Optimizing Sleep for Residents in Long-term Care Without Sedatives

While aging can result in a mild decrease in sleep quality, the causes for impaired sleep can be multifactorial. Contributing factors can include medical comorbidities and medications, behavioral changes such as daytime napping and sedentary lifestyle, and environmental factors such as noise and lack of sunlight exposure.⁹ Environmental factors within facilities can include frequent nursing care interruption and less bright light exposure.^10,11 Daytime bright light exposure is known to improve sleep quality in older adults.¹² Inconsistent bedtime routines and nighttime light and noise compromise sleep quality.^13-15 Patient-related factors such as pain, nocturia, depression, and polypharmacy can also negatively impact sleep in long-term care (LTC).^16-19 

Despite their multiple adverse effects, sedatives continue to be overprescribed for sleep concerns in older adults.²⁰ A systematic review found that, among people older than 60 years, sleep was modestly better with sedative hypnotics, but individuals experienced cognitive and psychomotor adverse events.²¹ Sedative hypnotics, such as benzodiazepines and z-drugs, have also been linked with an increase risk in hip fractures and falls.^22-24 Zolpidem can specifically be associated with complex sleep-related behaviors.²⁵ There is currently insufficient evidence for sleep benefits with diphenhydramine, antidepressants, or atypical antipsychotics despite the ongoing use of these medications for sleep in this population.²⁰ 

As in the community, nonpharmacologic treatments should be first line for poor sleep in LTC. The use of eye masks and ear plugs are simple interventions that have shown promise for easing insomnia in adults.²⁶ Resistance strength training and walking, combined with social activity, significantly improved sleep in nursing home and assisted living residents in one study.²⁷ Exercise for residents in wheelchairs has also been shown to increase sleep duration and efficacy at a 6-month follow-up.²⁸ Another study provided natural light exposure for at least 5 hours, either indoors or outdoors, for 21 days, which improved subjective sleep in LTC residents.²⁹ However, frailty can make these activities more difficult to implement among LTC residents.

Cognitive behavioral therapy (CBT) is just as effective as pharmacologic intervention for insomnia in older adults.³⁰ Buysse et al explored brief behavioral interventions targeted at good sleep hygiene. These interventions resulted in improved mood and anxiety ratings, insomnia remission, and improved sleep quality for residents.³¹ However, CBT and brief behavioral interventions are not easily implemented in LTC, given the prevalence of dementia and other practical and financial barriers.

Studies have explored strategies to decrease sedative-hypnotic use for sleep in group facilities. Ouslander et al conducted a clinical trial involving 160 residents aged >65 years from community nursing homes. Trained research staff provided the intervention, which included daytime physical activity to keep individuals out of bed, evening bright light exposure, consistent bedtime routines, nighttime routines with limited sleep disruption, and nighttime noise-reduction strategies. Unfortunately, no sleep improvement, as measured by actigraphy, was found in the intervention group.³² Although this study aimed to implement a sleep improvement strategy, the same strategy was implemented for all residents irrespective of individual resident needs. Alternatively, McCurry et al implemented a resident-centered sleep education program designed to teach caregiver staff about nonpharmacologic strategies to improve sleep in older adults with dementia living in adult family homes.³³ Adult family homes typically house two to six residents with a caregiver providing 24-hour supervision in a more home-like environment than LTC facilities. Over 6 months of follow-up, the study found that residents in the treatment group had longer sleep time, less daytime sleeping, fewer nighttime behavioral problems, and less depression. While the LTC setting is different than an adult family home, this study demonstrates that educating caregivers can provide a resident-centered, feasible approach to improving sleep. 

Given the variable potential of nonpharmacologic options and the known adverse effects of sedatives for insomnia in older adults living in LTC, our objective was to evaluate the impact of an evidence-informed, resident-centered staff educational intervention on nighttime sedative hypnotic use and on sleep patterns among LTC residents.

Setting and Participants

A randomized stepped-wedge controlled trial was conducted on five units in one LTC facility in Calgary, Alberta, Canada.³⁴ This LTC facility has 226 resident beds. The majority are government funded, with only a few private beds. There are approximately 25 residents per unit. Units are either regular LTC units or LTC advanced-dementia units, and overall about 80% of the residents have cognitive impairment. About three-quarters of residents are women (166), and the average age is 87.08 (83.95 for men, 88.22 for women).

