Employee Sleep and Workplace Health Promotion: A Systematic Review

Abstract

Objective

Workplace-based employee health promotion programs often target weight loss or physical activity, yet there is growing attention to sleep as it affects employee health and performance. The goal of this review is to systematically examine workplace-based employee health interventions that measure sleep duration as an outcome.

Data Source

We conducted systematic searches in PubMed, Web of Knowledge, EMBASE, Scopus, and PsycINFO (n=6,177 records).

Study Inclusion and Exclusion Criteria

To be included in this systematic review, studies must include: (1) individuals age >18 years; (2) a worker health-related intervention; (3) an employee population; and (4) sleep duration as a primary or secondary outcome.

Results

Twenty studies met criteria. Mean health promotion program duration was 2.0 months (SD=1.3) and mean follow-up was 5.6 months (SD=6.5). The mean sample size of 395 employees (SD=700.8) had a mean age of 41.5 years (SD=5.2). Measures of sleep duration included self-report from a general questionnaire (n=12, 66.6%), self-report based on PSQI (n=4, 22.2%), and self-report and actigraphy combined (n=5, 27.7%). Studies most commonly included sleep hygiene (35.0%), yoga (25.0%), physical activity (10.0%), and cognitive behavioral therapy for insomnia (CBTI, 10.0%) interventions. Across the interventions, 9 different behavior change techniques (BCTs) were utilized; the majority of interventions used three or fewer BCTs, while 1 intervention utilized 4 BCTs. Study quality, on average, was 68.9% (SD=11.1). Half of the studies found workplace-based health promotion program exposure was associated with a desired increase in mean nightly sleep duration (n=10, 50.0%).

Conclusions

Our study findings suggest health promotion programs may be helpful for increasing employee sleep duration and subsequent daytime performance.

OBJECTIVE

Poor sleep is prevalent among working-age adults, [1,2] and is associated with numerous adverse workplace outcomes[3] Several national entities, including the American Heart Association, the National Institute for Occupational Health and Safety, and the Centers for Disease Control call for attention to workplace-based health promotion as a promising tool for promoting population health practices, such as sleep.[4-7] We systematically review the characteristics and outcomes of workplace health promotion interventions that measure employee sleep duration.

Insufficient sleep among employees has significant workplace consequences. For instance, insufficient employee sleep is associated with lower information processing,[8] impaired cognition,[9-12] and reduced task performance.[13] Litwiller and colleagues[3] found a significant association between employee sleep quality and duration and poor work-related outcomes, including workload, depression, and fatigue. Finally, Mullins and colleagues[14] have found that sleep mediates the relationship between job demands and job performance, suggesting that employee sleep is vital for an efficient workforce.

Sleep deprivation in the workplace comes at a cost to employers. Research has found employees with untreated insomnia cost employers $2,280 more on average than an employee without insomnia each year in terms of absenteeism, presenteeism (showing up to work but underperforming), reduced performance, accidents, and injuries.[15] It is estimated that total direct and indirect healthcare costs associated with insufficient sleep duration in the US range from $30 to $40 billion annually, an important consideration for employers with employer-based health insurance.[16,17] Another study using data from a state-wide employee health program promotion showed employee sleep difficulty is linked with absenteeism, lower workplace productivity, and increased healthcare costs.[18]

Although health programs for clinical populations diagnosed with insomnia, such as Cognitive Behavioral Therapy for Insomnia (CBTI) and sleep hygiene improvement efforts, the effectiveness of these approaches among general audiences, such as worksite-based groups, is less well known.[19] Workplace-based health promotion is a compelling approach for several reasons, such as the ability to reach a large proportion of the general population with evidence-based programs.[20] The field of workplace-based health promotion has become increasingly prevalent among US worksites. In 2004, national data demonstrated nearly all US employers with ≥750 employees offered policies and programs related to employee health; yet, only 6.9% of worksites offered what was termed ‘comprehensive’ worksite wellness initiatives.[21] According to Baker and colleagues, ‘comprehensive’ programs include both employee health promotion as well as employee risk reduction.[22] Although one report shows over 90% of employers with ≥50,000 employees were offered some health promotion, sleep was not a focus of any of the selected programs.[23] Although workplace health promotion for employees is increasingly common in the workplace, few are comprehensive and very little attention in programming has been given to employee sleep duration.

