Background
The field of obstetrics and gynecology requires complex decision-making and skills because of unexpected high-risk situations. These skills are influenced by alertness, reaction time, and concentration. Night shifts result in sleep deprivation, which might impair these functions, although it is still unclear to what extent.
Objective
This study aimed to investigate whether a night shift routinely impairs the obstetrics and gynecology consultants’ and residents’ fitness to perform and whether this reaches a critical limit compared with relevant frames of reference.
Study Design
Residents (n=33) and consultants (n=46) in obstetrics and gynecology conducted multiple measurements (n=415) at precall, postcall, and noncall moments with the fitness to perform self-test. The self-test consists of an adaptive pursuit tracking task that is able to objectively measure alertness, reaction time, concentration, and hand-eye coordination and Visual Analog Scale tests to subjectively score alertness. The test is validated with a sociolegal reference of a 0.06% ethanol blood concentration (the peak level after 2 units of alcohol, the legal driving limit). This equals −1.37% on the objective score and −8.17 points on subjective alertness. Linear mixed models were used to analyze the difference within subjects over a night shift, integrating repeated measures over time.
Results
The overnight objective difference between postcall and precall measurements was −0.62 ( P <.05) for residents and 0.28 ( P =NS) for consultants, both not exceeding the sociolegal reference as a group. Objective impairment exceeded the reference for 31% of the residents and 28% of the consultants. Subjective alertness decreased in residents (−18.26; P <.001) and consultants (−10.85; P <.001), both exceeding the reference. No residents had to continue work postcall versus 7.8% of the consultants. None of the consultants that had to continue work were in an objective critically impaired state.
Conclusion
This study provides insight and awareness of individual performance after night shifts with clear frames of reference. The performance of residents is negatively and significantly affected by night shifts; therefore, a scheduled day off after a night shift is justified. Consultants showed no overall impairment; however, a quarter did exceed the alcohol limit reference after their night shift. If not logistically feasible to schedule a protected day off after a night shift, our group recommends safe shift scheduling, including options to transfer care after a demanding night shift to prevent working in a compromised state.
Introduction
The field of obstetrics and gynecology requires around-the-clock complex decision-making and prompt surgical and obstetrical interventional skills. These skills are influenced by alertness, reaction time, and concentration, which are impaired by sleep deprivation and fatigue. This impairment can exceed the impairment in performance observed after alcohol intoxication above the legal limit for safe driving. Therefore, sleep deprivation could increase the rate of attentional failures and clinical misjudgment. As obstetrical care continues around the clock, impaired performance owing to sleep deprivation is a likely factor affecting patient safety in obstetrics and gynecology.
Why was this study conducted?
This study was conducted to investigate whether a night shift routinely impairs the obstetricians’ or gynecologists’ fitness to perform and whether this reaches a critical limit compared with relevant frames of reference, such as the legal alcohol limit.
Key findings
Resident performance is negatively affected by night shifts, by both objective and subjective measures. Overall, consultants showed no objective impairment, although 28% of consultants did exceed the alcohol limit reference. Subjectively, they showed significant impairment overall.
What does this add to what is known?
Subjects served as their own control and data included objective and subjective measurements of fitness to perform, therefore gaining valuable insight into clinical practice, the objective influence of night shifts, and the self-perception of night shifts. Safe shift scheduling, including options to transfer care after a demanding night shift, is advised to prevent working in a diminished state.
In response to concerns about sleep deprivation in medical care, the Accreditation Council for Graduate Medical Education limited work hours to 80 hours per week and 24 hours per shift and the European Working Time Directive limited work hours to 48 hours a week. Although these initiatives were intended to improve patient safety, the actual effect has been debated. , In addition, published studies on the effect of fatigue on physician performance show varying results, possibly because of the absence of a real-time, objective measurement tool to use in prospective clinical setting. Therefore, the Fit to Perform (FTOP) tool was designed and validated to assess the state of alertness.
