholders, and the project will be revisited house-wide after the pandemic in hopes of realizing greater benefit for patients at every level of acuity.

Inclusion and Exclusion Criteria
Initial inclusion criteria for participants in this project were as follows: a) 18 years of age, b) conscious, as indicated by any RASS other than -4 or -5, c) non-delirious as indicated by a negative CAM-ICU score, d). able to communicate in English, e) length of CCU stay >24 hours.
Exclusion criteria for participation were: a) age younger than 18 years, (b) exhibition of primary dementia, (c) unconscious as indicated by a RASS score of -4/-5, or (d) refusal to participate.
Ultimately, the chemically paralyzed and proned patients were also excluded. Also, as a result of the pandemic, CCU had to be expanded into the adjoining 8-bed, open-bay Post Anesthesia Care Unit. The only partition between beds were curtains, and there was no buffer for noise or light in that area. Those patients could not be included in the project because there was no way to alter the environment for one without altering it for all.

Data Collection Details
Throughout the project, all patients meeting the criteria for the sleep enhancement protocol implementation were tracked and trended by the project manager/ graduate student. A pre-implementation rate of delirium was determined by calculating the rate for 30 days prior to implementation to provide a baseline measurement. After implementation, for 30 days, the rate was calculated again to ascertain change. The outcome was a slight reduction in delirium for patients in the CCU.
The collection phase answered the measures associated with the PICOT. Outcome measures that were collected include a reduction in positive CAM-ICU scores and, anecdotally, LOS in the CCU. (LOS is calculated by subtracting the date of admission from the date of discharge.)

Types of Measures
For any quality improvement project, measurements validate the project results (Institute for Healthcare Improvement [IHI], 2020). The measures chosen for studying processes and outcomes of the intervention are the CAM-ICU and the RASS, which are valid and reliable delirium assessment tools recommended by the SCCM in its 2013 Pain, Agitation, and Delirium PADI guidelines (see Appendices H and I). The independent variable of this project is the sleep enhancement protocol implementation. The dependent variable is the rate of delirium. Table   2 includes details for each category of measure for the project with the type of statistical test to be used for each measure. This figure is also represented in Appendix F

Outcome Measures
The primary outcome measure for this project was reduced incidence of delirium pre-and post-implementation. Secondary outcome measure was CCU LOS. Average LOS was calculated by dividing the total days of ICU stay for all patients included in the data collection pool by the number of discharges. There was a pre-and post-intervention pool. A smaller number in the post-implementation data pool would have indicated a shorter ICU LOS, but this did not occur.
At staff meetings and daily huddles, the nurse manager and project manager provided updates, opportunities for improvement, and progress related to the initiative.

Process Measures
Process measures are defined as the steps that lead a project to its outcome (IHI, 2020).
The process measure for staff completion of the sleep enhancement protocol inservice was done by assessing the percentage of staff that attended inservices before the roll-out date. The goal was 90%. Proper use of the CAM-ICU was reinforced to staff nurses in the ICU and evaluated using occasional direct observation and random chart review. The Donna Wright model of competency assessment helped to identify champion staff that were able to reinforce and support the use of the protocol during the implementation phase and beyond (Wright, 2005).

Balancing Measures
A balancing measure asks whether changes designed to improve one part of a system causes new problems in other parts of the system (IHI, 2020). The balancing measure for this project was the percentage of the staff that find satisfaction with the use of the sleep enhancement protocol. The project manager assessed staff satisfaction by way of random interviews following the completion of the project. The question asked was "Do you feel as though the Sleep Enhancement Protocol was helpful for our patients?" Staff members were invited to provide specific feedback in an email reply, but it was not required. All of the 25 staff who were interviewed saw value in the protocol and agreed that it will be an easy intervention to provide in the future.

Financial Measures
The

Validity and Reliability
Validity confirms that the data results represent the proposed measure for the project (reduction of delirium) and reliability demonstrates stability and consistency over time (Sylvia & Terhaar, 2018). The project manager, alongside the statistics team, both played an active role in data collection, evaluation, and chart reviews. The RASS (see Appendix I) is a 10-point scale for measuring sedation, with four grades of anxiety or agitation above 0 ("restless" to "combative"), and five levels of sedation below zero ("drowsy" to "unarousable") (Sesser et al., 2002). A score of zero indicates a calm and alert state. A score > -3 means the patient can be CAM-ICU assessed. A score of -4 or -5 means the patient is too sedate to assess for delirium.
This data will be measured both pre-and post-intervention. The CAM-ICU (see Appendix H) is the most commonly used instrument for diagnosing delirium (Devlin et al., 2018;Ely et al., 2002). It takes about 2 minutes to complete and can be used with both verbal and non-verbal patients (Ely et al., 2002). Typically, RASS and CAM-ICU are used together: the former to assess sedation and latter to assess delirium. The CAM-ICU is well-validated and showed highinterrater reliability (k=0.94, p<0.001) (Devlin et al., 2018;Ely et al., 2002). The CAM-ICU tool included four parts, which assess in order: mental status changes, inattention, disorganized thinking, and level of consciousness (Ely et al., 2002). Patients were either positive, negative, or unable to be scored. Deficits in features judgement or attentiveness indicate that the patient was delirious (Ely et al., 2002). These have been measured pre-and post-implementation of the sleep enhancement protocol.

