Alcohol Use Predicts Longer But More Fragmented Sleep: A Daily Diary Study of Alcohol, Sleep, and PTSD in Nurses

Abstract

Background

Due to the demanding nature of their profession, nurses are at risk of experiencing irregular sleep patterns, substance use, and fatigue. Evidence supports a reciprocal relationship between alcohol use and sleep disturbances; however, no research has examined such a link in a sample of nurses. One factor that may further impact the dynamic between alcohol and sleep patterns is posttraumatic stress disorder (PTSD) symptoms. We investigated the daily bidirectional associations between alcohol use and several sleep domains (i.e., self-report and actigraphy-determined sleep), and moderation by baseline PTSD symptom severity.

Method

Over a 14-day period, 392 nurses (92% female; 78% White) completed sleep diaries and actigraphy to assess alcohol use and sleep patterns. Within-person bidirectional associations between alcohol and sleep were examined using multilevel models, with symptoms of PTSD as a cross-level moderator.

Results

Daily alcohol use (i.e., ≥ 1 alcoholic beverage; 25.76%) was associated with shorter self-reported sleep onset latency (b = −4.21, p = .003) but longer self-reported wake after sleep onset (b = 2.36, p = .009). Additionally, days with any alcohol use were associated with longer self-reported sleep duration (b = 15.60, p = .006) and actigraphy-determined sleep duration (b = 10.06, p = .037). No sleep variables were associated with next-day alcohol use. Bidirectional associations between alcohol consumption and sleep were similar regardless of baseline PTSD symptoms.

Conclusion

Our results suggested that on days when nurses drank alcohol, they experienced longer but also more fragmented sleep.

Introduction

Nursing is a demanding profession, and the stressors nurses face may have a significant negative impact on their health. Such high stakes working environments can interfere with nurses’ sleep quality [1–3]. On average (49–79%), nurses report sleeping fewer than the recommended minimum seven hours per night [4], and 31–57% meet criteria for chronic insomnia [5–8]. Additionally a number of studies have indicated that nurses often report lower levels of sleep satisfaction [9, 10] and poor sleep quality [9, 11, 12]. Nurses’ sleep may also be affected by behavioral factors, such as alcohol use [13]. Approximately 30% of nurses report regularly consuming alcohol, and 13% demonstrate alcohol use problems [14]. Indeed, many nurses report using substances to cope with work-related stressors [15, 16]. Alternatively, sleep disturbances have been found to facilitate drinking behavior (i.e., cravings, urges) [17]. Despite research on associations between sleep and alcohol usage, little work in this area has been conducted in nurses [18–21]. This represents a critical public health gap to address, as nurses are at high risk for insomnia and substance use disorders [22–24]. Furthermore, these issues may not only affect nurses’ own health and well-being, but also patient outcomes.

There is considerable evidence supporting a bidirectional link between alcohol use and sleep disturbances in the general population. On a physiological level, alcohol consumption affects the quality and architecture of sleep [25], reducing REM sleep and increasing slow-wave sleep [13]. Further, alcohol consumption adversely affects sleep-related neuroendocrine and neurobiological responses, such as inhibiting excitatory activity NMDA receptors and facilitating GABA receptors, thus promoting sedative effects [13, 26]. Alcohol, however, also has a paradoxically stimulating effect, which may cause sleep onset to be delayed [27]. Whether alcohol induces or increases sleep latency depends on the dose of alcohol and interval between drinking and going to sleep [27]. One systematic review found that drinking heavily has been linked to a longer time to fall asleep, shorter time in bed sleeping, and longer time waking up after onset of sleep [28]. Long-term or large quantities of alcohol consumption can also adversely affect sleep quality [29]. Through daily diary assessment, increased alcohol consumption was related to later bedtimes and rise times [30]. On the other hand, researchers have found that consuming as little as 1–1.5 drinks of alcohol can decrease sleep quality by approximately 9.3% [31]. Together, these studies support the notion that alcohol can disrupt sleep onset, wakening, and duration at various consumption levels.

In terms of whether sleep plays a role in subsequent alcohol consumption, some studies suggest that disturbed sleep (i.e., insomnia associated with distress) increases the likelihood of alcohol-related problems [32]. Individuals with poor sleep quality may turn to alcohol to for its sedative effects to induce sleep [33]. Furthermore, poor sleep quality may impair executive function, increase impulsive behaviors, and increase fatigue, making substance use more likely [34]. Evidence regarding a relationship between sleep duration and alcohol consumption, however, is mixed [30, 35]. Cross-sectional reports have found that short-duration sleepers (i.e., self-report ≤ 6 h of sleep) were significantly more likely to consume alcohol [35]. However, self-reported daily diary assessments showed no relationship between sleep duration and alcohol consumption [30]. The results of these studies provide important stepping-stones to explore the directionality of the alcohol use and sleep link; however, most samples consisted of college students, making findings less generalizable to other populations.