This study aimed to investigate the effect of night shifts on objective and subjective performance in a cohort of Dutch residents and consultants in obstetrics and gynecology and to determine whether this reaches critical limits.
Materials and Methods
Study design
For this national prospective multicenter study, residents and consultants were recruited in collaboration with the Association of Gynecologists in the Netherlands. To acquire a representative sample of Dutch gynecologic care, we chose to include doctors from both academic and affiliated teaching hospitals. All residents and consultants working in these hospitals were approached to participate. Participation was voluntary. Informed consent was obtained, and a personal anonymized demographic questionnaire was conducted.
Here, 8-minute self-measurements were performed using the FTOP testing tool on site, which consist of objective and subjective parts. The objective part is an adaptive pursuit tracking task reflecting alertness, concentration, hand-eye coordination, and reaction time. The score reflects a modified percentage of success of the task. Subjective effects were assessed using Visual Analog Scale (VAS) questionnaires, measuring alertness, contentedness, and calmness, with scores ranging from 0 to 100 ( Supplemental Table 1 ). This paper focused on VAS alertness scores only. This tool has been validated in laboratory and clinical settings, using sociolegal (alcohol legal driving limit, 0.06% ethanol blood concentration) and professional (operative skills) frames of reference, and was used before in the FTOP study on surgeons. , The sociolegal reference is used for clinical relevance. This equals −1.37% on the tracker and −8.17 points on subjective alertness. Demographic and shift characteristics were assessed by a questionnaire ( Supplemental Table 2 ).
Obstetrics and gynecologic shifts were organized differently for Dutch consultants and residents. Residents were in-housed for a maximum of 12 hours during the night. They have protected time off after a night shift and usually perform floats of consecutive shifts. Consultants supervise the residents and work only on demand. Consultant shifts may be scheduled differently among hospitals. Most consultants perform on-call night shifts following their regular day shift. Some consultants are scheduled to continue work directly following their night shift, performing another regular day shift, whereas others have a protected day off. By asking shift characteristics via a questionnaire after the night shift, the activities during a shift and necessity to continue work afterward were clear.
A measurement block consisted of 3 measurements: noncall (on a random workday), precall, and postcall. Subjects functioned as their own control, assessing personal impairment or improvement. Subjects received a reminder (telephone call, text message, and/or email) during on-call days to perform the FTOP test.
Precall measurements for by residents were performed before the night shift (usually around 10 pm ), and postcall measurements were performed at the end of their shift (around 8 am) . Consultants performed precall measurements at the change of day shift to being on-call for the night (around 5 pm ). Postcall measurements were performed at the change of night shift to day shift (around 8 am ). Participants were asked to perform the precall and postcall measurements within 1 hour before or after the change of shifts. There were no dietary or behavioral restrictions before the measurement. Noncall measurements were performed during a dayshift at the participants’ convenience, with the advice to perform this on a Monday morning after a free weekend.
The subjects received direct feedback on the tracker score, without interpretation of the results. The participants were allowed to perform the test during several shifts. Test sets were formed, consisting of combined precall and postcall measurements during 1 night shift supplemented with a noncall measurement taken near the shift.
According to the Dutch Medical Research Involving Human Subjects Act, this study did not require submission to an ethics committee.
Statistical analysis
Outcome measures were reported as mean and standard deviation (SD). Differences are given as the average and the 95% confidence interval of the difference. SPSS Statistics (version 23; IBM, Amsterdam, NL) was used for acquiring linear mixed models to analyze the difference within subjects over a night shift, integrating repeated measures over time with subject factor and including the potential learning effect. Factors of influence for an impaired alertness after a night shift were determined using the calculated difference between precall and postcall alertness with linear mixed model regression analysis. The Pearson correlation analysis, without subject factor, was used to describe the relationship between objective and subjective performance on a group level. Chi-square and McNemar tests were conducted on categorical variables. Posthoc sensitivity analyses were performed to determine the impact of outliers if present. Outliers were defined as values greater than the mean plus or minus 3 SD.
