Sleep disturbance, among older adults, often co-exists with other symptoms and can exacerbate their condition. Several factors, including residents' physical and emotional symptoms, and environmental characteristics of the long-term care facilities (LTCFs), can contribute to sleep disturbance [1]. Physical factors include physical disability, cognitive impairment, chronic disease [1, 2] and decreased activities of daily living (ADL)[3], pain and fatigue, physical activity, and poor self-management [3, 4]. Psychological factors include depression, anxiety, loneliness, worry, and other emotional disturbances [5]. Environmental factors, such as a decrease in physical activity, restriction of social interaction, reduced exposure to sunlight, poor sleep hygiene [2], noise at night, and night lighting, can also interfere with sleep onset and maintenance [3, 6]. Thus, older adults at LTCFs are physically, mentally, and environmentally vulnerable groups owing to their high dependence on ADLs, complex diseases, multi-drug administration, and restrictions on social participation [7, 8]. Therefore, they have more sleep disturbances than community-dwelling older adults owing to the combination of said factors [9]. Studies report that more than one-third of older residents in LTCFs have a sleep disorder [10], and approximately 72% of them have poor sleep quality [11]. The risk of sleep disorders among older residents in LTCFs is higher than that among community-dwelling older adults because of the poor environment and physical factors that affect sleep hygiene.

Multiple studies have reported the negative effects of poor sleep on residents in LTCFs. For instance, sleep disturbances such as breathing difficulties, nightmares, poor sleep quality, and difficulty concentrating during sleep can cause depression in LTCF residents [12, 13]. Furthermore, persistent sleep deprivation can lead to cognitive impairment [14, 15], which is a major factor in re-admission to LTCFs [16]. Changes in sleep patterns can adversely affect neurological health by increasing the risk of dementia [17], leading to decreased functional recovery, social separation, increased risk of falls, and even death [18]. Thus, sleep disorders can have varied effects on the older adults' lives, which highlights the importance of investigating sleep management interventions for residents in LTCFs [19].

Previous evidence suggests that non-pharmacological approaches should be prioritized over drug use for the treatment of sleep disorders in older adults, and drug access is recommended only when non-pharmacological approaches have proved ineffective [20, 21]. However, drugs used for sleep disorders can lead to difficulties with tolerance and dependence [21], and long-term use increases the risk of dementia in older adults [22] and significantly increases the risk of falls [23]. Thus, non-pharmacological interventions including acupressure, auricular acupuncture, behavioral therapy, massage, hand massage, mindfulness meditation, tart cherry juice, sleep education, relaxation, and attention control, should be preferred. These interventions are crucial for improving older adults' sleep quality [24, 25], and it is necessary to establish evidence for their effectiveness in older adults.

Moreover, although systematic reviews of non-pharmacological interventions for older adults in LTCF settings have been conducted in other studies, and results regarding sleep improvement have been identified and evaluated, no meta-analyses have been performed [25, 26]. This could be because meta-analysis is not feasible due to considerable heterogeneity when including various countries and intervention methods. With more than a decade since the introduction of long-term care insurance in 2008, many non-pharmacological interventions for improving sleep have been employed in Korean LTCFs. In Korea, most studies have reported the impact of non-pharmacological interventions in patients with sleep disorders; however, not specifically in older adults [27]. Other studies that performed meta-analyses on the quality of sleep in older adults in the community were limited to physical activity programs [28]. For non-pharmacological interventions, most existing reviews analyzed participants with a specific type of dementia or those in a community setting. To bridge this gap in the existing literature, it is necessary to investigate all types of non-pharmacological interventions for older adults living in LTCFs. Therefore, the aim of this study was to provide a systematic review of the best available evidence for interventions suitable for older adults in Korean LTCFs, by integrating non-pharmacological sleep intervention studies and analyzing their effects.

This study investigated the effects of non-pharmacological sleep interventions among older adults in Korean LTCFs. In this review, we included one RCT, 11 NRCTs (78.6%), and 2 single-group pre-post studies. The results indicated that the RCT research was limited, possibly because it was difficult to create a controlled environment as required for randomized experimental studies in Korean LTCFs [40].

Among the types of non-pharmacological interventions, the most common interventions are those with aromatherapy with five studies (28.6%), possibly because aromatherapy can be used in multiple ways, such as skin application, inhalation, and massage [41], and can be safely applied to older adults without serious complications [42]. Meanwhile, in other countries, physical activity interventions, such as high-intensity physical strength training, walking, and elastic band exercises, have been implemented to improve sleep quality [43]. In Korea, there have been few physical activity interventions owing to the limited environment for patient safety management.

The most common intervention characteristics were intervention time of 30~60 minutes (six articles, 42.8%), intervention frequency of once a week (three articles, 21.4%), and a total intervention period of 4~8 weeks (six articles, 42.8%). These findings are consistent with previous studies where interventions were performed typically for 30~60 minutes per session, once a week, for a total of 4~8 sessions [44]. Furthermore, the improvement of sleep quality was confirmed in 12 articles (85.7%). Therefore, it can be inferred that non-pharmacological interventions are effective in improving sleep quality.