The goal of this review is to summarize the evidence on workplace health promotion programs that measure employee sleep duration as an outcome. Evidence suggests that sufficient sleep duration is an essential and modifiable determinant of health,[24-26] and as such, sleep is a powerful potential target for workplace wellness programs. By obtaining a more complete summary of the role of workplace-based employee health programs and sleep duration-related outcomes, our hope is that future workplace wellness programs can incorporate evidence-based health promotion programs to help employees maintain or obtain healthy sleep habits, such as sufficient sleep duration on a regular basis.

METHODS

We conducted a systematic review of studies that include sleep duration as an outcome of a health promotion program to promote various health behaviors among employees in a workplace setting. The search adheres to the Preferred Reporting for Systematic Review Protocols (PRISMA) guidelines. This review has been registered with PROSPERO: CRD42016037748.

Eligibility Criteria

Studies were included if they were interventional in nature (e.g., randomized controlled trial, RCT, 2-arm non-randomized intervention, 1-arm pre-test/post-test); carried out in employed adults (>18 years old); were conducted in a workplace context; and measured sleep duration as an outcome. Although this review did not have an English language restriction, all included articles were published in English. We excluded cross-sectional studies, studies that did not measure sleep duration as an outcome, and studies carried out in non-humans. Also, studies where target populations were shift workers were the target workplace population were excluded as these individuals need specialized recommendations due to circadian misalignment imposed by shift work schedules.[27]

Search Strategy

The following databases were searched: PubMed/Medline, Embase, Web of Science, the Cumulative Index to Nursing and Allied Health (CINAHL), PyscINFO, BIOSIS Citation Index, and the Cochrane Library. The New York Academy of Medicine Grey Literature, WorldCat’s OAISster and OpenGrey databases also retrieved relevant literature. Additionally, a search was performed within the table of contents of the following journals: J of Occup Environ Med, Sleep, Int Arch Occup Environ Health, Occup Med (London), J Sleep Res, Occup Environ Med, Sleep Med, and J Occup Health. The date of the search included all published literature in press on or before September 1, 2018 (date last searched). From the articles that were identified as eligible, a search of the articles within their bibliographies was also conducted. We used combinations of text words and thesaurus terms i.e., work [MeSH], workplace [MeSH Terms], occupational health [MeSH Terms], occupational health [MeSH Terms], employee [All Fields], health promotion [MeSH Terms], interventions [All Fields], and sleep [MeSH Terms].

Data Screening

Records were identified using the search strategy, then exported to EndNote X7. Irrelevant records were screened out by two trained research assistants (RR and PU) based on titles and abstracts. Full text retrieval was conducted for potentially eligible articles. Bibliographies of the selected articles were also analyzed for potentially eligible articles. The studies that were retrieved for detailed analysis were assessed by the independent reviewers to ensure they satisfied the inclusion criteria. Any disagreements were resolved through consensus or discussion with all co-authors.

Data Extraction

Extracted data were analyzed using RevMan (Version 5). Data were extracted in several general categories, including study design and characteristics; study population demographics (age, race/ethnicity); intervention characteristics (e.g., workforce population, intervention components, such as behavior change techniques, BCTs, and duration of the intervention), sleep duration measurement (e.g., wrist actigraphy, or self-report) and outcomes (e.g., change in sleep duration associated with exposure to the intervention).[28] Wrist actigraphy refers to a device that offers high correspondence to the gold standard for sleep measurement, which is polysomnography. When the specific Pittsburgh Sleep Quality Index (PSQI) item for sleep duration was provided in selected studies,[29] it was included in the review summary; otherwise, the PSQI global score is displayed, as it is constitutive of sleep duration, among other factors. In some cases, the baseline and follow-up sleep times were reported. In other cases, change in sleep duration was reported. The data extracted in this review reflect the published findings.