Poor sleep quality, short sleep duration, and alcohol usage are common among nurses in the United States [7, 36]. Sleepiness and circadian disruption are typically associated with shift work, which is required for essential occupations such as nursing. However, shift work can compromise work-related health, safety, and performance [36, 37]. Consuming alcohol has been linked to delayed sleep–wake times among night shift workers [38], and nurses may use alcohol to combat shift-work related fatigue and stressful work conditions. Yet little research has examined the relationship between sleep and alcohol consumption among nurses. From the few studies conducted, alcohol use among nurses has been associated with increased risk of sleep disturbances [18], which is consistent with findings in other samples [39, 40]. However, further research is needed to clarify the nature of the association between alcohol and sleep in nurses, as well as to examine other factors that may exacerbate this association.

A wide range of problematic health conditions, including sleep disruption and alcohol misuse, are associated with exposure to trauma and posttraumatic stress disorder (PTSD). PTSD symptoms can be classified into four categories: intrusive thoughts or re-experiencing, avoidance of situations that are reminders of the event, negative changes in thoughts and feelings, and hyperarousal and reactivity symptoms [41]. In fact, sleep problems are now widely recognized as a core symptom of PTSD [42, 43], and exposure to traumatic events has been positively associated with both alcohol use and alcohol use disorders [44, 45]. The reinforcing cycle between sleep problems and alcohol use may be exacerbated by PTSD symptoms in nurses [44]. Nurses are continuously exposed to workplace trauma [46, 47] and approximately 18% of nurses meet the diagnostic threshold for PTSD [47] with many more experiencing subthreshold symptoms. PTSD may make it more difficult to cope with disrupted sleep [48] and regulate emotions [49] leading to subsequent alcohol use [50]. More specifically, poor sleep can lead to increased irritability and negative affect, impairing one’s ability to effectively manage and cope with emotional distress [51]. This may be especially true for individuals who have experienced trauma, as they may turn to alcohol as a way of alleviating their emotional distress caused by sleep problems or other symptoms of PTSD [52, 53]. Conversely, alcohol may have a more pronounced detrimental effect on sleep duration and continuity among those nurses with existing PTSD-related sleep problems. Furthermore, exposure to traumatic events predicts alcohol use in nurses [15]. Staffing shortages, growing patient needs, and reduced capacities for providing safe care are often associated with higher demands and may also increase the probability of substance use and sleep problems [54–56]. Other research in samples similar to nurses (e.g., firefighters) has demonstrated that PTSD symptom severity and sleep disturbances interact to predict increased alcohol use [57]. Thus, it is possible that nurses experiencing elevated PTSD symptom severity may exhibit strong reciprocal associations between alcohol use and sleep issues, as substances are often utilized as a coping mechanism for managing symptoms associated with both PTSD and sleep disturbances. Yet, little is known about the reciprocal relationship between alcohol use and sleep on a day-to-day level and whether PTSD symptoms may exacerbate this association among nurses.

To expand previous research, we examined: Aim 1) whether daily alcohol consumption had a bidirectional relationship with several sleep indices (i.e., total sleep time, sleep efficiency, sleep onset latency, wake after sleep onset, sleep quality) measured across 14 days using sleep diaries and actigraphy in a sample of nurses and Aim 2) whether these associations differed by PTSD symptom severity. We hypothesized that greater daily alcohol use would be associated with poorer sleep that night (e.g., shorter total sleep time, lower sleep efficiency, poorer self-reported sleep quality, shorter sleep onset latency, and greater wake after sleep onset). We also hypothesized that these same metrics of poorer sleep would be associated with greater alcohol use the next day. Finally, we hypothesized that greater PTSD symptom severity at baseline would augment these bidirectional associations between daily alcohol use and poor sleep.

Method

Procedures

Interested individuals were first screened for inclusion/exclusion criteria, and eligible participants were invited to participate in the study. Inclusion criteria were: 1) not yet received the current season’s influenza vaccine, 2) between the ages of 18 and 65, and 3) registered nurses actively working at least part-time at one of two regional hospitals. Exclusion criteria were: 1) pregnant/breastfeeding or planning to become pregnant or 2) having an egg allergy (due to the egg compounds in the vaccine).