Data Analysis
The data collection tool (see Appendix F) is a descriptive table and provides each data collection point for the project: the date, numeric patient code, admission and discharge dates, LOS, CAM-ICU scores for each shift, RASS score, protocol initiation time, whether there were any interruptions, and a block for extraneous notes.
Data collection originated from CPRS, a reliable and valid EMR utilized by all medical centers in this healthcare system. Descriptive statistics were used to measure the frequency of delirium (Sylvia & Terhaar, 2018). Random chart audits were completed for the sleep enhancement protocol usage among the nursing staff with a goal of ≥ 90% which was the measurement of compliance for the hospital. Pre-implementation data was collected using chart audits during the four weeks prior to starting the protocol. All qualified CCU patient charts were reviewed and CAM-ICU scores were collected daily during the post-implementation phase. Table 2 provides a complete explanation of the measures of frequency, central tendency, and inferential statistics that were utilized.

Project Results
Data were collected by the project manager, according to timeline and plan. It was collected on the attached data collection tool, then shredded after being imported into Excel. A security chip-accessed federal computer was used, which is in an interior, locked office. Data were evaluated in collaboration with a statistician at Intellectus, using their proprietary software.

Chi-square Test of Independence
A Chi-square Test of Independence was conducted to examine whether Time and CAM_ICU were independent. There were 2 levels in Time: Pretest and Posttest. There were 2 levels in CAM_ICU: Yes and No.

Assumptions
The assumption of adequate cell size was assessed, which required all cells to have expected values greater than zero and 80% of cells to have expected values of at least five (McHugh, 2013). All cells had expected values greater than zero, indicating the first condition was met. A total of 50.00% of the cells had expected frequencies of at least five, indicating the second condition was violated. When the assumptions of the Chi-square test are violated, Fisher's exact test can be used to produce more reliable results with small sample sizes. Logit models such as binary logistic regression can be used for larger sample sizes.

Chi-square Results
The results of the Chi-square test were not significant based on an alpha value of 0.05, χ2(1) = 0.95, p = 0.329, suggesting that Time and CAM_ICU could be independent of one another. This implied that the observed frequencies were not significantly different than the expected frequencies. Four out of 10 patients exhibited signs of delirium pre-intervention, and only two of 10 showed signs of delirium post-intervention. While this looks like a reduction of 50%, it is not statistically significant due to the small sample sizes both before and after the sleep enhancement intervention. Table 3 presents the results of the Chi-square test.

Two-Tailed Independent Samples t-Test
A two-tailed independent samples t-test was conducted to examine whether the mean of LOS-in-days was significantly different between the pre-test and post-test categories of time.

Assumptions
Shapiro-Wilk tests were conducted to determine whether LOS-in-days could have been

T-test Results
The result of the two-tailed independent samples t-test was not significant based on an alpha value of 0.05, t(18) = 0.01, p = 0.993, indicating the null hypothesis cannot be rejected.
This finding suggested the mean of LOS-in-days was not significantly different between the pretest and post-test categories of time, and the sleep enhancement protocol did not have any impact on LOS. The results are presented in Table 4. A bar plot of the means is presented in Figure 3.

Two-Tailed Independent Samples t-Test for LOS-in-days by Time
13.05 14.52 13.00 11.93 0.01 .993 0.00 Note. N = 20. Degrees of Freedom for the t-statistic = 18. d represents Cohen's d.

Figure 3
The

Missing Data and Storage
Data regarding the interventions of the project have been thoroughly documented and followed daily through the entirety of the project timeline. When noncompliance with the assessment tools threatened the project, staff members were reminded in person to complete them. All relevant information has been added to the database at twice weekly intervals in the form of an Excel spreadsheet. The need for accurate data and consistency with CAM-ICU assessment and the sleep enhancement protocol implementation was reinforced to the project team members at huddle twice weekly. Data was obtained from the electronic health record and stored on a computer physically located inside an interior office with two locked doors and accessible only with a key card and passcode. The collection tool reflected numeric codes in lieu of actual patient identifiers.

