Acute and long‐term sleep measurements produce opposing results on sleep quality in 8 and 12 hour shift patterns in law enforcement officers

The occupational demands of law enforcement increase the risk of poor‐quality sleep, putting officers at risk of adverse physical and mental health. This cross‐sectional study aimed to characterise sleep quality in day workers, 8 and 12 h rotating shift pattern workers. One hundred eighty‐six officers volunteered for the study (37 female, age: 41 ± 7). Sleep quality was assessed using the Pittsburgh sleep quality index, actigraphy and the Leeds sleep evaluation questionnaire. The maximal aerobic capacity (VO2max) was measured on a treadmill via breath‐by‐breath analysis. There was a 70% overall prevalence of poor sleepers based on Pittsburgh sleep quality index scores, where 8 h shifts exhibited the worst prevalence (92%, p = 0.029), however, there was no difference between age, gender, or role. In contrast, 12 h shifts exhibited the poorest short‐term measures, including awakening from sleep (p = 0.039) and behaviour following wakefulness (p = 0.033) from subjective measures, and poorer total sleep time (p = 0.024) and sleep efficiency (p = 0.024) from the actigraphy. High VO2max predicted poorer wake after sleep onset (Rsq = 0.07, p = 0.05) and poorer sleep latency (p = 0.028). There was no relationship between the Pittsburgh sleep quality index scores and any of the short‐term measures. The prevalence of poor sleepers in this cohort was substantially higher than in the general population, regardless of shift pattern. The results obtained from the long‐ and short‐term measures of sleep quality yielded opposing results, where long‐term perceptions favoured the 12 h pattern, but short‐term subjective and objective measures both favoured the 8 h pattern.

. Poor sleep quality in occupations such as law-enforcement has been associated with an increased risk of work-related injuries, higher job stress, and poorer mental health (Costa, 2015;Rajaratnam et al., 2011). The deleterious effects of poor sleep quality have far reaching implications for the long-term health prospects of police officers, a population already at an increased risk of stress-related physical illnesses such as cardiovascular diseases (Garbarino et al., 2019) and metabolic stress syndrome (Yong et al., 2016). Poor sleep quality is also a risk factor for metabolic syndrome and obesity and impacts lifestyle behaviours which influence cardiovascular health (Garbarino & Magnavita, 2015;Mota & Vale, 2009). The cyclical relationship between poor sleep quality and job stress has been observed to impact severely on health variables (Gerber et al., 2010;Lin et al., 2012). A possible mediating factor in this relationship could be exercise, since cardiorespiratory fitness and sleep have been shown to positively influence each other in healthy individuals (Strand et al., 2013), but not in elite athletes (Nedelec et al., 2018).
Few data sets exist on sleep quality in law enforcement, none of which have investigated its relationship with fitness.
Law-enforcement officers (LEOs) play a pivotal role in society, often facing high pressure scenarios where they must make rapid decisions during all hours of the day. The need to cover this 24 h period has led to the development of rotating shift schedules, which require officers to work at times misaligned with their internal circadian rhythms (Fekedulegn et al., 2016;Ma et al., 2019).
Given the unavoidable nature of their working hours, it is important to work within the constraints of their occupational requirements to optimise sleep quality in police officers. There is a large body of research concerned with how shift patterns can be manipulated to provide better outcomes for workers. For example, traditionally shift patterns rotated counter-clockwise (morning-nightafternoon), whereas now that research shows delaying sleep is easier than advancing it, many studies recommend clockwise rotation schedules (morning-afternoon-night) (Burgess, 2007). There have also been debates on the merits and disadvantages of different shift lengths.
Twelve-hour shift schedules mean that workers can have a compressed work week with a greater time to rest. However, longer shift lengths may lead to lapses in concentration and may therefore be dangerous for officers. Therefore, the aims of this study were: (1) to examine the effects of different shift patterns on sleep quality, specifically comparing day only, 8 h and 12 h night shift patterns; and (2) To examine the relationship between sleep and maximal aerobic capacity (VO 2max ) among LEOs.