All study procedures were approved by the hospital and university Institutional Review Boards. This study was part of a larger investigation on the effects of sleep on antibody response to the influenza vaccine (R01AI128359–01) that occurred between September 2018 to November 2018. Participants were recruited from two Dallas, Texas regional hospitals through nursing staff presentations, notification through employee email systems, and flyers that directed them to an initial online consent form. Nurses (N = 461) provided online consent and were asked to complete initial online Qualtrics surveys to collect demographic information as well as retrospective self-report estimates of recent health. Participants were then invited to enroll in the main portion of the study in the early fall (i.e., the start of the influenza season), which included completion of in-person informed consent and initial data collection approximately one month later. Of the 461 nurses, some chose not to enroll in the larger part of the study (N = 69). Remaining participants (N = 392) were given instructions on completing the alcohol consumption surveys, sleep diaries, and wearing the actigraph, which they completed for the subsequent 14 days.

Participants

In general, the participants in this study reflected the demographics of the United States nursing population [58]. Table 1 reports demographic and health characteristics for the entire sample. Most participants identified as White (77.8%), female (91.8%), and non-Hispanic (89.2%, n = 350), with a mean age of 39.54 (SD = 11.15; age range: 22–65). Overall, participants matched national demographics of nurses in the United States [58]. Twenty-six of the sample reported working at least one night shift (working between 9 pm and 6am) during the 14-day daily diary period. On average, nurses drank 0.57 (SD = 0.67) alcoholic beverages a day. In total, 15 nurses (3.8%) demonstrated probable PTSD (> 33 symptom score), with an average severity score of 9.12 (SD = 10.59) among the entire sample. For more information about participant characteristics, please refer to Slavish and colleagues [59].

Table 1.

Participant Characteristics for the Entire Sample

	Entire sample
	(n = 392)
	N (%)	Mean (SD)
Age		39.54 (11.15)
Gender (female)	360 (91.8)
Ethnicity (Hispanic/Latinx)	42 (10.8)
Race
White	305 (77.8)
African American/Black	26 (6.6)
American Indian/Alaska Native	6 (1.5)
Asian	41 (10.5)
Multiracial	7 (1.8)
Other	7 (1.8)
Marital Status
Married	248 (63.3)
Single	101 (25.8)
Divorced	33 (8.4)
Separated	7 (1.8)
Widowed	3 (0.8)
Health Conditions
Sleep Apnea	8 (2.0)
Heart Disease	2 (.05)
Cancer	5 (1.3)
High Blood Pressure	35 (8.9)
Breathing Problem	13 (3.3)
Diabetes	10 (2.6)
Chronic Pain	4 (1.0)
Autoimmune Disease	7 (1.8)
Menopausal Status
Pre	123 (31.4)
Peri (current)	53 (13.5)
Post	69 (17.6)

Daily Measures

Daily Alcohol Use

In the daily sleep diary, participants reported on the number of drinks consumed during the previous day using the item “How many alcoholic drinks did you have yesterday?” In which the amount of alcohol in a standard drink was described (e.g., 12 fl oz of beer, 8–9 fl oz of malt liquor, 5 fl oz of wine, 1.5 fl oz shot of 80-proof spirits). Daily alcohol use was dichotomized as 0 = no drinking and 1 = drinking (defined as consuming 1 or more alcoholic beverages) for the present study, given that on most days when nurses reported drinking (46.18%, n = 653 observations), they only reported consuming one alcoholic beverage.

Daily Actigraphy-Determined Sleep

For 14 days, participants were instructed to continuously wear an Actiwatch Spectrum Pro (Philips Respironics, Bend, OR USA) on their non-dominant wrist. The Actiwatch is a watch-like device used to infer objective sleep/wake patterns. Participants were asked to push an “event marker” button when they intended to go to sleep and when they got out of bed. Rest intervals were manually set in Actiware software (Version 6.0.8) by two trained individuals using a protocol that systematically relies on a combination of event markers, sleep diary data, activity data, and light levels [60]. Discrepancies between the two scorers in setting the rest intervals were resolved by a third person. Data were exported using default settings (10 immobile minutes for sleep onset and offset, medium wake threshold [40 activity counts]). Exported actigraphy data were used to determine total sleep time (i.e., total number of minutes in a rest interval that are scored as sleep by the sleep interval detection algorithm) and sleep efficiency (total sleep time, divided by time elapsed between the start and end time of a given rest interval, multiplied by 100), which were used in the current analyses.

