The Negative Impact of Night Shifts on Diet in Emergency Healthcare Workers

We performed a prospective nationwide observational study in five French hospitals—two university hospitals and three non-university hospitals. The main inclusion criterion was to work as an emergency HCW. Exclusion criteria were refusal to participate and pregnancy. This study is part of the SEEK protocol [20,22]. We obtained the ethical approval from Ethics Committee South-East I (DC-2014-2151) and the protocol was registered on ClinicalTrials.gov as number NCT02401607↗. Each participant had their food intake monitored twice over 24 consecutive hours: (1) during a day shift (from 8:30 a.m. to 6:30 p.m.) + the night (no work, 6:30 p.m. to 8.30 a.m.), and (2) during a rest day (from 8:30 a.m. to 6:30 p.m.) and a night shift (6:30 p.m. to 8:30 a.m.) (Figure 1). Participants also had to complete a questionnaire on their sociodemographic characteristics.

Participants were asked to complete a 24-h dietary recall that was explained to them by a member of the investigation team. The participants were asked to indicate as precisely as possible all the details regarding the food ingested at each meal and in-between meals. The diaries were reviewed afterward with the participants during a dedicated interview. We next used Nutrilog^® (version 3.2, Nutrilog, Marans, France), a diet and nutrition software for health care professionals, to translate participants’ answers on nutrients intake using nutrition table Ciqual^® on this software [26]. We retrieved total energy intake (kcal), glucids (grammes and % energy intake, and sugar), lipids (grammes and % energy intake, and simple, mono- and poly-unsaturated fatty acids), protids (grammes and % energy intake), calcium, cholesterol, fibres, and water. We further looked at time of intake and more specifically we calculated number of emergency HCWs that did not eat or drink for more than 4, 8, and 12 consecutive hours.

Participants had also to complete a short questionnaire to retrieve their sociodemographic (age, gender, marital status) and work characteristics (occupation, job seniority, and setting—university hospital or not).

The main judgment criterion was food intake as quantitative variables. From personal experience and observation in a preliminary non-published study among 10 emergency HCWs, there is at least a 20 ± 20% decrease in food intake during a night shift compared to a day shift, because of the increased working demands during a night shift due to fewer staff. Using this difference as the main outcome, we calculated that a sample size of 10 participants allowed a statistical power greater than 80% with an alpha level less than 5%.

Statistical procedures were performed with Stata software (v17, College Station, TX, USA). Quantitative variables were expressed as mean ± standard deviation (SD), and qualitative variables were expressed as number (%). The Gaussian distribution for each food intake variable was assessed by a Shapiro–Wilk test. Mean levels of each 24 h food intake variable were compared between night and day shifts using Wilcoxon (signed rank) tests. Association between food intake, sociodemographic (age, gender, marital status) and work characteristics (occupation, job seniority, and setting—university hospital or not), were assessed using Chi2 for categorical variables, and Spearman correlations for quantitative variables. The influence of shift condition, sociodemographic and work characteristics with each food intake variable were further assessed using univariate regressions—linear regressions for quantitative variables (coefficient and 95 confidence intervals—95% CI) and logistic regression for qualitative variables (odds ratio and 95% CI). Significance was set at the p < 0.05 level.

Emergency HCWs working during a night shift have a slighter 24-h energy intake compared to those working during a day shift, but also compared to daily recommended diet [27]. Literature is not consistent about these results. Although we did not find any study about the impact of night shifts among emergency physicians, there is some among nurses with controversial results. Some found a lower energy intake among night workers versus day workers [28,29] while others did not find any significant differences [30,31,32]; however, another study found higher food intake among night workers versus day workers [33]. Although we do not study the energy intake the day after the shift, it seems that this low food intake is compensated for on the day of recovery after work [29]. We can explain this low food intake in our study. First, emergency HCWs are subjected to daily stress due to overcrowding of EDs [34], lack of beds available in the hospital, and life-threatening emergencies [35,36]. However, the impact of stress on eating is controversial. Indeed, while acute stress is able to induce dietary restriction, chronic stress will induce food intake signal thus providing a better response to future stressful events [37]. Many pathophysiological pathways have been studied, especially the impact of ghrelin and leptin as biomarkers of acute stress [38,39,40]. It would be interesting to study those biomarkers among emergency HCWs. Secondly, sleep restriction and disturbed nychthemeral rhythm promote increased energy intake [19,41].