Results
Demographics
The study included 79 participants (33 residents and 46 consultants) from 4 centers, completing 415 measurements ( Supplemental Table 3 ). Consultants were 13 years older than residents. Sex, work, and sleep hours per week did not differ between groups ( Table 1 ). All gynecologic specialties involved are shown in Supplemental Table 4 .
| Variable | Consultants | Residents | 95% CI | P value |
|---|---|---|---|---|
| Subjects | n=46 | n=33 | ||
| Sex | ||||
| Male | 12 (26.1) | 8 (24.2) | ||
| Female | 34 (73.9) | 25 (75.8) | ||
| Age | 47.1 (8.2) | 33.8 (2.9) | 10.7–15.9 | <.001 |
| Position | ||||
| Partnership | 12 (26.1) | 0 | ||
| Salaried employment | 34 (73.9) | 33 (100.0) | ||
| Hours of work per week a | 43.9 (7.7) | 46.9 (6.2) | −6.4 to 0.5 | .088 |
| Hours of sleep per night b | 6.8 (0.7) | 6.8 (0.8) | −0.4 to 0.4 | .993 |
| Experience in years c | 11 (9) | 7 (3) | 1.3–7.1 | .009 |
a Consultants (n=39) and residents (n=30)
b Consultants (n=40) and residents (n=39)
c Consultants, experience in years as specialized attending; residents, experience in years as resident.
Shift characteristics
At the beginning of a night shift, consultants had already worked 7 hours more than residents and were awake for 11.40 hours vs 6.25 hours. Shift duration differed by 8 hours, and hours of sleep during a shift differed by 4.5 hours ( Table 2 , Supplemental Figure 2 ). All activities during the night shift showed significant differences between groups, with residents acting more of their shift in the emergency room or delivery room and consultants performing more surgeries ( Figure 1 ). Consultants were able to take more rest and were scheduled to continue work in 7.8% (n=5) after a shift vs none of the residents. All residents in this cohort performed floats of consecutive shifts.
| Characteristic | Consultants | Residents | 95% CI | P value |
|---|---|---|---|---|
| Noncall measurement | ||||
| Hours slept | 6.7 (1.7) | 7.5 (1.3) | −1.9 to 0.3 | .119 |
| Precall measurement | ||||
| Hours worked | 7.6 (4.6) | 0.5 (1.1) | 6.2–8.0 | <.001 |
| Hours awake | 11.4 (3.7) | 6.3 (3.6) | 4.1–6.2 | <.001 |
| Night shift a | 1 (0.0) | 3 (2.0) | −2.3 to −1.3 | <.001 |
| Postcall measurement | ||||
| Hours awake | 3.6 (6.5) | 15.5 (5.4) | −14.0 to −9.8 | <.001 |
| Hours slept | 4.9 (1.9) | 0.4 (1.2) | 4.0 to 5.1 | <.001 |
| Night shift a | 1 (0.0) | 3(2.0) | −2.3 to −1.4 | <.001 |
| Total length of shift | 18 (6.0) | 10 (1.0) | 6.2 to 9.4 | <.001 |
Objective measurements
Comparing precall and postcall measurements, there was a decrease in objective alertness over the course of a resident night shift (−0.62; P =.035) ( Figure 2 , A; Table 3 ), but this did not exceed the alcohol limit reference. Residents showed more impairment than consultants who had no significant decrement in performance (consultants, 0.28 [ P =.293]; difference in the performance of residents vs performance of consultants, −0.90 [ P =.024]) ( Figure 2 , B; Table 3 ) . The difference between noncall and precall measurements was 0.67 ( P =.293) for residents and 0.043 ( P =.928) for consultants. Objective impairment after the night shift exceeded the alcohol limit reference for 31% of all resident shifts and 28% of the consultant shifts, but none of them had to continue working a day shift immediately following their night shift.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
