Morning-evening type and burnout level as factors influencing sleep quality of shift nurses: a questionnaire study

Aim To assess the relationship between sleep quality and demographic variables, morning-evening type, and burnout in nurses who work shifts. Methods We carried out a cross-sectional self-administered study with forced choice and open-ended structured questionnaires – Pittsburg Sleep Quality Index, Morningness-eveningness Questionnaire, and Maslach Burnout Inventory. The study was carried out at Gazi University Medicine Faculty Hospital of Ankara on 524 invited nurses from July to September 2008, with a response rate of 89.94% (n = 483). Descriptive and inferential statistics were applied to determine the risk factors of poor sleep quality. Results Most socio-demographic variables did not affect sleep quality. Participants with poor sleep quality had quite high burnout levels. Most nurses who belonged to a type that is neither morning nor evening had poor sleep quality. Nurses who experienced an incident worsening their sleep patterns (P < 0.001) and needlestick or sharp object injuries (P = 0.010) in the last month had poor sleep quality. The subjective sleep quality and sleep latency points of evening types within created models for the effect of burnout dimensions were high. Conclusions Nurses working consistently either in the morning or at night had better sleep quality than those working rotating shifts. Further studies are still needed to develop interventions that improve sleep quality and decrease burnout in nurses working shifts.

Some studies investigated how circadian type influenced the performance of individuals (14). The circadian types are morning-type, evening-type, and intermediate type.
Morning-types go to sleep early and wake up early, while evening-types are active during the night and cannot get up early (15, 16). The shift of the nurses should be adjusted to their living and especially sleeping habits, which is unfortunately currently not the case.
Burnout is another important problem for nurses working shifts (17). It is a psychological response to chronic emotional and interpersonal job-related stressors, and is defined by the three dimensions -emotional ex-haustion, depersonalization, and reduced personal accomplishment (18,19). It seems that burnout and sleep quality may affect each other. Although there is no sleep study on nurses, Grossi et al have recently demonstrated the relationship between burnout scores and sleeplessness in women (20).
The aim of this study was to assess the relationship between sleep quality and demographic variables such as arterial blood pressure, body mass index, heart beat rate, tea and coffee consumption, the relationship between sleep quality and morning-evening type, and the relationship between sleep quality and burnout in shift-worker nurses.

Design and setting
A cross-sectional study was carried out between July and September 2008 by a self-administered, forced choice, and open-ended structured questionnaire at Gazi University Medicine Faculty Hospital, Ankara, Turkey. The research population comprised all 637 nurses working in the hospital with a bed capacity of 1085. We excluded the nurses diagnosed with psychotic, neurological, metabolic, and sleep disorders, as well as nurses who either worked less than one year or who were on vacation or on medical leave during July and September 2008. Therefore, 524 nurses were invited to participate, but 26 refused and 15 did not complete the questionnaires, which left 483 nurses who were interviewed, with an overall 89.94% (483/524) response rate.

Participants
The study included 483 female nurses (

Pilot trial
For pilot trial, the first part of the question-form was given to 20 nurses who worked in another hospital and were not a part of the main research population. The pilot trial provided a test of comprehensibility and clarity of the questions, and based on it, self-administered, closed-ended, structured questionnaire interview was revised.

ethical considerations
Hospital ethics committee did not require an ethical approval for the study since it included no invasive practices for humans or animals. Instead, hospital management provided written approvals. The author contacted the chief nurse in each ward, explained the purpose of the study, and obtained a verbal permission. The participants signed the informed consent, after they had been informed in detail on characteristics and aim of this study.

