Implementation Outcomes for the SLUMBER Sleep Improvement Program in Long-Term Care

Abstract

Objectives:

To describe the implementation of a mentored staff-delivered sleep program in nursing facilities.

Design:

Modified stepped-wedge unit-level intervention

Setting and Participants:

This program was implemented in two New York City nursing facilities, with partial implementation (due to COVID-19) in a third facility.

Methods:

Expert mentors provided staff webinars, in-person workshops and weekly sleep pearls via text messaging. We used the integrated Promoting Action on Research Implementation in Health Services (i-PARiHS) framework as a post-hoc approach to describe key elements of the SLUMBER implementation. We measured staff participation in unit-level procedures and noted their commentary during unit workshops.

Results:

We completed SLUMBER within five units across two facilities and held fifteen leadership meetings prior to, and during program implementation. Sessions on each unit included three virtual webinar presentations and four in-person workshops for each nursing shift, held over a period of 3–4 months. Staff attendance averaged >3 sessions per individual staff member. Approximately 65% of staff present on each unit participated at any given session. Text messaging was useful for engagement, educational reinforcement and encouraging attendance. We elevated staff as experts in the care of their residents as a strategy for staff engagement and behavior change and solicited challenging cases from staff during workshops to provide strategies to address resident behavior and encourage adoption when successful.

Conclusions and Implications:

Engaging staff, leadership, residents, and family of nursing facilities in implementing a multicomponent sleep quality improvement program is feasible for improving nursing facilities’ sleep environment. The program required gaining trust at multiple levels through presence and empathy, and reinforcement mechanisms (primarily text messages). To improve scalability, SLUMBER could evolve from an interdisciplinary investigator-based approach to internal coaches in a train-the-trainer model to effectively and sustainably implement this program to improve sleep quality for facility residents.

Brief summary

We describe a modified stepped-wedge unit-level nursing facility intervention to improve resident sleep. Environmental noise and light feedback and staff input improved the sleep environment and health.

INTRODUCTION

Many negative health conditions are attributable to poor sleep,^1–8 and rates of poor sleep among nursing facility residents exceed similarly aged community-living adults.^9,10 Contributors include sleep disorders,^{11, 12} pain,^13,14 depression,^15,16 anxiety,^17,18 medications,^19,20 and circadian (24-hour) rhythm disruption.²¹ Environmental factors and staff practices negatively impact sleep, including nighttime noise, nighttime light, and limited daytime bright light exposure.²² Behavioral factors include reduced physical activity, social activities and social engagement as well as daytime napping. The additive effect of environmental and behavioral factors exacerbates the severity and impact of sleep disturbances. Importantly, facility characteristics and nursing practices that drive daytime and nighttime routines can be targeted for quality improvement.^23,24

Facility-level approaches to reduce daytime sleeping can improve nightime sleep. This includes increasing daylight exposure, daytime physical activity, mentally stimulating activities and social engagement.^25–27 Modifications to improve nighttime sleep environments, in addition to reducing disruptions from staff-provided care, include reducing environmental light and noise.^22,28 Components of cognitive behavioral therapy for insomnia (CBT-I)^29-32––stimulus control, sleep hygiene, sleep restriction, relaxation, and education––can also improve sleep. Researcher-implemented interventions, such as increased light exposure, resident activities and socialization, have successfully decreased daytime sleeping;^10,33 however, impacting nighttime environments has been challenging and may require sustained changes in staff-directed care. Staff-level adoption and integration into routine workflows may make interventions more sustainable but likely require repeated reinforcement.