| Participants
City based law-enforcement officers across two different locations in the United Kingdom were recruited through an email internal to their organisation. Two hundred fifty-two officers volunteered to be a part of the study, of which 186 participants (age: 41 ± 8, 137 male) completed the PSQI, while 64 of these (age: 40 ± 8, 44 male) completed the Leeds sleep evaluation questionnaire (LSEQ) every morning for at least 3 shift days and 54 (age: 42 ± 8, 37 male) completed the continuous assessment with actigraphy. The number of participants who completed each measure varied (Table 1). Participants were included in the study if they were working in law enforcement in the United Kingdom, did not present any cardiovascular or neurological conditions, and were above 18 years of age. No formal assessment of sleep disorders was made, however, participants were asked for details of use of sleeping pills and pre-existing medical conditions. Four participants recorded using over the counter "Nytol" and one reported melatonin tablets "occasionally" or "once weekly". Ethical approval was granted by UCL Ethics Committee (13985/004) in line with the declaration of Helsinki.
All participants provided informed consent prior to taking part. This study was registered on clinicaltrials.org (13985/004) and given the ClinicalTrials.gov Identifier: NCT04204486.

| Study design
This study was originally designed as a randomised control trial to investigate sleep quality and changes in sleep quality following an exercise intervention programme. Due to interruption of the study due to the COVID-19 pandemic at the time of post-testing, we provide a cross-sectional analysis of the factors affecting sleep quality from the participants' baseline data, and any associations between sleep quality and physical fitness. These data were collected as follows: Participants were emailed the Pittsburgh sleep quality index (PSQI) upon recruitment to record their long-term subjective perception of sleep quality.
They then attended the Institute of Sport, Exercise and Health (ISEH) to collect anthropometric and fitness data. To record acute objective measurements of sleep quality, accelerometers were then provided at the participants' workplace and activated in person by researchers.
Participants were asked to wear the accelerometer for 8 consecutive days to cover a full cycle of shift days and rest days, and to provide

| Participant characteristics
Baseline characteristics (height, weight) were measured before the study began and used to calculate body mass index (BMI). Physical fitness was tested through a VO 2max test on a treadmill, using the Bruce protocol (Bruce, 1971) which consisted of 3 min incremental stages on a treadmill (h/p/cosmos, Nussdorf, Germany), at each stage the inclination and speed of the treadmill increased. The participant progressed between stages and was verbally encouraged to reach volitional exhaustion. Maximal oxygen uptake (VO 2max ) was measured via breath-by-breath analysis through the Vyntus CPX Metabolic Cart (Vyaire Medical, Chicago, USA). Participants were then categorised into percentiles with a fitness grading based on their age and gender following the American College of Sports Medicine Guidelines (Ferguson, 2014). VO 2max was chosen as the gold standard for cardiopulmonary fitness (Strand et al., 2013).

| Sleep measurements
2.4.1 | Subjective long-term assessment -PSQI The PSQI is a widely used tool for the self-assessment of sleep quality over the previous month (Buysse et al., 1989). It provides a global sleep score between 0 and 21, calculated from adding seven component scores together, with higher scores indicating a worse subjective sleep quality. Global scores equal to or above 5 are considered a sensitive and specific measure indicative of poor sleep quality in adult populations (Buysse et al., 1989).

| Statistical analysis
A Kolmogorov-Smirnov test confirmed that all data were not normally distributed, therefore non-parametric tests were implemented. Chi-squared  Table 3.