Daily Self-Reported Sleep

An electronic version of the Consensus Sleep Diary—Core[61] was completed by participants each morning upon awakening using REDCap [62]. Diaries were used to determine total sleep time (time in bed [with the intention of sleeping] minus the sum of sleep onset latency, wake after sleep onset, and terminal wakefulness) and sleep efficiency (total sleep time divided by time in bed, multiplied by 100). Sleep onset latency and wake after sleep onset were examined separately as well. Participants were asked to rate their subjective sleep quality on a scale of 0 (very poor) to 4 (very good). Sleep diaries provide reliable and valid assessments of total sleep time and sleep efficiency and correlate significantly with actigraphy (rs = 0.36 to 0.60), EEG (rs = 0.18 to 0.63), and polysomnography (rs = 0.36 to 0.59) measures [63–65].

Daily Work Schedule

To classify work schedule, in the daily sleep diaries, participants reported whether they worked a night shift (“Did you have to be at work past 9 pm OR before 6 am?”), or another shift schedule (e.g., day shift, or were off work during the previous 24-h period). Night work schedule was used as a covariate in the present study’s analyses.

Baseline Measures

PTSD Symptoms

The PTSD Checklist for DSM-5 (PCL-5) is a 20-item self-report measure used to assess PTSD symptom severity in the past month [66]. We used the version of the PCL-5 that examines past month PTSD symptoms without assessing Criterion A trauma exposure [66]. The measure is summed to obtain a total score ranging from 0 to 80, with higher scores indicating greater symptom severity. A score ≥ 33 indicates a positive screen for PTSD (i.e., moderate to severe symptoms) [67]. The PCL-5 has good psychometrics [41]. In the current study, the PCL-5 demonstrated good internal consistency (α = 0.94).

Demographics

At baseline, participants reported on age, gender, ethnicity, and marital status. For more information regarding the larger study and sample, see Slavish et al. [59].

Statistical Analysis Plan

All analyses were conducted in the open-source statistical program R (R Core Team, 2013). Multilevel models were conducted using the R package lme4 [68], and tables were created using the R package sjPlot [69]. For all multilevel models, level 1 days were nested within level 2 people. Restricted maximum likelihood (REML) was used, which is a robust method for handling missing data using all available information to estimate the model [70]. Intercepts were allowed to vary randomly across people. All models controlled for daily work schedule (0 = any other schedule; 1 = night shift schedule [work between 9 pm and 6am]) given previous studies showing robust differences in sleep [71] and alcohol use [72] by work schedule. Alcohol use was dichotomized (0 = no drink; 1 = drink). For analyses examining sleep predicting subsequent alcohol use, sleep data were lagged back one day (as alcohol was reported in the morning, reflecting alcohol experienced the previous day). For analyses examining alcohol predicting subsequent sleep, data were not lagged (as previous day’s alcohol and previous night’s sleep were reported simultaneously). Holm-Bonferroni [73] sequential corrections were applied to p-values to adjust for multiple comparisons (20 analyses in total; false discovery rate = 15%) [74].

We examined the moderating role of baseline PTSD symptom severity on the bidirectional relationship between daily alcohol intake and sleep. An example equation for between-person (level 2) baseline PTSD symptoms moderating the within-person (level 1) slope between daily alcohol use on daily sleep (i.e., Aim 2) is shown below:

$$\mathit{\text{Level}}1\left(\mathit{\text{days}}\right):{\mathit{\text{Daily sleep}}}_{ij}={\beta }_{0j}+{\beta }_{1j}\left({\mathit{\text{Daily alcohol use}}}_{ij}\right)+r_{ij}$$

$$\mathit{\text{Level}}2\left(\mathit{\text{people}}\right):\left.{\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}\left({\mathit{\text{Baseline PTSD}}}_j\right)+{\mu }_{0j}\right)$$

$${\beta }_{1j}={\gamma }_{10}+{\gamma }_{11}\left({\mathit{\text{Baseline PTSD}}}_j\right)$$

where: γ₀₀ is the overall mean of daily sleep, γ₁₀ is the overall slope between alcohol and sleep, γ₀₁ is the overall effect of between-person PTSD symptoms on the overall mean of daily sleep, γ₁₁ is the cross-level moderation effect of between-person baseline PTSD symptoms on the within-person slope between alcohol and sleep, and μ_0j is the random deviations of the j^th person’s mean daily sleep from the overall mean.