We implemented SLUMBER (Improving Sleep Using Mentored Behavioral and Environmental Restructuring), a mentored staff-education and empowerment program, in three nursing facilities in New York City to test its impacts on sleep, depression, anxiety, physical activity and cognitive function. We specify resident-level measurement procedures and outcomes elsewhere.³⁴ Here, we focus on strategies, rationale and delivery––implementation of SLUMBER––a complex and multi-model educational program provided to nursing facility staff, to improve sleep in nursing homes. Although health behavior research benefits greatly from methods to monitor fidelity to intended interventions,^35,36 we reframed resident care and provided sleep-related behaviors and tools for improving the sleep environment using strategies for staff empowerment. We did not measure fidelity of strategies individual staff may have used with individual residents. In a separate outcomes paper,³⁷ we describe the subset of residents who consented to participate in sleep-related data collection (surveys and actigraphy watches). In this paper we describe the SLUMBER implementation as a roadmap for others as strategies for improving the quality of facility care.

METHODS

Conceptual Framework:

We use the integrated Promoting Action on Research Implementation in Health Services (i-PARiHS)³⁸ framework as a post-hoc approach to highlight key elements of the implementation and to identify insights and interrelationships between four main constructs of this framework (Innovation, Recipients, Context, Facilitation). The i-PARiHS model (Table 1) specifies facilitation should enable and reinforce recipients within their particular context to adopt and use evidence-based practice (EBP). SLUMBER utilized blended facilitation, an implementation strategy providing context-specific coaching, problem-solving, and technical assistance required to implement EBP.^39,40 We applied blended facilitation to four SLUMBER program components: 1) leadership buy-in/support; 2) staff/research team relationships; 3) accessible education/training; and 4) staff motivation for action. Our methods combined intentional planned strategies and iterative in-the-moment responsive strategies. We report resident-level outcomes of SLUMBER in Martin et al.³⁷

Table 1:

Implementation Strategies and their Associated iPARiHS Constructs³⁸

Implementation Strategy	iPARiHS Construct
Blended Facilitation	Facilitation (to align context, recipients and innovations)
Informatics support	Context (systems)
Staff training and academic detailing	Recipients (knowledge, skills, goals, acceptability, motivation)
Audit and feedback	Recipients (motivation, goals)

Implementation Overview:

Successful implementation required close attention to: 1) clear messaging, highlighting benefits to all stakeholders (residents, families, staff, and leadership); 2) identifying stakeholders and deploying effective engagement strategies; 3) forming trusting stakeholder relationships, especially staff, to encourage adoption of proposed work strategies; and 4) providing effective, actionable, and sustainable work strategies through education and reinforcement.

Clear Messaging:

We developed brochures and platform presentations for leadership meetings. We repeated facility leadership presentations as listener feedback suggested challenges in understanding study procedures and leaders’ schedules required multiple meetings to include all critical participants. For greatest clarity, accessible to lay audiences, and a reliably reproducible presentation, we created a 3-minute animated video (see YouTube site: https://www.youtube.com/@slumberstudy2018). We later revised this video with exemplars of facility staff-described success to ground the SLUMBER approach.

To implement facility program changes and a research study requiring resident consent for obtaining data, family awareness and buy-in was critical for adoption. Clear and consistent family messaging occurred through directed mailings and community meetings, while offering opportunities for facility ownership of SLUMBER as a quality improvement activity. We provided draft letters and handouts for each facility to revise and place on their letterhead with leaders’ signatures for clear study support and ownership. We created the name “SLUMBER” and its logo to be memorable, self-evident and to engender trust, believing this was as important for residents and families as it was for staff.

Leadership Engagement:

We identified three not-for-profit New York City nursing facilities based on prior professional relationships, which informed understanding facility culture and receptiveness to research engagement and quality improvement. We met with facility leadership, explained planned interventions and highlighted anticipated benefits including improvements in facility care quality, academic partnership, and enhanced institutional reputation. We used these meetings to corroborate and gather additional information on facility structure (number of units, unit size), resident characteristics (age range, length of stay, and likely capacity to consent to participate), stability of unit staffing, ability to accommodate unit-level sessions with our team, willingness to install technology, support for Wi-Fi connectivity, and willingness to provide MDS 3.0 data for consenting residents.