Behavior Change Techniques (BCTs)

BCTs were coded in the current systematic review according to the 40-BCT taxonomy published by Michie and colleagues.[30] This taxonomy of BCTs outlines a wide array of BCTs and strategies, ranging from goal setting to role modeling. Two reviewers assigned BCTs to each study, then discussed any discrepancies until agreement was reached. The BCT taxonomy was adhered to strictly with the exception of “educational seminars” that lacked a direct corollary BCT taxonomy. Consequently, the study team added “educational seminars” as a BCT in this review. Figure 1 displays the flow diagram of study screening according to PRISMA guidelines.

Figure.

Flow Diagram of Study Screening and Selection according to PRISMA guidlines

Intervention Type

Coders were trained to extract the ‘intervention type,’ or the health promotion outcome that was the focal point of the intervention, such as sleep hygiene, which refers to evidence-based recommended healthy sleep practices like obtaining sufficient sleep duration and practicing a relaxing bedtime routine. Data extraction identified 7 types of interventions including : 1) sleep hygiene; 2) yoga; 3) schedule control; 4) physical activity; 5) Cognitive Behavioral Therapy for Insomnia, CBTI; 6) stress reduction; or 7) napping. For instance, studies that offered educational interventions to improve sleep behavior and hygiene were coded as “sleep hygiene”[31] whereas yoga interventions offered instructions in specific stretching and exercise techniques, and physical activity interventions included teaching or coaching exercise routines or habits broadly. CBTI interventions were delivered in accordance with previous efforts to address insomnia-like symptoms.[32] Finally, stress reduction focused on exercises such as meditation techniques for reducing stress, while napping interventions included policies and initiatives to encourage employee napping.

Study Quality Evaluation and Data Analysis

The quality of the studies included in the systematic review was assessed using the Downs and Black checklist.[33] The Downs and Black Checklist is a 27-item scoring system assessing the following domains: reporting, external validity, internal validity/bias, internal validity/confounding, and power. While there is some discordance on how to use the quality scores from the Downs and Black approach,[34] we used a quality rating that determined 21 (80.8%) and higher as high quality, 11 – 20 (42.4-80.8%) moderate quality, and 10 or lower as poor quality (<42.2%).[35] Quality ratings were determined for the studies in this systematic review independently by two reviewers (RR, PU). Discrepancies were adjudicated through discussion until consensus with coauthors was reached. We quantitatively summarized worksite-based interventions measuring employee sleep duration as an outcome by intervention type (e.g., sleep hygiene, yoga) and examined the effect of intervention on sleep duration measured via self-report, actigraphy, or both.

RESULTS

Description of Study Characteristics

The general characteristics of the studies are shown in Table 1. Studies used one of three designs, including: RCT, 2-arm non-randomized pre-post, or 1-arm pre-post. Among the studies, industries where interventions were conducted included financial services (n=2, 10.0%), educational services (n=2, 10.0%), media (n=2, 11.1%), and manufacturing (n=2, 10.0%). Unfortunately, 6 studies did not list the industry. Target populations were mostly office employees (n=9, 45.5%) and factory workers (n=2, 10.0%), with the remainder of studies including manufacturing workers, lab workers, cleaning staff, and nursing home employees and managers. The remaining studies did not describe the target population (n=5, 25.0%). The duration of the workplace interventions ranged from several hours[36] to 4-months[37] (mean=2.0, s.d.=1.3). The longest follow-up range from several weeks post-intervention[38] to two years post-intervention[39] (mean=5.4, s.d.=6.5).

Table 1.

Summary of study characteristics (n=20).