Results

Descriptive Results

Point-biserial correlations for the study variables are provided in Table 2. For repeated measures variables, intraclass correlation coefficients (i.e., ratio of between-person variation to total variation) were 13% for self-reported total sleep time, 17% for actigraphy-determined total sleep time, 25% for self-reported sleep efficiency, 32% for actigraphy-determined sleep efficiency, 30% for self-reported sleep quality, 26% for self-reported sleep onset latency, 21% for self-reported wake after sleep onset, and 31% for self-reported alcohol use. Thus, the results suggest that there was more variation at the within-person level (i.e., from day-to-day) than at the between-person level (i.e., from person-to-person) for each of these constructs. In terms of missing data by measure, 0.07% (n = 366) of alcohol use observations were missing; 0.07% (n = 366) of daily sleep diaries were missing; 0.08% (n = 454) of actigraphy records were missing/not usable. We have now added this information to the methods.

Table 2.

Between-Person (14-Day Average) Correlations Among Key Study Variables

Variable	M	SD	Range	1	2	3	4	5	6	7	8	9	10
1. Actigraphy total sleep time	402.10	50.37	237.7–551.8
2. Actigraphy sleep efficiency	86.97	4.87	62.55–95.90	.41** [.32, .49]
3. Self-reported total sleep time	432.21	49.20	246.2–559.9	.81** [.77, .84]	.10* [.00, .20]
4. Self-reported sleep efficiency	91.07	5.10	65.48–98.85	.09 [−.01, .19]	.33** [.24, .42]	.32** [.23, .41]
5. Self-reported sleep onset latency	17.42	14.75	.92–152.25	.08 [−.02, .18]	−.26** [−.35, −.16]	−.09 [−.19, .01]	−.73** [−.78, −.68]
6. Self-reported wake after sleep onset	13.17	10.86	0–79.38	−.03 [−.13, .07]	−.12* [−.21, −.02]	−.11* [−.21,−.01]	−.56** [−.63, −.49]	.16** [.06, .25]
7. Self-reported sleep quality	2.57	0.57	0−4	−.03 [−.13, .07]	.15** [.05, .24]	.11* [.01, .21]	.42** [.34, .50]	−.37** [−.45, −.28]	−.11* [−.21,−.01]
8. Daily self-reported alcohol use (total)	0.57	0.67	0–16	.11* [.02, .21]	−.04 [−.14, .05]	.15** [.05, .25]	.04 [−.06, .14]	−.06 [−.16, .04]	.03 [−.07, .13]	−.04 [−.14, .06]
9. Self-reported alcohol use (drink)	0.29	0.45	0–1	.03 [−.07, .13]	−.00 [−.10, .10]	.09 [−.02, .19]	.04 [−.06, .14]	−.10 [−.20, .01]	.06 [−.04, .16]	−.00 [−.11,.10]	.57** [.49, .63]
10. Total PTSD severity	9.12	10.59	0–63	−.00 [−.10, .10]	−.09 [−.19, .01]	−.03 [−.13, .07]	−.22** [−.32,−.13]	.22** [.12, .31]	.15** [.05, .24]	−.22** [−.31,−.12]	.16** [.06, .25]	.02 [−.09, .12]
11. Shift work status (nights)	0.10	0.30	0–1	−.13* [−.23, −.02]	.01 [−.10,.11]	−.17** [−.27, −.06]	−.11* [−.21, −.00]	.09 [−.02, .19]	.00 [−.10,.11]	.02 [−.09, .13]	−.03 [−.14, .07]	−.15** [−.25, −.04]	.05 [−.06, .16]

M and SD are used to represent mean and standard deviation, respectively. Range includes the minimum and maximum value of the variables. Values in square brackets indicate the 95% confidence interval for each correlation. Self-reported alcohol is reported as total number of drinks and dichotomous drinking (0 = no drinks consumed that day, 1 = one or more drinks consumed that day). Shift work schedule was coded as 0 = any other schedule (days off and daytime work) and as 1 = night shift work. PTSD = posttraumatic stress disorder

^*p < .05.