We asked facility leadership to identify two units in their facility having at least 40 residents per unit of whom at least 20 would be likely to be able to consent and participate in actigraphy and surveys. However, all unit residents were potential recipients of staff-delivered strategies for sleep quality regardless of whether they consented to participate in sleep quality evaluations and related impacts. (Please see Supplementary Table 1 for resident demographic description and a separate outcomes manuscript³⁷ for details.) While we excluded dementia-designated units, we recognized that cognitive impairment was prevalent in non-dementia units; however, we anticipated many residents with cognitive impairment would retain capacity to consent and participate in research.

Relationship Building:

Leadership meetings were also used to establish trust in our team and to forge sustainable relationships. We offered accessibility to our team whenever requested and planned meetings for updates and study findings, conveyed our shared goal of improving facility quality, and provided assurances that we were visitors who would always approach our entry as a privilege founded in: sensitivity to and avoidance of any disruption in care delivery; understanding the unpredictable nature of nursing care; and confidentiality.

Webinar and workshop content development:

Content was based on prior multi-component efforts in facilities⁴¹ and other evidence-based strategies for sleep improvement in institutional environments organized by three learning domains:

1. Improve nighttime sleep environments.

2. Increase resident daytime activity levels and light exposure, reduce daytime sleeping.

3. Identify and assist individual residents with sleep difficulties.

To reinforce learning, we developed 15-minute didactic webinars interspersed between in-person mentorship workshops. This strategy reinforced content, allowed for customization for each facility, unit, shift and situation, and over time, encouraged staff to recognize the importance of care coordination across nursing shifts.

Development team and steps:

As described elsewhere³⁴ and in Table 2, a team of investigators with expertise in sleep, geriatrics, psychology/behavioral health, nursing, implementation science, nursing education and facility care met weekly and drafted educational materials iteratively. During mentoring workshop sessions, mentors led discussions about how staff actions on each shift impacted residents on subsequent shifts (e.g., helping residents stay awake during the day would help them sleep better at night, and helping residents sleep better at night would help them engage with daytime visitors). We encouraged between-shift hand-offs that included information about residents’ sleep, similar to vital signs, medications, and planned treatments.

Table 2:

Outline of education and strategies delivered to each nursing shift (day, evening, night) on each participating unit.

Pre-implementation: Study team/leadership meetings to identify participating units, develop plan for each webinar and mentoring sessions, and identify potential champions. Facility notifies families of residents on participating units about the study
Week 1: Webinar 1: What is SLUMBER? 1. Introduction of mentors 2. Basics of how sleep works 3. Why good sleep at night is important for residents 4. Benefits to facility and staff of healthy resident sleep 5. Assignment for this week: Notice sources of noise (download free decibel meter app)	Week 2: Workshop 1 1. Introduction to mentors/their roles 2. Ground rules and confidentiality 3. SLUMBER goals 4. Answer questions about webinar 1 5. Discuss sources of nighttime noise (evening, night shifts) 6. Discuss/Review common sleep issues unit residents
Week 5: Webinar 2: What happens during the day? 1. How and when do residents sleep? 2. Time in bed asleep during the day 3. Behavioral activation/social engagement 4. Light exposure 5. Assignment for this week: Observe daytime sleeping behavior.	Week 6: Workshop 2 1. Staff feedback/questions on webinar 2 & text messages 2. Tips to help staff manage their own sleep 3. Brainstorm meaningful activities (day, evening shifts) 4. How to identify resident preferences 5. Introduce Noise Monitoring Apps
Week 9: Webinar 3: How to help residents who cannot sleep 1. Individual resident factors   a. Sleep disorders   b. Medical problems and medications   c. Feel pain or discomfort at night   d. Feeling upset, stressed or depressed 2. Assignment: Identify specific residents to discuss with mentors.	Week 10: Workshop 3 1. Staff feedback/questions on webinar 3 & text messages 2. Dealing with difficult residents (behavioral challenges and poor sleep), including “Switch and Back off” method 3. Increasing activities for residents; having residents choose and conduct their own activities
	Week 13: Workshop 4 1. Feedback related to SLUMBER program 2. Discuss any changes observed on the unit 3. Discuss helpful tools and strategies and how to sustain over time
Post-implementation: Leadership close-out meetings, ongoing text messages “tips and reminders”

Notes: Webinars presented live by JLM and recorded. Recordings made available to all unit staff who opted in for text messaging. Workshops were attended by full mentoring team and unit staff. The intervention team presented each workshop and webinar three times: day, evening and night shifts.