First author	Year	Study Design	Industry	Workforce Population	Intervention Duration (months)	Longest Follow- Up (months)
Adachi [46]	2003	1-arm pre-post study	Not specified	Office employees	1 mo	1 mo
Adachi [47]	2008	1-arm pre-post study	Not specified	Not specified	1 mo	12 mo
Adler [41]	2017	1-arm pre-post study	Military	Soldiers	1 mo	2 mo
Chen [31]	2010	1-arm pre-post study	Not specified	Not specified	1 mo	1.5 mo
de Bruin [60]	2017	1-arm pre-post study	Not specified	Not specified	1.5 mo	6 mo
Hallman [61]	2017	RCT	Maintenance	Cleaning staff	4 mo	4 mo
Itani[62]	2018	2-arm non-randomized intervention	Manufacturing	Factory Workers	4 mo	4 mo
Järnefelt [43]	2012	2-arm non-randomized intervention	Media	Office employees	1 mo	6 mo
Järnefelt [39]	2014	2-arm non-randomized intervention	Media	Office employees	1 wk	2 years
Kakinuma [36]	2010	RCT	Financial Services	Office employees	1 hr	1 mo
Klatt [45]	2009	RCT	Educational Services	Office employees	2 mo	1 mo
Klatt [29]	2017	RCT	Financial Services	Office employees	2 mo	2 mo
Li [48]	2017	2-arm non-randomized intervention	Healthcare	Office employees	1 mo	1 mo
Marino [37]	2016	RCT	Nursing Homes	Extended care workers	4 mo	12 mo
Nakada[63]	2018	Quasi randomized trial	Pharmaceutical	Office employees	5 mo	1 mo
Olson [44]	2015	RCT	Information Technol.	Office employees	3 mo	12 mo
Schiller [42]	2017	RCT	Not specified	Not specified	2 mo	18 mo
Suzuki [38]	2008	RCT	Not specified	Not specified	3 wk	3 wk
Takahashi [40]	2004	1-arm pre-post study	Manufacturing	Factory Workers	1 wk	3 wks
Trousselard [64]	2014	1-arm pre-post study	Educational Services	Lab Workers	3 mo	3 mo

Demographics of Study Samples

Table 2 displays characteristics of the populations in the studies selected for this review. Sample size ranged from 8[40] to 2,932[41] (mean=395.3, s.d.=700.9; median=53.5). Regarding participants who were lost to follow-up, the studies ranged from 0 (0.0%)[31] to 240 (29.3%)[42] participants (mean=51.6, s.d.=82.4; median=6.0). Participant mean age across the studies ranged from 25.9 years[41] to 49.0 years[30] mean=41.5, s.d.=5.2). The mean proportion of females was 26.8% and males was 73.1%. Race/ethnicity information was provided by only 7 studies (35.0%). Among those studies reporting race/ethnicity, mean proportion reporting being White was 34.0%, 2.2% Latino/Hispanic, 0.6% Black/African American, 56.8% Asian, and 6.4% other.

Table 2.

Study sample demographics (n = 20 studies).

First author	Year	Sample Size	Lost to Follow- Up	Mean Age at Baseline (years)	Males (%)	Females (%)	White (%)	Latino/ Hispanic (%)	Black/ African American (%)	Asian (%)	Other (%)
Adachi [46]	2003	114	39	37.5	45.6%	54.4%
Adachi [47]	2008	47	0	38.7	36.2%	63.8%
Adler [41]	2017	2932	52	25.9	92.4%	7.6%
Chen [31]	2010	37	0	38.3	0.0%	100.0%
de Bruin [60]	2017	26	3	44.9	15.4%	84.6%
Hallman [61]	2017	173	34	45.1	24.3%	75.7%
Itani[62]	2018	800	15	42.3	596	118				783
Järnefelt [43]	2012	33	6	44.4	18.2%	81.8%
Järnefelt [39]	2014	59	15	46.0	50.8%	49.2%
Kakinuma [36]	2010	581	190	44.4	80.8%	19.2%
Klatt [45]	2009	48	6	33.8	19.0%	81.0%	88.9		4.4	6.7
Klatt [29]	2017	81	54	45.0	24.4%	75.6%	87.8	2.4	2.4	4.9	2.4
Li [48]	2017	33	6	42.0	23.1%	76.9%
Marino [37]	2016	1285	503	38.5	8.2	91.8	66.6	11.5	12.7		6.7
Nakada[63]	2018	77	7	43.1	42	28				70
Olson [44]	2015	701	227	46.7	59.7%	40.3%	67.7	7.0	1.4	3.4	20.4
Schiller [42]	2017	820	240	44.6	0.0%	100.0%	100.0
Suzuki [38]	2008	43	5	39.6	0.0%	100.0%				100.0
Takahashi [40]	2004	8	0	49.0	100.0%	0.0%				100.0
Trousselard [64]	2014	9	0	39.7	100.0%	0.0%