^**p < .01

On 25.76% (n = 1414) of the total 5488 possible measurement occasions (i.e., 14 days × 392 participants), nurses consumed at least one alcoholic beverage. On 46.18% (n = 653) of the 1414 total drinking days, nurses consumed only 1 alcoholic beverage. Of the 1414 days where nurses reported consuming at least one alcoholic drink, only 178 (12.5%) of those observations were days where nurses reported consuming 4 or more drinks (i.e., heavier drinking). On most days, that nurses drank, they only consumed 1 or 2 alcoholic drinks (n = 1047 out of 1414 total drinking days, or 74% of all drinking days). On average, only 25% (n = 99) of the nurses in the full sample were classified as heavy drinkers vs. light or non-drinkers (i.e., > 7 drinks/week for women, and > 14 drinks/week for men). However, the heavy drinkers still drank relatively infrequently: they consumed an average maximum of 13.24 drinks (SD = 4.76) in any consecutive 7-day period, as opposed to light drinkers who consumed an average of 2.60 drinks (SD = 2.58) over the same period. 93 nurses (23.72%) did not consume any alcohol during the 14-day period. On average, participants completed 13.07 (SD = 1.6, median = 14, range = 4–14) out of the 14 possible daily surveys, for an average compliance rate of 93.4%. As expected, most actigraphy and sleep diary measures were moderately positively correlated. PTSD symptom severity was moderately correlated with self-report but not actigraphy measures of sleep. Total number of alcoholic drinks per day was weakly correlated with actigraphy and sleep diary total sleep time, but not other sleep variables.

Aim 1: Main Effects of Within-person, Bidirectional Associations Between Alcohol Use and Sleep

Alcohol Use Predicting Subsequent Sleep

See Table 3 for details. Days with alcohol use were associated with shorter self-reported sleep onset latency (b = −4.21, 95% CI [−5.96—−2.60], p = .003) and longer self-reported wake after sleep onset (b = 2.36, 95% CI [0.95–3.76], p = .009), covarying for daily work schedule. More specifically, nurses reported falling asleep faster by approximately 4.21 min and waking up later by 2.36 min on days they consumed alcohol. Compared to nurses who do not work or work a day shift (the reference group), night shift workers reported falling asleep faster by 5.39 min on days when alcohol was consumed. The Marginal R² for these models indicated that the fixed effects of alcohol use and daily work schedule predicted less than 1% of variance in subsequent sleep respectively (R² marginal = .010 sleep onset latency; R² marginal = .003 wake after sleep onset).

Table 3.

Effects of daily alcohol use on self-reported and actigraphy-determined total sleep time, self-reported sleep onset latency, and self-reported wake after sleep onset

	Self-reported total sleep time (min.)			Actigraphy total sleep time (min.)			Self-reported sleep onset latency (min.)			Self-reported wake after sleep onset (min.)
Predictors	Estimates	95% CI	p	Estimates	95% CI	p	Estimates	95% CI	p	Estimates	95% CI	p
(Intercept)	438.95	433.77–444.14	<.001	409.68	404.35–415.01	<.001	19.26	17.66–20.86	<.001	12.63	11.43–13.82	<.001
Self-reported alcohol use (drink)	15.59	8.34–22.84	.006	10.06	3.03 – 17.09	.037	−4.21	−5.96–−2.46	.003	2.36	0.95–3.77	.009
Daily work schedule (night shift)	−104.69	−115.77–−93.62	<.001	−94.88	−105.86–−83.89	<.001	−5.39	−8.09–−2.70	<.001	0.13	−2.03 – 2.29	.906
Random Effects
σ2	9889.55			8228.51			524.66			344.75
τ00	1289.20 ID			1565.83 ID			179.51 ID			93.57
N	392 ID			390 ID			392 ID			392
Observations	4797			4466			4797			4797
Marginal R2 / Conditional R2	.100/NA			.089 / NA			.010/NA			.003 / NA

Self-reported alcohol is reported as dichotomous drinking (0 = no drinks consumed that day, 1 = one or more drinks consumed that day). Shift work schedule was coded as 0 = any other schedule (days off and daytime work) and as 1 = night shift work. CI = 95% confidence intervals. For random effects: σ2 represents level 1 variance (within-person), τ00 represents level 2 variance (between-person), N is the total Level 2 sample size, and Observations are the number of Level 1 daily observations. ID = participant ID, the Level 2 grouping variable. Marginal R2 is the variance of the fixed effects only; Conditional R2 is the variance of the fixed and random effects. For random effects: σ2 represents level 1 variance (within-person), τ00 represents level 2 variance (between-person), ICC represents the intraclass correlation coefficient, N is the total Level 2 sample size, and Observations are the number of Level 1 daily observations

Bold values represent p < .05 estimates; p-values were adjusted for multiple comparisons using Holm–Bonferroni correction