Protection of Human Rights
The project proposal was presented to the University of St. Augustine's EPRC committee for approval prior to institutional review board (IRB) submission via this hospital's affiliated university. Final review by the facility's research and development committee has also verified that no human subject data is involved in the project prior to granting their final approval. Letters of support and approval are included (see Appendix L). Consent was not required for this project because delirium assessment is standard practice in the CCU. No patient harm occurred.

Impact
This evidence-based practice project could not be determined to be successful in terms of reliable reduction of ICU delirium, nor did it show any correlating reduction in LOS. The project has addressed the problem of ICU delirium by applying an evidence-based intervention, and while the data pool was reduced by the Covid pandemic, the protection of sleep has become a standard of care in the ICU as well as the acute care areas of this facility. Implementation occurred according to plan; it was measurement that was hindered by the pandemic, in that there were very few patients who met inclusion criteria. This huge reduction in qualifying patients secondary to Covid-19 posed the biggest limitation to the success of this project. ICU nurses reported satisfaction with the autonomy of applying the protocol, and they conveyed a sense of empowerment in using it. Several nurses reported ideas for other potential projects involving nurse-initiated protocols.
Future implications of this project include the alteration of practice across the facility by the addition of sleep enhancement measures for all patients, not just the critically ill. To further improve the practice problem, sleep enhancement will be included in annual unit education, as well as the onboarding and orientation processes. To sustain this project, chart audits will continue to be done by unit peer leaders. Plans are in place to include "sleep enhancement as a nursing intervention to reduce delirium" in an upcoming journal club discussion for continuing nursing education credit. The ICU nursing staff has been supportive of this project, and fully compliant in implementing the protocol on qualifying patients. The process measure reflecting 90% or better compliance with inservice training was met. Because super-trainers educated their peers, and advocated for sleep enhancement, they reported a deeper interest in the project outcome. One nurse said, "This wasn't another mandate; it wasn't done to us, it was done with us." Another said, "This feels like the most normal thing we have done since the pandemic began." All seem hopeful that a post-pandemic reimplementation will be successful.

Dissemination
Project culmination included a comprehensive review and evaluation of strengths, weaknesses, and opportunities for refining future use of nurse-initiated protocols in the ICU.
Internal dissemination of the project results within the institution happened at several levels.
Initially, results were presented in the unit, to nurses at their shift huddles, and then at their monthly staff meeting. First string attendees included bedside nursing staff, resident physician staff, and the nurse managers. Results were presented to facility leadership remotely using Microsoft Teams. The ICU leadership team included (in ascending order) the assistant nurse manager and manager, the chief nurse of medicine, the medical director, and the facility director. An all-staff email invitation was sent facility-wide so that every interested employee was able to attend the presentation on Teams. Results were also published in the facility newsletter.
The project manager disseminated results at their local AACN chapter monthly meeting.
Even though the project did not demonstrate statistical significance the value of Clinical Significance is realized for planning future projects and the staff recognized its essentialness for patient care to reduce the potential for delirium. Journal submission provides an opportunity to disseminate information validating the benefits of NIPs and to corroborate that evidence-based practices can improve patient outcomes (Wolf, 2015). A full text will be submitted to the SOAR@USA institutional repository as a part of the DNP program completion requirements, which will also enhance the discoverability of this EBP project. Final submission to the Sigma Repository will be completed in an effort to distribute the project material worldwide.

Conclusion
The intention of this paper has been to introduce the rationale behind and the plan for an evidence-based practice change project to reduce delirium by protecting sleep and rest in an ICU.
As indicated in the literature, utilizing a well-organized, evidence-based sleep enhancement protocol should reduce fragmentation of sleep and decrease the incidence of delirium with all its negative sequalae for this patient population. The protocol is simple and includes favorable environmental manipulation, frequent reorientation, and clustered care. There were no earplugs, eye masks, aromatherapy, or medications used in this project, though there are several published articles that advocate for them. This project sought to facilitate a protocol for the consistent use of sleep enhancement interventions in hopes of promoting uninterrupted sleep in ICU patients as evidenced by improved CAM-ICU scores and a decreased incidence of delirium. Unfortunately, the global COVID-19 pandemic erupted which severely restricted the pool of eligible patients and interrupted the expected outcome. Thankfully, the facility plans to implement the Sleep Enhancement Protocol on every floor going forward. They will reassess the effectiveness after the pandemic ends.