Intervention Effects on Employee Sleep Duration

Table 3 summarizes the interventions and effects on employee sleep duration. The most common focus for interventions was on sleep hygiene (n=7, 35.0%), followed by yoga (n=5, 25.0%), schedule control (n=2, 10.0%), physical activity (n=2, 10.0%), Cognitive Behavioral Therapy for Insomnia (CBTI, (n=2, 10.0%), general healthy lifestyle (n=1, 5.0%), and napping (n=1, 5.0%). Measures of sleep duration included self-report from a general questionnaire (n=12, 66.6%), self-report based on PSQI (n=4, 22.2%), and both self-report and actigraphy combined (n=5, 27.7%).

Table 3.

Impact of interventions on employee sleep behavior and quality ratings (n=20 studies).

First Author	Year	Intervention	Sleep Duration Measure	Behavior Change Techniques (BCTs)	Main Findings (Detailed)*	Main Findings (Brief)	Study Quality
Adachi	2003	Sleep hygiene	Self-report	a) self-monitoring b) goal setting (behavior) c) rewards contingent on behavior	Participation in the intervention was associated with an increase in sleep duration (5.96 to 6.1h/night, p-value <0.01)	+	53.8%
Adachi	2008	Sleep hygiene	Self-report	a) instructional seminar b) follow-up prompts c) self-monitoring d) rewards contingent on behavior	Participation in the intervention was associated with an increase in sleep duration (5.52 to 5.97h/night, p-value <0.05)	+	57.7%
Adler †	2017	Sleep hygiene	Self-report	a) self-monitoring b) follow-up prompts	Participation in the intervention was associated with an increase in sleep duration (p-value =0.05)	+	86.5%
Chen	2010	Sleep hygiene	Self-report (PSQI)	a) instructional seminar	Participation in the intervention was associated with an increase in sleep duration (PSQI sleep duration subcomponent at baseline=1.0 v. follow-up:0.6: p-value =0.002)	+	57.7%
de Bruin	2017	Yoga	Self-report	a) instructional seminar b) behavioral modeling c) practice session(s)	Participants who received the intervention reported increased sleep duration (7.0 to 7.4 h/night), although no p-value was provided to determine significance	NC	53.8%
Hallman	2017	Physical activity	Self-report	a) instructional seminar(s) b) practice session(s)	Participation in the intervention was not associated with an increase in sleep duration (7.20 v. 7.30h/night, p=n.s)	NC	82.7%
Itani	2018	Sleep hygiene	PSQI	a) instructional session b) follow-up prompts	Participation in the intervention, compared to control, was not associated with a significant increase in sleep duration as measured by PSQI (p=.218)	NC	49.4%
Jarnefelt	2012	CBTI	Self-report, actigraph	a) instructional seminar (CBTI)	Participation in the intervention was associated with an increase in reported duration (7.3 to 7.7h/night, p=0.027), but not according to actigraphy (p-value =n.s.).	+	80.8%
Jarnefelt	2014	CBTI	Self-report	a) instructional seminar(s) (CBTI) b) one-on-one coaching	Participation in the intervention was associated with an increase in sleep duration (7.1 to 7.5h/night, p-value =0.05)	+	57.7%
Kakinuma	2010	Sleep hygiene	Self-report	a) instructional seminar(s) b) self-monitoring c) follow-up prompts	Participation in the intervention was not associated with an improvement in sleep duration (0.03 v. 0.07h/night, p-value =0.69)	NC	78.8%
Klatt	2009	Yoga	Self-report (PSQI sleep duration)	a) instructional seminar(s) b) self-monitoring	Participants in the intervention was associated with an improvement in PSQI global (6.73 to 5.00, p-value: 0.0018); but not in PSQI sleep duration (0.91 to 0.82, p-value=.3287)	+	80.8%
Klatt	2017	Yoga	Self-report (PSQI)	a) instructional seminar(s) b) goal-setting	Among participants who received the intervention, PSQI scores improved (5.93v. 3.89, p-value=.005)	+	71.2%
Li	2017	Physical activity	Self-report	a) environmental restructuring (sit/stand workstations provided) b) follow-up prompts	Participation in the intervention as not associated with an increase in sleep quantity (p-value=.708)	NC	71.2%
Nakada	2018	Mental health intervention	Self-report, actigraph	a) instructional session b) self-monitoring	Participation in the intervention, compared to control, was associated with a significant increase in sleep duration as measured by actigraphy intervention: 5.1 v 4.6 h; p=.036.)	+	54.32%
Marino	2016	Schedule control	Self-report, actigraph	a) instructional session(s) (result-based work culture) b) behavioral modeling c) goal setting	Participation in the intervention was not associated with an increase in sleep duration (employees p-value=.179, managers=.321).	NC	76.9%
Olson	2015	Schedule control	Self-report, actigraph	a) instructional session(s) (result-based work culture) b) behavioral modeling c) goal setting	Participation in the intervention was associated with an increase in sleep duration according to actigraphy (p-value=.041)	+	80.8%
Schiller	2017	Schedule control	Self-report	a) environmental restructuring (reduced working hours per week by 25%	Participation in the intervention was not associated with an increase in sleep duration (p-value=.263)	NC	76.9%
Suzuki	2008	Sleep hygiene	Self-report (PSQI)	a) instructional seminar(s) (sleep hygiene) b) self-monitoring c) follow-up prompts	Participation in the intervention was not associated with an increase in sleep duration (5.83 v. 5.88h/night, p-value =.09); but PSQI global did improve (7.70 v. 7.40, p-value =.009)	NC	65.4%
Takahashi	2004	Napping	Actigraph	a) practice session(s) (napping)	Participation in the intervention was not associated with an increase in sleep duration according to actigraphy 5.6 v. 5.6h/night; p-value =n.s.)	NC	69.2%
Trousselard	2014	Stress reduction	Self-report	a) instructional seminar(s)	Participation in the intervention was not associated with an increase in sleep duration (6.5 v. 6.5hh/night, p-value=.890)	NC	61.5%