Days with alcohol use were associated with longer self-reported total sleep time (b = 15.60, 95% CI [8.34–22.84], p = 0.006) and longer actigraphy-determined total sleep time (b = 10.06, 95% CI [3.03–17.09], p = .037) with the same covariate in the model. More specifically, nurses reported longer sleep times by 15.59 min, while actigraphy measured 10.06 min longer total sleep time on days when alcohol was consumed. Compared to nurses who do not work or work a day shift (the reference group), night shift workers reported shorter sleep times by 104.59 min on days when alcohol was consumed. The Marginal R² for these models indicated that the fixed effects of alcohol use and daily work schedule predicted approximately 9–10% of variance in total sleep time respectively (R² marginal = 0.100 self-reported total sleep time; R² marginal = 0.089 actigraphy-determined total sleep time). There were no significant associations between alcohol use and subsequent self-reported sleep quality, self-reported sleep efficiency, or actigraphy-determined sleep efficiency.

Sleep Predicting Subsequent Alcohol Use

When examining reverse associations at the within-person level, no associations were found between any of the sleep indices and next-day alcohol use.

Aim 2: Cross-Level Moderation by Baseline PTSD Symptoms

At the between-person level, baseline PTSD symptoms did not moderate within-person associations between alcohol use and any sleep indices, or vice versa (see supplemental material for nonsignificant models).

Discussion

This was the first study to investigate daily bidirectional associations between alcohol use and various sleep indices among nurses. The results indicated that only alcohol use predicted several sleep indices, however, not the reverse. These daily associations did not vary by an individual’s baseline levels of PTSD symptoms. Together, these results characterize the daily cycle of alcohol use and sleep in a sample vulnerable to high levels of occupational stress.

During the two week period, on average, 37% of nurses reported sleeping less than 7 h per night while the actigraphy-determined total sleep time indicated 61% of nurses did not obtain the recommended hours of sleep. In general, a smaller proportion of nurses reported sleeping less than 7 h in our study, compared to previous estimates based on nurses’ reports [5–8]. Nevertheless, the total sleep time determined by actigraphy does correspond with the suboptimal sleep durations found in previous research. Among nurses who did drink (76% consumed > 1 beverage during the two week period), most reported low levels of alcohol consumption (e.g., 1–2 drinks). This may have limited the generalizability of our results and our ability to detect significant effects. Other studies similarly show that ~ 30% of nurses report regularly consuming alcohol [14]; however, higher rates of problematic alcohol usage have been found [75].

In line with extant work, we found that alcohol use was associated with longer sleep duration [40, 76] and shorter time falling asleep [27, 77], but also more frequent wakefulness during sleeping time [28, 77, 78]. Previous polysomnography studies have generally found that alcohol improves sleep during the first half of the night, but disrupts it during the latter half [79, 80]. While these effects were relatively small, it’s possible that the adverse effects of alcohol use on sleep continuity may accumulate across time. If these patterns persist, they may increase an individual’s risk for developing insomnia and other negative health outcomes. For example, after experiencing transient alcohol-related sleep disturbances, an individual may adopt compensatory behaviors (e.g., napping, excessive caffeine use to stay awake) that further reinforce sleep difficulties. Alcohol may also have a greater negative impact on sleep particularly among lighter drinkers (e.g., 1–2 drinks on average) [77]. The distinction between light/moderate and heavy drinking, in relation to sleep health among nurses, warrants further investigation. Likewise, future research should examine if results replicate in other samples of nurses, particularly those endorsing higher levels of alcohol use.

Though sleep has been considered a risk factor for future alcohol use and related problems [81], we did not find support for this directional association at the daily level. A previous daily diary study found that sleep had reciprocal effects on alcohol use the following day, such that better self-reported sleep efficiency predicted heavier drinking the following night [82]. It is possible other factors are involved in the sleep-alcohol link, including affect. Sleep disruption has been associated with negative affect [34, 83, 84]; thus, it may influence motives for drinking alcohol (e.g., to cope) [85]. Based on qualitative findings from a study of nurses’ perspectives, alcohol is widely used to cope with emotional distress, particularly work-related stress [16]. The current study adds to the larger literature examining the alcohol use-sleep link by examining these associations among a large, understudied sample (i.e., nurses) over a longer time-frame (i.e.,14 days), increasing power and external validity of results. Future research should investigate potential non-linear relationships between sleep health and alcohol consumption, as this may provide a more comprehensive understanding of their complex interplay.