The quality and choice of food during the night shift differs from the day shift [19]. This could induce restrictive behavior. Our results are in accordance with this. Indeed, emergency HCWs eat fewer carbohydrates and lipids during night shifts than their daytime colleagues [28,42]. Those results are also in agreement with the literature. Furthermore, it seems that the main carbohydrate intake during night shifts include the excessive consumption of sweets and sugary drinks [43,44,45]. Concerning lipids, night workers, and especially permanent night-shift workers, are at risk of dyslipidemia [9]. However, in our study, emergency HCWs eat less lipid and fat during night shifts than during day shifts. This is in line with literature [28,29,43]. Another possible explanation is the lack of time to eat during both day and night shifts. It seems that this lack of time leads to an increase in consumption of cold foods and fast foods [31,44,45], which predisposes individuals to obesity and cardiovascular disorders [46,47]. We also showed a low protein consumption during night shifts—type of position held, overwork and the quality of sleep are risk factors for low protein consumption [27,47]. Our results showed a low consumption of water consumed by emergency HCWs. Some of them do not even consume fluids during the entire night shift, i.e., during more than 12 h [28,48]. It seems that coffee/tea and sugary drinks are the preferred drinks mostly consumed during night work [28,33,49].

We showed a link between age, emergency HCWs’ position held and the consumption of lipids, fat, and cholesterol. The more experience they have, the more lipids, fat, and cholesterol they consume. This is in line with a study that shows that eating behavior is influenced by seniority, and age [50]. It seems that experienced workers, with staggered schedules, eat more during the night, but mainly snacks rich in sugar, fat, and energy [50,51]. We also found that males have a higher energy, cholesterol, and protein intake, but also are less able to spend more than 4 h without eating. This finding is not in agreement with the literature. Indeed, night work induces an increase in the consumption of sugary and fatty foods, possibly in response to stress among emotional eaters, who are most often women [52,53]. University hospitals are larger than non-university hospitals, so we could imagine that emergency HCWs have less time to eat because ED is more overcrowded. However, we found that working in a large health center would have a positive influence on the diet of emergency HCWs. According to the literature, night workers in large hospital centers tend to consume more “junk food”, which is easily accessible and more convenient to eat [28]. Our study is the first to examine the influence of night work on the diet of emergency physicians. Indeed, the subject often explored nurses or even more often industries in Asia with a population having a very high energy expenditure and a very low socio-economic level. Most of these studies have a low level of evidence (non-representative samples, unsystematic literature reviews) and are difficult to transpose in our population due to very different eating habits and high standards of living [19]. Another potential variable could be the sleep that we did not study the impact of. Indeed, previous literature suggests that night-shift workers obtain less sleep compared to day-shift workers. Thus, if night-shift workers have a longer duration awake, they will be hungrier, and so more intake [54].

Our study has some limitations. Firstly, the analyses of the dietary data were not necessarily done in the same subjects for day and night as we had more day shifts and fewer night shifts. Secondly, we worked on questionnaires that utilized self-reported information that may have contained some imprecision in the description of the food portions consumed as well as over- or under-declarations on the part of the participants completing the questionnaires. Although our data were collected via a self-assessment questionnaire with self-reported responses, the general methodological quality of our study was good and the risk of bias across participants was reduced through the use of a validated scale presented to every emergency HCW before the beginning of the data collection session. Another limitation is that we did not study the impact of cognition. Indeed, the impact on memory in case of a decrease of attentiveness or responsiveness could impact negatively the ability to fill in the questionnaire. By and large, while quality assurance protocols were put in place for the data collection process, it was difficult to micro-manage the different centers. Lastly, strict inclusion criteria were applied in order to extract data from a targeted population of emergency HCWs and thus answer our research questions. Continuous monitoring of food intake may be relevant and mandatory in future research. On the other hand, few studies consider the diet of emergency HCWs, and when this is the case, it concerns nurses [55]. Our study is therefore the first to look at the nutrition of emergency physicians during night shifts. Unfortunately, we did not study the food intake the day following the night shift. We previously showed that a night shift has a prolonged effect on some outcomes such as biomarkers of stress [23,24]. Further studies assessing the food intake during rest after a night shift could be relevant. Finally, very few studies have looked at the impact of night work on emergency physicians [24]. It would, therefore, be interesting to conduct more prospective and interventional studies to try to explore all the underlying mechanisms and subsequently put in place health-promotion and/or preventive strategies.

All relevant data are in this manuscript.