This article did not require institutional review board approval. This was a quality improvement project evaluating the implementation of evidence-informed sleep optimization strategies into current practice within the participating LTC facility. It scored 2 (low-risk score) on the Alberta Research Ethics Community Consensus Initiative (ARECCI) Screening tool and, therefore, formal ethic approval was not sought.³⁵

Study Design

The intervention was a multifaceted educational intervention that involved: (1) the development of opinion leaders among physicians and nursing staff (health care aides, licensed practical nurses, and registered nurses); (2) brief, in-person, frontline nursing education and clinical-decision support (Supplementary Box 1); (3) a written letter to physicians explaining the project (Supplementary Box 2); and (4) implementation of sleep logs (Supplementary Box 3). 

The intervention was rolled out sequentially at 1-month intervals on each unit, in turn, by random assignment, for an overall study period of 5 months. An investigator not familiar with the involved units randomly assigned the order of intervention rollout. The first month acted as baseline for all units. By the end of the study, all units had received the intervention. Two of the units acted as a single unit for the randomization process and implemented the intervention at the same time. Nurses on all five units were asked to complete sleep logs for all 5 months. This study was not blinded. 

The in-person, brief (15-20 minutes), in-service meetings were conducted within the first week of implementation on a unit and provided information on reasons behind poor sleep in LTC residents and possible practical solutions to improve sleep. Nursing staff on the units who were unable to attend the in-service were offered one-on-one education with a nurse educator. As part of the in-service, staff were educated on how to further use the sleep logs as a visual guide for conducting resident-centered insomnia evaluations and developing individualized nonpharmacologic sleep management plans. Specifically, part of the intervention aimed to empower nursing staff involved in direct resident care to chart resident sleeping patterns, brainstorm possible reasons for poor sleep, and record nonpharmacologic interventions used. Regular check-ins were conducted by one of the study authors working at the facility to ensure sleep logs were completed, answer questions, and informally collect nursing feedback.

Outcomes and Data Collection

The primary outcome was the monthly change in sedative hypnotic use, both scheduled and as needed (PRN), for insomnia. The secondary outcomes, using data from the sleep logs, included exploration of changes in the daily mean number of hours slept per month, the mean number of nighttime (8 pm to 7 am) hours slept per month, and the mean number of daytime (7 am to 8 pm) hours slept per month. Three (24 hours each) sleep logs were collected per resident per month. Residents were recorded as awake or asleep on q2 hour nursing rounds by visual inspection. Number of hours of daytime sleep was logged from 0700h to 2000h, while number of hours of nighttime sleep were logged from 2000h to 0700h. We performed the analysis on the average daytime and nighttime sleep across the three sleeps logs per resident.

Data Analysis

The data was first explored with descriptive statistics, including means, medians, and standard deviations. McNemar’s test was used to evaluate the change in the number of patients with prescriptions between baseline and end of study. Mann-Whitney U tests were calculated to determine the effect of intervention on the average monthly number of tablets across 5 months and all units, and also to compare the differences in average hours slept as recorded in the sleep logs. To test the effect of intervention, while adjusting for unit and month given the stepped-wedge study design, generalized estimating equations with a normal distribution were conducted for average sleep logs, and with generalized estimating equations with a Poisson distribution for tablet count data.

 There were 20 residents from five units with sedative hypnotic prescriptions, ranging between three and five residents (12%-20%) with prescriptions per unit. Three residents had prescriptions for nighttime sedative hypnotics discontinued prior to day 1 of our month 0 baseline cycle and were therefore excluded from analysis. Over the 5-month study period, we observed 41 medication administration periods during pre-intervention and 44 during the intervention period among the 17 residents (Table 1).

Among the 17 residents with prescriptions at baseline (month 0), 14 had prescriptions for scheduled doses of sedatives, and four residents had prescriptions for PRN doses of sedatives. At the end of the study, only 12 residents still had prescriptions for sedatives—10 for scheduled doses, and three for PRN doses. (One resident had both a scheduled and as needed medication; we chose to break these up separately in measuring our primary outcome of monthly changes in sedative hypnotic prescriptions.) The reduction in prescription for scheduled doses was 29% (4/14; P=.125), and the reduction in PRN prescriptions was 25% (1/4; P=1.00) (Table 2).

A total of 1126 tablets were given as scheduled doses, and 40 were PRN during pre-intervention, for an average of 33.12 and 4.44 tablets per patient, respectively. After intervention, there were a total of 797 tablets given as scheduled doses and eight were as PRN, for an average of 27.48 and 1.33 tablets per patient, respectively. The differences between the average number of tables between pre- and post-intervention periods for both scheduled and PRN doses was not statistically significant, even when adjusted for unit and month (Table 1).