Contrary to our predictions, the results did not support the moderating role of PTSD symptom severity on the bidirectional associations between daily alcohol use and various sleep indexes. In other words, the links between daily alcohol consumption and sleep were similar regardless of the severity of an individual’s PTSD symptoms. Other studies have suggested that specific PTSD symptom clusters may be more closely associated with alcohol use and sleep. In particular, hyperarousal symptoms have been found to be positively associated with alcohol use and consequences [86], as well as sleep disturbances [87]. It is also likely that PTSD symptoms contribute to the alcohol-sleep relationship since individuals with PTSD have a tendency to cope with distress (i.e., self-medication hypothesis) [88] and sleep issues [89] through alcohol consumption. Indeed, excessive use of alcohol to mitigate symptomology increases the risk of worsening health problems [52]. In the current study, most nurses fell well below the probable PTSD cut-off score of 31–33 on the PCL-5 [66]. Only 15 nurses screened positive for PTSD, possibly contributing to the null moderation findings. Though nurses are generally at increased risk for developing PTSD, not all nurses who experience traumatic events develop PTSD. However, significant bivariate correlations were found between PTSD symptom severity and alcohol use, as well as several sleep indices (e.g., self-reported sleep efficiency, sleep onset latency, wake after sleep onset, and sleep quality). In order to support nurses’ well-being, additional research is required to better understand the nature and scope of these associations.

Limitations and Future Directions

There are several strengths of this study (e.g., large sample size; repeated measures over 14 days; corroborative objective sleep measures, within-person analyses); however, some limitations may warrant further investigation. First, the larger study was centered on vaccine responses in nurses; thus, the sample was not assessed for trauma exposure and included a small number of nurses endorsing elevated PTSD symptom severity and alcohol use. This may have affected our ability to determine: 1) whether sleep problems predicted alcohol use and 2) the interplay of PTSD symptom severity on bidirectional links between alcohol and sleep. Examining these links in nurses with clinical levels of symptoms may help clarify the interplay between PTSD symptoms, alcohol use, and sleep disturbances. It is also plausible that reported and actual alcohol consumption may differ, especially at higher doses when alcohol interferes with self-reporting. For instance, there is a possibility that some of the daily surveys were completed during work hours, potentially leading to underreported alcohol consumption. Typically, however, nurses have a higher level of health consciousness than the average population [90], which may also impact their substance use behavior. Future research should investigate the total number of drinks consumed, explore alternative methods of assessment (e.g., passive monitoring via transdermal devices), as well as additional substance intake including co-use, in order to gain a deeper understanding of the effects of substance use on sleep and vice versa. Consideration of specific motives for using alcohol (e.g., coping, sleep problems) may also shed light on the dynamic interplay between trauma, sleep, and alcohol use as it relates to the onset and maintenance of problematic drinking behaviors and sleep issues. Thus, inclusion of repeated measures of PTSD symptoms, affect/stress, and alcohol use motives should be included in future investigations in order to further understand the day-to-day interplay of these factors. To maximize generalizability, the larger study did not exclude individuals with sleep disorders (e.g., insomnia, obstructive sleep apnea), mental health conditions, or chronic health conditions. We did, however, assess and investigate insomnia and circadian rhythms symptoms using daily sleep diaries, which are arguably more reliable than a single-time point retrospective questionnaires [59, 91]. It may be important to investigate the other factors as moderators in future studies, as research shows chronic health conditions are related to increased substance use [92, 93]. Covarying of these variables (e.g., apnea) at baseline was considered; however, our sample endorsed very low rates of all health conditions. Furthermore, it is important for future research to investigate potential variations in workplace trauma and symptoms of PTSD based on different healthcare disciplines (e.g., preventative, intensive care units). Lastly, due to the predominantly white and middle-aged sample, this study lacked the necessary statistical power to adequately analyze potential variations in sleep patterns and alcohol consumption based on age and race/ethnicity. Therefore, it is imperative to consider these factors in future research to delineate these patterns.

Conclusions

The current study provides support for the negative impact of alcohol use on sleep continuity in nurse’s daily lives. More specifically, we found that although alcohol may help facilitate sleep onset and lead to longer sleep duration, it may also lead to more fragmented sleep in nurses. The development of preventative programs should address the lack of knowledge about the effects of alcohol on sleep architecture and promotion of responsible alcohol use (e.g., use of protective behavioral strategies) [94]. Future researchers should consider other factors (e.g., affect, stress) in addition to including more intentional sampling of nurses who experience symptoms of PTSD daily to elucidate the link between alcohol use and sleep. Nursing plays an essential role in healthcare; therefore, maintaining nurses’ health is crucial to their well-being and the delivery of patient care.

Data Availability

Materials and data for this study are available by emailing the corresponding authors.