Quality ratings: High >80.8%; Moderate 42.2-80.8%; Low: <42.2%.

^†Adler et al.,[41] used biofeedback from actigraphy in the intervention, yet did not report actigraphy output, thus self-report only is noted as the sleep duration measure.

^*Main findings are reflected from the published papers. In some cases, authors present change scores, in other cases authors compare follow-up sleep duration between intervention and control. Findings are displayed, consequently, either as sleep quantity (e.g., h/night), differences (e.g., 0.2 h/night more than control), or PSQI global from 0 to 21 (higher scores indicate work sleep outcomes), or PSQI sleep duration from 0 to 3 (0: <7h/night; 1: 6-7h/night; 2: 7-8h/night; 3:>8h/night.

Regarding BCTs included in the 20 studies, there were 9 individual BCTs from the Michie et al., taxonomy[30] that were used in the eligible interventions included in this review. The most common BCTs were instructional seminar (n=14, 38.9%), self-monitoring (n=5, 13.2%), goal setting (n=4, 10.5%), follow-up prompts (n=5, 13.2%), practice sessions (n=3, 7.9%), behavioral modeling (n=2, 5.3%), rewards contingent upon behavior (n=2, 5.3%), and one administered one-on-one coaching (n=1, 5.3%). Across the studies, 13 utilized two BCTs in their intervention (76.5%), 7 used three BCTs (7=6, 41.2%), and 1 intervention featured 4 BCTs (n=1, 5.9%).

In terms of effects of the interventions on employee sleep duration, this review identified a positive effect of intervention exposure on employee sleep duration in 10 studies (50.0%). Examining effects by intervention type, it appears CBTI interventions, Järnefelt et al.[43] and Järnefelt et al.[39], improved employee sleep duration (n=2/n=2, 100%). Also, one schedule control study conducted by Olson et al.[44] improved employee sleep duration (n=1/n=3, 33.3%). Among the yoga interventions, Klatt et al.[29] and Klatt et al.,[45] achieved improvement in sleep duration (n=2/3, 66.7%). Similarly, among the sleep hygiene interventions, the following studies improved sleep duration (n=4/n=6; 66.7%): Adachi et al.,[46] Adachi et al.,[47] Chen et al.,[31] and Li et al.[48] None of the stress reduction, schedule control, or napping interventions were associated with a statistically significant improvement in sleep duration among employees.