The average daytime hours slept decreased as the intervention was rolled out (Table 1). Once units received the intervention, there was a significant drop in the average hours slept during the day (1.41 vs 2.68 hours; P=.049), which was sustained through the end of the study period (Figure 1). Conversely, the average nighttime hours slept significantly increased over time (8.96 vs 7.91 hours; P=.049) (Figure 1). The average time slept over a 24-hour period did not change dramatically (10.57 vs 10.37 hours), although the effect of intervention was found to be statistically significant (P<.001). 

This project sought to evaluate an evidence-informed, educational intervention aimed at reducing use of nighttime sedative hypnotics among residents in an LTC facility. Sedatives have been associated with more harms than benefits in older adults, and those living in LTC are particularly at risk for adverse outcomes given the high prevalence of frailty and complex comorbidities.²⁰ This educational intervention, which emphasized a nonpharmacologic resident-centered approach to care, was successful at shifting sleep patterns to more nighttime hours spent asleep. It also decreased hypnotic sedative prescriptions, although not statistically significantly. This study builds on earlier literature that found nonpharmacologic, resident-centered interventions targeted at caregivers to be effective.³³ However, this is the first study that we are aware of to explore this type of intervention within a large LTC facility.  

This educational intervention was relatively simple, using available nurse educators to conduct brief in-servicing that was further supported by opinion leaders, simple clinical-decision support, and by using sleep logs as a method of developing and auditing individualized sleep strategies. This pragmatic intervention could be easily spread to other LTC facilities. 

Comorbid medical and psychiatric illnesses, polypharmacy, and pain can all negatively impact sleep and are common among LTC residents.⁹ Environmental factors can also interfere with sleep and need to be considered in facility living. Instead of using sedating medication, a resident-centered behavioral approach to managing sleep can empower health care providers to consider and optimize factors that might be impacting a resident’s sleep. Implementation of individualized behavioral strategies can not only decrease the overuse of sedatives, but it can also decrease excessive daytime napping and nighttime behaviors.³³

A risk to quality improvement interventions like ours that focus on getting physicians to reduce sedative prescriptions is that they can potentially substitute with other classes of medication, which carry their own risks. Therefore, effort needs to be taken to ensure the entire health care team is aware of the intervention focus of prioritizing behavioral sleep strategies over pharmacologic solutions.

Limitations

This study was limited to one LTC facility, and therefore the results may not be generalizable to facilities with different resident populations, staffing ratios, or workplace cultures. However, the intervention used was informed by evidence, which makes it more likely to be successfully implemented in other facilities. The number of residents on sedative prescriptions at the start of the study was small, which may have impacted statistical power.

A simple, resident-centered educational intervention showed a trend toward reduced sedative use and significantly improved sleep patterns among LTC residents. Given the high prevalence of sleep issues among frail older adults, nonpharmacologic interventions should be systematically incorporated into LTC. 

Supplementary Materials

Authors:
Adina M Constantinescu, MD, CCFP, COE¹• Jadden R Warness, MD ²• Navjot Virk, RN, MN ³• Grace Perez, MSc⁴ • MaryJane Shankel, MD, CCFP, COE⁵ • Jayna Holroyd-Leduc, MD, FRCPC⁶

Affiliations:
¹Calgary West Central Primary Care Network, Calgary AB
² Ross University School of Medicine, Calgary AB
³ Brenda Stafford Foundation, Calgary AB
⁴ Department of Family Medicine, University of Calgary, Calgary AB
⁵ Clinical Assistant Professor, Department of Family Medicine University of Calgary; and Medical Director, Brenda Strafford Foundation, Calgary AB
⁶ Head and Professor, Division of Geriatric Medicine, Departments of Medicine and Community Health Sciences, University of Calgary; and Medical Director, Calgary zone Specialized Geriatric Services, Alberta Health Services, Calgary AB

Disclosures:
N. Virk and MJ Shankel work for the Brenda Strafford Foundation, who operates the LTC facility where this study took place. J Holroyd-Leduc is the University of Calgary Brenda Strafford Foundation Chair in Geriatric Medicine. The Brenda Strafford Foundation provided in-kind support for the project but had no role (outside of employing N Virk and MJ Shankel) in the data analysis or preparation of the manuscript.

Address correspondence to:
Adina Constantinescu
Email: amconsta@ucalgary.ca