The development team engaged in the following steps:

Step 1: Incorporation of findings from prior interventions:

Our team identified proven strategies from prior research in long term care.^41,42 We incorporated learning about increased daytime light exposure to improve circadian rhythms and facilitate better sleep. For daytime activities, we supported staff to encourage residents to eat meals out of bed, participate in scheduled activities and converse with residents while providing care. Moreover, we provided educational content about sleep apnea and other symptoms highlighting the need for medical intervention when indicated.⁴³

Step 2: Incorporation of other common-sense approaches:

We incorporated attention to other potential causes of nighttime behavior disturbances such as sleep interruption during bedside care and the need to build trust to reduce nighttime anxiety. In addition, nursing staff suggested we develop resources for families. We created a brief “pocket card” with tips about healthy sleep that were included in webinars and mentoring session content. These were provided to families of enrolled residents.

Step 3: Implementation:

We implemented three webinars for each nursing shift, which were recorded for wide dissemination to individuals who were unable to attend in-person sessions. We conducted four workshops for day, evening and night shift units, adjusting the agenda for each workshop based on questions or requests from staff.

Strategies for Engagement:

During workshops, we presented ourselves as “facilitators” rather than experts. We stated, “We are here to help.” We elevated staff as experts in the care of their residents, recognizing that providing educational content was necessary but insufficient for changing staff behavior. We connected consequences of poor sleep with benefits gained from better sleep but recognized this was less likely to be absorbed without reinforcing strategies. During workshops, we intentionally solicited challenging cases from staff to provide strategies that might provide a notable difference in resident behavior and encourage adoption when successful. Attentiveness to staff challenges was an engagement strategy recognizing that staff appreciated when someone listened and were then more likely to be engaged. We kept notes of staff comments to revisit at subsequent workshops. We (DH, JC, JM, AB, MC) reviewed these notes after each workshop to identify key themes for reinforcement strategies.

Because of competing demands for staff time and turn-over, we anticipated the need for frequent retraining and unit champions to help engage new staff. Engagement strategies also included: 1) use of an interdisciplinary sleep team, including nursing, psychology and medicine; 2) in-person meetings on each unit during every nursing shift (day, evening, night) and providing beverages, snacks and tokens with the study logo such as canvas bags and penlights; 3) flexibility in meeting time and duration dependent on staff availability; 4) pragmatic use of technology and data; and 5) direct access to educational materials and mentors between sessions.

Reinforcement:

We provided regular text messages (a strategy suggested by staff during an early workshop meeting) as educational reinforcement and as reminders about upcoming meetings and webinars. We sent weekly text messages as “sleep pearls”––short, focused communications about promoting sleep and daytime activity, “Sleep in the News”––based on new, relevant research findings, and sleep-focused self-care messages for staff. Text messages also included a link to recorded webinars. We reinforced educational content with pocket cards containing key sleep promotion strategies.

Environmental monitoring and feedback:

We measured noise to characterize the sleep environment and to provide an important window into nursing care practices. We custom-developed decibel sound meters and transmitted data to a central server using wi-fi. We placed these meters on each intervention unit for 24-hour monitoring near resident rooms and at nurses’ stations where greatest noise would be captured, while not obstructing traffic or presenting potential hazards. We converted noise data into an accessible and easily understood graphical format. We were explicit with staff that our intervention target was to reduce the amount of time decibel (dB) levels exceed 60 dB during nighttime hours (10pm to 6am), representing a noise level likely to awaken residents and prevent falling asleep. To inspire staff to reduce noise, we provided staff with noise data from one unit that demonstrated hoped-for reductions over time. We placed computers on nursing stations where staff could access noise data and training materials. We also facilitated staff in downloading a noise meter app onto their smartphones and used momentary data from their own unit.