According to the meta-analysis results of 11 articles, an effect size of 1.0 was obtained, which indicates a very large intervention effect. Direct comparison was not possible, as no other study has analyzed the effect of non-pharmacological sleep interventions on older adults in Korean LTCFs. Chae et al. [45] reviewed studies investigating sleep disturbance intervention in cancer patients; Hedges' g was.78, with a medium effect size. Regarding the intervention type, music intervention showed the largest effect size, followed by massage, exercise (walking), and aroma therapy. Meanwhile, Chung and Park [46] meta-analyzed the effects of non-pharmacological interventions on adults with sleep disorders in Korea; Hedges' g was 2.36, indicating a large effect size. Among intervention types, auricular acupuncture therapy showed the largest effect size, followed by foot reflexology massage, aromatherapy, and laughter therapy. Previous studies have indicated that non-pharmacological interventions to improve sleep quality are effective; however, the most effective intervention type differs for each participant [47].

In this study, the aromatherapy's effect on sleep quality showed a medium effect size, whereas Lin et al. [48]'s study with adults over 20 years of age showed a large effect size. Therefore, aromatherapy can be regarded as effective in improving sleep quality. However, aroma therapy tends to show different effects depending on the intervention method; for instance, it has been reported that aroma inhalation therapy has a larger effect size than other application methods, such as massage or intake [41]. Herein, the one study on aroma inhalation could not be confirmed through subgroup analysis. Therefore, it is necessary to compare the effects of inhalation, non-inhalation, and mixing methods in future intervention studies [42].

As in the study of Kim and Oh [34], the effect size by group vs. individual interventions confirmed that group-based interventions had a larger effect size than individual interventions. While individual interventions were centered on physical approaches, in group interventions, interactions with other participants were made, and therefore, psychosocial effects are thought to be involved. This is similar to previous studies where a psychosocial approach was found to be effective for sleep intervention for older adults in a facility [49]. Furthermore, psychological approaches can induce physical relaxation that effectively alleviates physical symptoms [50].

A subgroup analysis was performed for the intervention period (within 4 weeks vs. more than 4 weeks); interventions within 4 weeks showed a medium effect size, whereas interventions longer than 4 weeks showed a very large effect size. Previous studies have reported similar findings with continuous interventions for 4 to 8 weeks after the start of intervention [51, 52]. Additionally, in 10 of 11 articles included in the meta-analysis, sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI) and KSS. The PSQI and KSS assess sleep disorders based on the "last one month;" thus, they are inappropriate for testing the effectiveness of interventions of less than four weeks [51], and it is thought that these results have been reflected. Therefore, for future intervention studies, it is necessary to plan an intervention period that can sufficiently secure the outcome measurement tool's validity or to select a measurement tool suitable for interventions of less than four weeks. Furthermore, most non-pharmacological interventions with an intervention period of more than 4 weeks were group interventions, such as Korean dance, mindfulness meditation, and therapeutic recreation. Therefore, it can be inferred that group intervention and intervention period exceeding 4 weeks had a mutual effect, and thus, the effect size should be interpreted accordingly. Additionally, this study analyzed only the intervention period because of the high heterogeneity in intervention time, frequency, and total sessions among the intervention methods. Thus, it is necessary to analyze the various aspects of intervention methods in the future.

As for improvement in sleep quality by control intervention type (sham intervention group vs. no treatment group), the experimental and sham intervention group showed a small effect size, whereas the experimental and no treatment group showed a very large effect size. As there are limited studies analyzing the effect size according to control intervention type, it is necessary to establish a clear rationale by repeatedly confirming the effect size for the control group intervention types in the future.

Certain limitations of this review would encourage further research. First, the effect size may be overestimated, as 10 of the 11 meta-analyzed articles employed NRCT interventions. Second, generalization is limited, as certain studies had an unclear or high risk of bias, for instance, non-blinding of assessors. Therefore, we suggest further randomized control studies targeting older adults in LTCFs, and a more objective and clear effect size should be established based on future RCT studies. Third, although there were various types of non-pharmacological interventions in the selected literature, there was a limit to analyzing the effect size for each intervention type, because the number of studies for each intervention type was insufficient for a subgroup analysis, except for aroma therapy. Moreover, the number of studies that measured objective sleep quality among the selected studies was insufficient; thus, the size of intervention effect on objective outcomes such as sleep time, sleep efficiency, and sleep-onset delay could not be estimated. Therefore, in future studies, it is recommended to objectively analyze sleep quality using objective measurement tools, such as actigraphy. Fourth, in the case of subgroup analysis according to the study design in other intervention types, there is a limitation in generalization because heterogeneity is rather high. Finally, a mixed analysis of NRCT and RCT showed a positive effect; however, the results of one RCT in our review did not report a statistically significant difference. Consequently, it limited the possibility to draw a clear conclusion about the intervention effect. Therefore, it is recommended that more RCTs be included in future studies.

The results of this systematic review and meta-analysis confirmed the possibility of non-pharmacological interventions affecting the improvement of sleep quality in the older adults in Korea LTCFs. The most frequently used non-pharmacological intervention was aromatherapy, used in five studies. The subgroup analysis revealed that group-based interventions and interventions ≥4 weeks were more positively associated with sleep quality. Therefore, further research investigating the different intervention types, durations, and control groups is needed to improve our understanding of sleep intervention for older adults in LTCF settings. Moreover, further studies are required to provide the best evidence for the standardization and effectiveness of specific interventions, along with detailed analyses of their feasibility in LTCFs in Korea.