Data collection
We collected data from nurses by using 4 anonymous self-administered questionnaires. Personal Information Form with forced choice and open-ended questions was designed by the authors to obtain information about demographic, socio-economic, individual, and work-related issues that were supposed to have effects on the scores of Pittsburg Sleep Quality Index (PSQI), Morningnesseveningness Questionnaire (MEQ), and Maslach Burnout Inventory (MBI) ( Table 1). Height, weight, arterial blood pressure and heart beat rate measurements of participant nurses were recorded in the first visit. Then, nurses were asked to fill in the Personal Information Form, PSQI, MEQ and MBI questionnaires, which were collected at the end of the shift. However, the questionnaires from the nurses who were not able to turn them in because of workload were collected as soon as completed, some even in the same shift. and inferential statistics. Statistical power was strengthened by defining poor sleep quality as a score of ≥5 on PSQI (21,22). Comparisons between socio-demographic variables, nursing work characteristics (ie, years employed in nursing, status of nurses, and work pattern), age, BMI, arterial blood pressure, tea/coffee drinking habit, needlestick injures in the last month, MBI, and morning-evening types were carried out by independent sample t-tests for continuous variables and Pearson χ 2 tests for categorical variables. Mann-Whitney U statistic was used as nonparametric test (Table 1).

Instruments
We tested contributions of demographic characteristics, work related issues, burnout level, and morning-evening type variables to PSQI by binary logistic regression analysis (with enter method, entry criteria P ≤ 0.05). The dependent variable was sleep quality ( Table 2). Linear regression analysis (adjusted for EE, DP, and PA) was conducted in order to investigate responsive components to individual morningevening types of the nurses. Before that, PSQI components were separated to ascertain the effect of morning-evening type for each component after dummy variables had been created for MEQ (Table 3). Pearson r bivariate correlations (two tailed) among PSQI and MBI subscale scores were calculated after the data had been split among continual day shift, night shift, and rotating shift (Table 4).

ReSUltS
The mean ± standard deviation global PSQI value of nurses was 7.32 ± 3.42 and the global PSQI of 79.1% (n = 382) of nurses was ≥5. The mean ± standard deviation total sleep time was 6.95 ± 0.99 hours. Most of the basic socio-demographic variables did not affect sleep quality (Table 1). EE (P < 0.001) and DP (P < 0.001), as well as PA decreased sleep quality (P = 0.001). Neither MEQ type also decreased sleep quality (P = 0.016). The events experienced in the last month affecting the sleep pattern (P < 0.001), sleepiness during the shift (P = 0.004), sharp object and needlestick injuries (P = 0.010), and lateness or failure to wake up in time were associated with sleep quality (P < 0.001) ( Table 1).
Emotional exhaustion, lateness to the work, disturbed sleep pattern in the last month, and having no roommate or no bed-partner were significantly associated with poor sleep quality ( Table 2).
Excluding the effect of EE, DP, and PA on sleep quality, we aimed to determine the effect of morning-evening type on PSQI. In order to investigate this, dummy variables were formed according to MEQ and the effect of these variables on PSQI components was identified using linear regression method after having made the necessary corrections according to EE, DP, and PA. In the models created by considering the effect of EE, DP, and PA, subjective sleep quality and sleep latency points of evening types were high ( Nurses working consistently either in the morning or at night had better sleep quality than those working rotating shifts (Table 4). Nurses working consistently night shifts had better sleep quality than all others.