RESULTS

We initiated the SLUMBER study at three facilities between March 22, 2018 and March 14, 2020. The study was terminated early due to restrictions during the COVID-19 public health emergency and associated regulations that prohibited the presence of the study team in facilities. This impacted implementation in the third facility. Over a 24-month period we held fifteen leadership meetings and conducted six environmental scans to determine optimal placement for decibel meters. Leadership meetings included the corporate chief operating officer (the first two facilities were under one corporation), medical directors, nursing leaders, recreational directors, and facilities management. Some meetings included the chief of social work and a psychologist. Meeting content focused on presenting the study rationale, identifying optimal units for intervention, and a review of all study procedures including subject consent. During later meetings we provided noise data and specific nursing care issues that impacted activities for daytime wakefulness and environmental issues that impacted nighttime sleep (e.g., operating hallway vacuum cleaners, excessive lighting, compressor noise from old appliances, and the need for increased daytime activities). Facility leadership noted opportunities to market their sleep improvement efforts for reputation enhancement.

At one facility, a 3^rd unit was added to increase the total resident sample at that facility. In the third facility, we ended our engagement prematurely due to COVID-19 restrictions. For each participating unit, three virtual webinar presentations and four in-person workshop discussions were held physically on that unit for day, evening and night shifts (see Table 2). Live webinars were recorded and made available to staff via text message as well. A total of 293 individual staff members participated in one or more activities across all three facilities (245 in the two fully engaged facilities), including 193 CNAs, 62 licensed professional nurses and/or registered nurses, 10 social workers and 28 other staff members (Table 3). Although some staff indicated watching recordings later, we were unable to quantify this level of participation. Session attendance was similar between evenings (3pm-11pm) and nights (11pm-7am) shifts (mean attendance across all sessions: 5.0 for both) with lower uptake on days (7am-3pm) shifts (mean: 3.8). Exact staffing denominators were unobtainable due to split shifts, cross covering and intermittent unit assignments. We noted two or fewer staff on any unit did not attend all or some portion of seminars and webinars due to demands for resident supervision or care. We therefore estimate ≥ 65% (3 to 6 out of 4 to 6 unit staff, respectively) attended sessions for some amount of session time across webinars and workshops.

Table 3:

Staff participation (# attendees) at in in-person SLUMBER sessions by shift, unit and activity at participating facilities. (Note that non-unit personnel––social workers, therapists, administrative staff–– often attended webinars as well.)

	Facility 1		Facility 2			Facility 3	***Mean # Attendees / Session type
	Unit A1	Unit A2	Unit B1	Unit B2	Unit B3	Unit C1
	N	N	N	N	N	N	N
	Day shift (7AM - 3PM)
Webinar 1	6	4	5	6	*	6	5
Workshop 1	4	2	3	4	6	6	4
Webinar 2	4	4	5	4	4	6	5
Workshop 2	3	2	5	3	2	6	4
Webinar 3	4	4	4	2	3	6	4
Workshop 3	6	3	4	2	6	**	4
Workshop 4	6	6	6	3	5	**	5
	Evening shift (3PM - 11PM)
Webinar 1	6	6	6	5	5	6	6
Workshop 1	8	6	5	3	4	6	5
Webinar 2	4	6	5	2	3	3	4
Workshop 2	3	6	*	4	5	4	4
Webinar 3	6	6	3	2	4	4	4
Workshop 3	7	6	6	5	1	**	5
Workshop 4	7	3	4	3	4	**	4
	Night shift (11PM - 7AM)
Webinar 1	4	5	4	4	2	3	4
Workshop 1	5	3	3	4	4	5	4
Webinar 2	4	4	3	3	2	4	3
Workshop 2	5	3	3	3	2	3	3
Webinar 3	4	4	3	2	3	5	4
Workshop 3	3	2	3	3	3	**	3
Workshop 4	4	3	4	4	3	**	4

^*Session cancelled due to unit-issues.