| Subjective long-term perception
Of the participants who completed the PSQI, 130 (70%) were categorised as poor sleepers (

| Relationship between subjective and objective sleep measures
Of all LSEQ and actigraphy outcomes, only TST was significantly correlated with BFW (Rho = 0.35, p = 0.013) ( Table 4). The PSQI global score was not associated with any of the subjective or objective acute sleep measures. Normalised total scores of the three methods ( Figure 4) showed that 12 h scored more poorly than the 8 h group on the LSEQ (p = 0.024) and actigraphy (p = 0.019) but scored better on the PSQI with marginal significance (p = 0.048).  Austrians aged above 15. The vast difference in prevalence of poor sleepers between the populations could be due to higher levels of stress and work pressure present in the LEOs compared with other occupations (Garbarino et al., 2019;Gerber et al., 2010;Ramey et al., 2012), as well as the demands of rotating shift work that the majority of the LEO population engage in. The higher prevalence in the present study compared with other LEO studies could be due to the sample selection, where 52% of volunteers in this study worked night shifts. Nonetheless, LEOs on days only showed a 67% prevalence of poor sleepers, which is still significantly higher than the general population. The high levels of stress and work pressure present in law enforcement, in all shift patterns, is likely to be a strong contributor to poor sleep in this population, far beyond their working schedule (Garbarino et al., 2019;Gerber et al., 2010;Ramey et al., 2012).

| Comparison between shift patterns
The present study included two types of night shift patterns: 8 and 12 h forward rotating shifts (Table 2). Both groups of officers on the rotating shift schedule had fast-rotating cycles of the same length (8 days). Rotating shift work can be classified as clockwise (morningevening-night) or counterclockwise (morning-night-afternoon) in the direction shifts change (Burgess, 2007). The nature of the freerunning human biological clock has a cycle of roughly 25 h when external environmental cues are removed (Escames et al., 2011). Thus making phase-delay (delaying sleep) easier than phase-advance (attempting to sleep earlier) (Åkerstedt, 1998) which favours clockwise shift patterns. 12 h patterns cannot be classed as clockwise due to the rapid swing from day to night shifts rather than a gradual phase delay of morning-afternoon-night.
This study presents contradictory results regarding the impact of To confirm these incongruences, the correlation matrix (Table 4) shows no relationship between the PSQI score and any of the acute measures.
While the validation of sleep measures is beyond the scope of this paper, it is important to note that this finding warrants further research on how sleep quality is measured in shift workers.
The current literature favours 12 h shift patterns, as they are suggested to allow a greater rest period within the shift cycle (one-third longer than 8 h shift patterns) (Costa, 2015). After more than one night shift, 3 days are necessary for recuperation in order to fully overcome the impairment in alertness caused by the shift in the sleep-wake cycle (Burgess, 2007). This might explain why the 12 h group yielded better PSQI scores, where participants are asked to provide an overview of 30 days of sleep quality rather than an immediate recall of sleep after a shift. The 12 h shift in this cohort included a 4-day rest period after night shifts, which might be enough to aid recovery from sleep loss (Table 2). This extended rest period may also contribute to the officer's welfare due to their ability to socialise more, spend more time with family, and to engage in hobbies between rosters (Gerber et al., 2010). A major complaint of shift workers across many professions is social isolation due to job-enforced self-exclusion from social activities (Costa, 2015), therefore, by providing longer breaks between shifts, officers on the 12 h shift might feel that their well-being is being impacted less than those on 8 h shifts who only have 2 days of rest every 6 days of work.  (Åkerstedt, 1998).
Further research is needed to investigate the long-term effects of either shift pattern on general health outcomes.