The average quality for selected studies was 70.2% (s.d.=10.9%, median: 71.2%, range: 53.8% to 86.5%). According to Downs and Black, 80.8% or higher indicates high quality, and rating between 42 and 80.8% indicates moderate quality. Therefore, most studies were in the moderate quality range.

DISCUSSION

We characterized the types of workplace-based employee health interventions that included sleep duration as a primary or secondary outcome, and examined intervention type, study design and quality, and intervention outcomes on employee sleep. Approximately half of the 20 interventions in this review achieved a statistically significant increase in sleep duration.

Intervention size ranged widely from a pilot study in a small employee group,[46] to a self-monitoring intervention using a large employee population.[41] Interventions were delivered to employees in a variety of industries, including financial services, maintenance, and education. Target employees included general office employees, cleaning personnel, lab workers, and manufacturing workers. Unfortunately, several studies did not specify their worker population or industry. The duration of the intervention in this review ranged widely from several hours to several months, presenting a wide range of intervention lengths.

While the participant mean age in this review was 41.5 years, it is interesting to consider the role of workplace-based health promotion as our population growth in older adults (e.g., 40 years and older) continues to accelerate.[49,50] The proportion of 50-65 year old individuals will double in the next 20 years, likely shifting the age of participants in these types of programs, presenting new challenges and opportunities for health promotion in the workplace to advance population-level outcomes.

In our review, the most common interventions were those focused on improving employee sleep hygiene. Among the sleep hygiene interventions, most were successful in increasing sleep quantity. This is interesting, for sleep hygiene practices are under-explored in non-clinical populations, such as working adults, but also not always sufficient for increasing sleep duration.[19] The next most common intervention type was yoga, two thirds of which were successful in increasing sleep among employees. The interventions focused on CBTI were all effective for increasing sleep duration. This is not surprising as there is a large body of evidence to show the effectiveness of CBTI approaches for improving sleep;[32,51,52] however, it should be noted that the studies testing CBTI were not in populations with clinical insomnia diagnoses. Consequently, future research may further explore the contributing factors to the increased duration among recipients of CBTI when an insomnia diagnosis is absent.

Also effective in increasing employee sleep duration was one intervention focused on reducing work-family conflict. In this RCT to reduce work-family conflict conducted by Olson et al.,[44] researchers evaluated a comprehensive program to deliver multiple targeted BCTs to employees (and managers) over 3 months. According to their results, the intervention–while not expressly addressing sleep– increased employee sleep duration. These results further emphasize the interconnectedness of social/contextual factors that matter for health outcomes, as well as the importance of multi-level and multi-modal interventions.[53]

Only 1 study meeting our criteria focused on napping in the workplace, although evidence suggests there are benefits of napping for reducing fatigue and improving alertness.[54] In light of the evidence to suggest a mid-day dip in alertness among the majority of the population,[55] napping in the afternoon may represent a healthy routine well-suited to attention from workplace based health promotion as most workers are at the office at the ideal time to nap. The shift work exclusionary criteria for our study excluded nap interventions tailored to such occupational groups as medical residents and nurses whose jobs often include shift work. Future research could examine the effect of napping interventions on employee sleep and workplace-based performance measures, such as productivity. That napping interventions are precluded in some industries where napping is a firing offense (e.g., air traffic controllers) further limits nap-focused-interventions in worksites.

Our findings show a range of BCTs utilized in health promotion to address employee sleep. Instructional seminars, self-monitoring, and goal setting were the most common BCTs applied in the studies that met criteria for this review. These findings are consistent with previous workplace research, but did not emphasize leadership support for employee health promotion, which has been a focus in other workplace health promotion efforts [23].

Overall, the findings in this study suggest the potential role for workplace-based health promotion as a tool for improving sleep among working adults. Our results do nevertheless highlight limited attention to sleep as an outcome in workplace-based interventions. Our results also illuminate future areas for interventions that seek to improve sleep duration.