^**Cancelled due to COVID-19 shutdown

^***Mean total attendance per session 4 ±SD 1

In workshops, we witnessed opportunities to reinforce webinar-derived educational material using staff-presented challenges with specific resident behaviors. We were able to develop and sustain researcher-staff relationships through proposed strategies for greater feasibility and relevance incorporating staff suggestions. We learned that staff did not access our computers, commenting they never sat at nurses’ stations and suggested we use text messages for reinforcement.

Environmental scans found qualitative nighttime environment changes on four of five units, although we lacked objective measures for some phenomena, we noted and staff agreed that the hallway lights were dimmed from being previously bright, there were fewer alarm bells and nighttime noise levels had decreased. Staff expressed: “The environment is quieter at night.” This impression was corroborated by measured decibel levels (Figure 1) gathered on one of the intervention units. Due to WIFI instability, we achieved complete data capture on only this one unit. We observed considerable reductions in percent of time that noise levels were greater than 60 decibels (equivalent to normal conversation) during hours between 11 PM and 7 AM. These reductions continuously progressed over time.

Figure 1:

Noise Data from September 2018 to January 2019 in Facility 1 Unit.

Workshop attendees provided details of their work experiences, offered solutions to sleep-related problems and gave us useful feedback. Staff acknowledged the utility of webinars and workshops: “We have heard similar strategies before, but with your [SLUMBER] program, we have learned to approach residents better.” They also expressed appreciation for mentoring: “I like that you [mentors] are really involved and what you are doing to help is great because it works.” In Figure 2, we present the most frequently used words and themes to reflect workshop discussions.

Figure 2.

Concept cloud and themes for the most frequently mentioned words/phrases during the workshops with unit staff.

DISCUSSION

We successfully implemented this multicomponent intervention, due to an intentional process working with families, residents, leadership and staff. Staff were highly engaged and shared that they utilized suggested sleep strategies for their residents. The intervention showed reductions in nighttime noise through objective measurement and qualitative improvements based on our environmental scans.

While development of SLUMBER did not engage residents, caregivers or staff, facility leadership participated in pre-implementation meetings. The implementation also evolved as a participatory strategy over time. It is unclear whether we might have arrived at similar strategies if staff were involved prior to launch but we suspect that our ability to modify content and logistics of program delivery will always yield somewhat different approaches due to nursing facility heterogeneity, the residents who inhabit them, staff who they employ and the care they deliver.

There were several limitations to our implementation and this study. First, we recognize that SLUMBER, including use of in-person experts, was geared towards maximizing effectiveness as a proof of concept rather than scalability. Our team’s intentional presence on all three shifts engendered trust and program engagement but raises questions about scalability. Future iterations will need to be modified for real-world spread, including use of facility champions and train-the-trainer approaches as implemented in other long-term care studies.^44,45 In this trial, we established an intervention possibly maintained without on-site experts; however, true sustainability requires long-term reinforcement, not addressed here. However, this program may be inherently self-reinforcing through immediate feedback when residents have better sleep (fewer nighttime disruptions and need for care) and are more awake, and subsequently more engaged and participatory during the day. Second, while we collected data on training and mentoring participation, but were unable to ascertain quantitatively, what interventions were implemented by staff or how frequently, despite qualitative signs of uptake. Third, while these facilities have different physical environment and resident characteristics, two are part of a single operation in a single city, with relatively stable unionized staff, thus limiting generalizability. A third non-profit facility from a different organization only received the initial implementation. We know that pragmatic implementation is challenging, particularly in facilities with disadvantaged populations;⁴⁶ thus future studies will require adapting lessons learned from this study to address the needs of other facilities.

CONCLUSIONS and IMPLICATIONS

Overall, this study showed that engaging staff, leadership, residents and family of nursing facilities in implementing a multicomponent sleep quality improvement intervention is feasible and impactful. This requires gaining trust through presence and empathy and through use of reinforcement (primarily text messages). Future work is required to scale the SLUMBER intervention, transitioning from an interdisciplinary investigator-based approach to internal coaches in a train-the-trainer model.