| Cardiorespiratory fitness
Cardiorespiratory fitness was the only variable that showed weak but consistent relationships across both long-term and acute measurements. In this study, officers with higher VO 2max ratings generally exhibited poorer sleep outcomes. They found it harder to fall asleep according to the PSQI; this was then confirmed through actigraphy, where a higher VO 2max was associated with more waking minutes during the night (WASO), regardless of age, gender, and shift pattern.
These results ( Figure 5) indicate that officers with higher fitness levels perceived that it was harder for them to fall asleep and exhibited more minutes of wakefulness after sleep onset based on actigraphy measurements. This was an unexpected result, but not inexplicable.
Sleep quality has largely been associated with better fitness (Lee & Lin, 2007;Mota & Vale, 2009;Strand et al., 2013). However, similar relationships to the one found in the present study, particularly in relation to WASO, have been echoed in studies on elite athletes (Nedelec et al., 2018). For example, in a systematic review, Gupta et al. (2017) found that prevalence of sleep disorders in athletes ranged from 13% to 70%. A potential explanation for this, especially for those with "excellent" fitness could be related to the acute and chronic stresses which can manifest in the variability of sleep quality of highly trained athletes (Nedelec et al., 2018). Long and intense periods of exercise can lead to inadequate recovery sleep sessions (Driver & Taylor, 2000). The overtraining principle manifests in the sleep pattern, leading to increased fragmentation (Härmä, 1996), as reflected by increased WASO scores. Another potential explanation for this is the "ceiling effect" (Youngstedt, 2003), where athletes and very fit individuals have exhausted the possible sleep-related adaptations that come with improved fitness, and other factors may more strongly influence their sleep quality such as stress or nutritional factors (Nedelec et al., 2018). This latter explanation may be true for the LEOs in this study. The fitter volunteers in this study, regardless of shift pattern, were vastly specialist LEOs, who are a select group of individuals required to maintain a higher fitness standard. These officers are typically older, hold more stressful operational responsibilities and are more experienced, which may also imply having experienced greater exposure to trauma. A combination of these factors with the requirement to maintain high fitness standards could therefore be impacting the quality of sleep of these officers. Tailored fitness strategies designed around the occupational demands and stressors of law enforcements could help mitigate these effects.

| Strengths and limitations
This study addressed the gap for objective measurements in sleep quality in LEOs and combined a variety of measurements (objective and subjective, short and long term) to provide insight into the effects of differing shift patterns. Another strength of the study was the inclusion of both genders, which is often lacking in the literature.
A limitation is the compliance with the LSEQ and actigraphy continuous measurement which varied among participants. Participants did not always fill the daily questionnaire leading to missing data for some participants, which made it challenging to determine sleep times in the actigraphy data in some cases. This study could have also benefited from taking into consideration the napping habits and caffeine intake of the participants. Garbarino et al. (2004)

| Implications and recommendations
These results do not present an obvious answer as to which shift pattern is better for sleep quality in LEOs. However, it is clear that there is a high prevalence of poor sleep among LEOs. There are many strategies officers can employ to improve this, such as strategic napping (Åkerstedt & Torsvall, 1985;Bonnefond et al., 2001) and anchor sleep (Burgess, 2007). The present study clearly supports the consensus in current PSQI literature that officers need a minimum of 3 rest days (Burgess, 2007) to recover from significant changes in shift pattern given that the 12 h shift type were allowed 4 days of rest, whereas the 8 h shift had 2 days of rest between shift cycles. The study also supports the need for education on sleep hygiene practices in order to optimise sleep quality (Garbarino et al., 2019;Zee & Goldstein, 2010). Garbarino et al. (2020) reported that promoting sleep health awareness in police officers led to a significant improvement in the quantity and quality of sleep reported. In the present study no formal assessment of sleep knowledge or disorders was conducted which could give a better clinical analysis. This is pertinent considering a previous study (Garbarino et al., 2004) found shift work may exacerbate intrinsic sleep disorders.

| CONCLUSION
This study reports a very high prevalence of poor sleepers among law enforcement workers, regardless of shift pattern. Long and short-term measurements provided opposing findings regarding shift outcomes.
The long-term perception of sleep quality, measured via the PSQI, indicated that the 8 h shift pattern produced the highest prevalence of poor sleepers, where this group scored worst on general sleep quality and efficiency, resulting in poorer mood during the day. However, LSEQ and actigraphy measured throughout a full shift cycle indicated that officers on the 12 h shifts instead had poorest objective sleep efficiency and felt less alert when awakening. Finally, LEOs with high cardiorespiratory fitness were more likely to exhibit more fragmented sleep, possibly due to multiple accumulated stressors. Therefore, the method implemented for measuring sleep quality appears to have substantial implications on study outcomes; the incongruence between the long-and short-term measures reported here warrants further research into how sleep quality is measured in shift workers in order to determine appropriate shift patterns to safeguard officer well-being.

AUTHOR CONTRIBUTIONS
All authors contributed to the study design and reviewed the

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.