Implications of Findings

This review identified workplace-based health interventions that measured employee sleep duration and found half of the selected studies improved employee sleep duration. Particularly important, the interventions focused on employee sleep hygiene and CBTI had the highest success rates of behavior change toward healthier sleep duration among employees, offering further support from previous literature for these approaches.[19,32,56] To the extent to which employers can meaningfully improve sleep, our findings suggests that workers stand to benefit from worksite-based programs aimed at improving sleep.[15,57] Worksite-based employee health interventions could potentially also serve a benefit to society in terms of improved work-related productivity in the aggregate.[17,18]

Limitations

A major limitation of the literature to date is the paucity of worksite-based intervention research that utilized objective measures of sleep. As the majority of studies measured self-reported sleep duration, the results may have been biased and possible associations between program exposure and sleep may have been missed. Our review identified 2 studies examining workplace health interventions that measured sleep objectively using actigraphy, while the remainder of studies relied on self-report of sleep duration. Although we assessed the quality of the studies, we were unable to evaluate intervention fidelity (or that the interventions were delivered as intended). It is also important to note that some interventions had multiple BCT components (e.g., one-on-one coaching and instructional seminars), Overall, the findings in this study illuminate challenging to compare outcomes of different BCTs across the studies, as the authors had originally planned a meta-analysis. It should be noted only published research is included, thus, there may be publication bias from studies that did not achieve significant findings. Furthermore, studies often had limited description of the industry or job descriptions of employees in their studies. Finally, studies in this review did not quantify the economic cost savings of improved employee health or sleep duration. As well-documented evidence can attest, sleep deprivation carries significant economic tolls in terms of reduced productivity (due to both absenteeism and presenteeism) as well as work-related injury.[17,58,59] These factors bear consequence for outcomes, but also for future interventions, as it would be useful to understand the types of employees who enroll in health promotion programs and their unique job demands as we consider effects of interventions.

Future Research

Future interventions may consider combining educational approaches and intervention components for behavior change among employees with regard to sleep identified in this review, such as yoga plus CBTI. Future research may also examine areas not specifically related to sleep, but with implications for sleep, such as reducing work-family conflict. There is need for future research to utilize objective measures of sleep duration, however, so as to better characterize the interventions specifically targeting sleep and their objective effects on employee post-intervention sleep. There is also need for future meta analytic work to quantify the impact of employee health promotion programming and its implications for employee sleep, but also the economic benefit (or disadvantage) of interventions that endeavor to increase employee sleep duration or improve employee sleep quality. Specifically, it would be particularly useful to identify the optimal duration of intervention exposure for optimal behavior change. Furthermore, research might include employee populations that are racially and ethnically diverse. Research should also consider the independent impact of workplace-based health promotion on sleep quality in addition to sleep duration, and determine the predictors of effective interventions (e.g., BCTs, intervention duration).

Conclusion

This systematic review provides evidence that workplace-based interventions can improve employee sleep duration. Further, it is noted that interventions focused on sleep hygiene, CBTI, and reducing work-family conflict may be promising ways to increase employee sleep duration. Given the strong link between employee sleep and alertness and work-related productivity, employers would likely be well-served to draw heightened attention to sleep in their health promotion efforts.

So What?

What is already known on this topic?

Workplace-based employee health promotion holds tremendous promise for advancing health promotional behaviors, such as exercise and nutrition. Indeed, these programs are rapidly expanding at worksites across the US.

What does this article add?

While workplace-based health promotion programs are increasingly popular, few aim to promote healthy sleep among employees. Sleep is directly related to workplace outcomes, including employee health, absenteeism, and productivity. This systematic review systematically examines workplace-based employee health interventions that measured sleep duration as an outcome.

What are the implications for health promotion practice or research?

This systematic review identifies the evidence on workplace-based health promotion aiming to improve sleep health among employees. Specifically, we articulate the behavioral change techniques employed in successful programs, and identify opportunities for improving employee sleep health using worksite-based health promotional activities.

List of Abbreviations

CBTI

Cognitive Behavioral Therapy for Insomnia

BCTs

Behavior Change Techniques

PRISMA

Preferred Reporting for Systematic Review Protocols

PSQI

Pittsburgh Sleep Quality Index

US

United States

PROSPERO

International prospective register of systematic reviews

RCT

Randomized Clinical Trial