DISCUSSIon
Most nurses in this study (79.1%) experienced poor sleep quality. This was expected since most of them (65.2%) worked rotating shifts. The finding is in accordance with previous studies (27)(28)(29)(30).
Nurses with poor sleep quality had higher burnout levels, experienced something that affected their sleep quality in the last month, were late for work because they could not awake up in the morning, had significantly more sharp object/needlestick injuries, and were sleepy at work. Alimoğlu and Dönmez (31) also found a higher burnout level in nurses with sleep disorders. The strongest predictor of sleep quality in our study was the participant's natural morningness-eveningness sleep pattern rather than work pattern.
Nurses working rotating shifts experienced more sleeping problems and sleepiness at work than nurses working continuous day/night shifts. Furthermore, nurses in rotating shifts had more accidents or errors. Therefore, rotating shifts should be avoided to assure safety of nurses and patients (32).
Sleep quality and work pattern had a low but significant correlation, but the MEQ type was not clearly associated with poor sleep quality. Admi et al showed that work pattern did not affect lateness, absence, or accidents and errors at work (33). On the other hand, some other studies indicated that sleeplessness or poor sleep quality increased work accidents (34,35). Different results on sleep quality were reported because of factors influencing biopsychosocial integrity of respondents that might affect sleep quality, such as stress and anxiety.
Healthcare workers in the night shifts or those who worked 60 hours a week were found to have higher risk of sharp object or needlestick injuries (36,37), and higher risk errors or near errors (38). Similar results were also found in this study.
Most nurses in the study had difficulty to sleep after the night shift. Bad sleep quality led to twitching in legs, adaptation problems, and confusion during sleep. Other studies also showed that nurses working rotating or night shifts were more tired than others (39) and that sleepiness decreased after they had started working morning shifts only (9).
Our study showed that emotional exhaustion, any incident influencing sleep pattern in the last month, and having no roommate or having no (bed) partner affected sleep quality. Such a finding in individuals with no roommate or bed partner might be explained by their feeling lonely and in need of emotional support. Exhaustion, on the other hand, decreases or diminishes self-confidence and interest in work, as well as causes fatigue and weakness (24). An in- dividual with chronic exhaustion might lose the initiative, grow limited working capacity, and develop a lack of stamina, fortitude, and toughness (25). It has been reported that interns with higher burnout levels experienced chronic sleep deprivation (40).
Morning-evening type and shift pattern of nurses did not affect sleep quality in our study. Chung et al, on the other hand, reported that chronotype affected sleep quality (11). These contradictory results require further studies on the topic. Admi et al found that shift pattern did not affect sleep quality (33). However, some studies showed that rotating shifts worsened circadian rhythm and risked the safety of both health care workers and patients (32,34,35,41,42).
In this study, evening type nurses had poorer subjective sleep quality and sleep latency, which were subscales of PSQI, than nurses of other two types. The administration in the hospital under study did not, however, arrange shifts according to chronotype. Evening type participants were shown to have more negative habits than morning types (14). Morning-types had early sleep schedules and circadian rhythms, and regular waking-, bed-, and sleep-time.
Evening types, on the other hand, had late sleep schedules and circadian rhythms, and irregular waking-, bed-and sleep-time. Also, evening-types experienced more common irregular sleep and lifestyle habits, and dissatisfaction with the sleep (14).
Poor sleep quality due to increased burnout level in nurses working rotating shift was an interesting finding, since it might have been expected that nurses working night shifts have lower job stress level than others. Similarly, nurses working fixed day shifts had increased EE and DP scores. Adaptation of biological rhythm would be a lot easier for nurses working fixed shifts than rotating shifts. Although separate studies were available, we did not find any study that investigated MBI, PSQI, and work patterns together. Jamal and Baba (42), for example, found no correlation between work pattern and burnout levels but found higher health problems for nurses working rotating shifts. Newey and Hood studied the relationships of work patterns and sleep/fatigue, and found that night shift was the worst shift followed by day shift and evening shift (41). Gold et al indicated that error rate related to sleepiness of nurses working rotating shift was twice as high as that in other shifts (32).
A limitation of this study is that it was carried out on a given nurse population in Ankara, with a self-reported ques-tionnaire, which might result in a biased reply from each respondent. Therefore, generalizations should be made carefully and be limited only to the investigated population. In conclusion, our findings showed that majority of nurses experienced poor sleep quality and had increased levels of burnout, especially rotating shift nurses. Most of the nurses who did not belong to either morning or evening type had poor sleep quality. Further studies are needed to plan interventions that decrease burnout and improve sleep quality for shift-work nurses. These interventions would likely improve nurses' overall well-being and working conditions and patients' safety. The hospital administrations should take workers' chronotypes into account when forming the shift lists.
We thank Assistant Professor Dr Pınar Özdemir for his guidance during statistical analysis of the study. This study was submited for oral presentation in the 14th International Nursing Research Conference on 9-12 November 2010 Burgos, Spain.
Funding None.
ethical approval Hospital ethics committee did not require an ethical approval for this study but hospital management provided written approvals. The participants signed the informed consent.
Declaration of authorship ADZ designed the study and performed data analysis and manuscript preparation. SA designed the study and performed data collection.