Introduction
Over the past two decades, a growing body of evidence has demonstrated an association between work outside of the Monday through Friday schedule and increased morbidity and mortality in hospital settings [
1]. Most studies that have found worsened outcomes in patients admitted during the weekend and during off-hour shifts have evaluated patients requiring urgent or emergency intervention, with conditions such as stroke [
2], myocardial infarction [
3], traumatic brain injury [
4], pulmonary embolism [
5], gastrointestinal hemorrhage [
6], pediatric surgery [
7], and mortality in critical care admission [
8]. In obstetrical and perinatal outcomes, several studies have indicated the unfavorable influence of weekend and night delivery on neonatal death [
9‐
14], birth-related injury,[
14,
15] asphyxia, [
16] intraventricular hemorrhage, [
11] neonatal encephalopathy [
17], maternal complication [
18], and perinatal complication [
12,
14,
18,
19]. Recent studies have analyzed the impact of staff fatigue [
20], lack of medical personnel [
17,
18], high volume [
21], and institutional characteristics (e.g., academic teaching hospital [
22] or secondary/tertiary hospital [
23,
24]) on weekend and off-hour outcomes. A report from Sweden found an increased risk of asphyxia in infants born at night [
16]. This indicates a possible negative weekend and off-hour effect on obstetrical care. Definitive conclusions have not been reached in this area due to narrowly focused research settings and the range of disease and symptom outcomes assessed [
11,
18,
21,
25‐
28].
This study focuses on cerebral palsy (CP) as an outcome to evaluate weekend and off-hour effects in Japanese settings. In Japanese obstetrical care, stillbirth outcomes have improved, but risk for CP remains among surviving infants [
29]. One study that used data from one hospital in Japan reported that adverse neonatal outcomes among cesarean births increased during the night [
30]. The number of obstetricians is decreasing in Japan [
31], and access to obstetric care is being reduced, especially on weekends and during off-hours. To tackle this urgent and problematic shortage, perinatal medical care has been consolidated, and tertiary (comprehensive and regional) perinatal care centers have been established. This consolidation may modify weekend and off-hour effects. This study evaluated whether there are weekend and off-hour effects and examined the potential effects of the type of perinatal medical care center on the incidence of CP, using nationwide Japanese databases.
Results
A total of 182 CP cases with normal spontaneous delivery without cesarean section and with emergent cesarean section (Table
1) were selected by year (2009, 64 [35.2%]; 2010, 49 [26.9%]; 2011–2012, 69 [37.9%]) from the JOCSC database (Table
1). These were matched to 1820 controls, also sampled by year (2009, 640 [35.2%]; 2010, 490 [26.9%]; 2011–2012, 690 [37.9%]). In all cases of either normal spontaneous delivery or emergent cesarean section, significantly more preterm labor (
P < 0.001), fewer gestational weeks (
P < 0.001), more fetus growth restriction (SD for birth weight;
P < 0.001), more weekend delivery (
P = 0.015), and more use of regional perinatal care centers (
P < 0.001) were seen in the CP cases.
Table 1
Demographics and prenatal participant characteristics
Birth year |
2009 | 64 (35.2%) | 640 (35.2%) | NA |
2010 | 49 (26.9%) | 490 (26.9%) | |
2011–2012 | 69 (37.9%) | 690 (37.9%) | |
Delivery |
Normal | 90 (49.5%) | 900 (49.5%) | NA |
Emergency CS | 92 (50.5%) | 920 (50.5%) | |
Age, years | 32.0 ± 5.1 | 31.9 ± 5.4 | 0.865 |
Parity |
One or more | 77 (42.3%) | 688 (37.8%) | 0.233a |
Prescription |
Prostaglandin | 22 (12.1%) | 156 (8.6%) | 0.112a |
Oxytocin | 41 (22.5%) | 479 (26.3%) | 0.266a |
PIH | 13 (7.1%) | 116 (6.4%) | 0.687a |
Preterm labor | 68 (37.4%) | 230 (12.6%) | < 0.001a |
Gestational week | 37.8 ± 2.3 | 38.8 ± 1.7 | < 0.001 |
SD for birth weight | -0.304 ± 1.032 | 0.003 ± 1.039 | < 0.001 |
Perinatal care centers |
Comprehensive | 60 (33.0%) | 992 (54.5%) | < 0.001a |
Regional | 122 (67.0%) | 828 (45.5%) | |
Day |
Weekday | 122 (67.0%) | 1344 (73.9%) | 0.015a |
Weekend | 60 (33.0%) | 476 (26.2%) | |
Time shift |
Day | 89 (48.9%) | 1015 (55.4%) | 0.088a |
Quasi-night | 44 (24.2%) | 435 (23.9%) | |
Night | 49 (26.9%) | 370 (20.3%) | |
Table
2 shows that significantly more cases than controls had night deliveries, adjusted for year and obstetric condition (Model 2: OR 1.56, CI 1.05–2.32). Significantly more cases than controls also had weekend deliveries, adjusted for year and obstetric condition (Model 2: OR 1.59, CI 1.12–2.25). In addition, significantly more cases than controls had delivery in regional perinatal care centers, adjusted for year (Model 1: OR 2.41, CI 1.74–3.33) and obstetric condition (Model 2: OR 2.50, CI 1.78–3.50).
Table 2
Odds ratios for weekend and off-hour effects on cerebral palsy
Perinatal care centers |
Comprehensive | 1.00 | 1.00 | 1.00 |
Regional | 2.40, 1.74–3.32 | 2.41, 1.74–3.33 | 2.50, 1.78–3.50 |
Day |
Weekday | 1.00 | 1.00 | 1.00 |
Weekend | 1.50, 1.08–2.08 | 1.41, 1.01–1.97 | 1.59, 1.12–2.25 |
Shift |
Day | 1.00 | 1.00 | 1.00 |
Quasi-night | 1.25, 0.86–1.83 | 1.28, 0.87–1.87 | 1.29, 0.87–1.92 |
Night | 1.48, 1.02–2.14 | 1.43, 0.98–2.08 | 1.56, 1.05–2.32 |
Hosmer–Lemeshow test | | P = 0.636 | P = 0.415 |
For those with normal spontaneous delivery (Table
3), significantly more cases than controls were delivered at regional perinatal care centers, adjusted for year (Model 3: OR 2.69, CI 1.69–4.30) and obstetric condition (Model 4: OR 2.92, CI 1.76–4.84).
Table 3
Odds ratios for weekend and off-hour effects on cerebral palsy for normal spontaneous delivery
Perinatal care centers |
Comprehensive | 1.00 | 1.00 | 1.00 |
Regional | 2.65, 1.66–4.21 | 2.69, 1.69–4.30 | 2.92, 1.76–4.84 |
Day |
Weekday | 1.00 | 1.00 | 1.00 |
Weekend | 1.20, 0.76–1.91 | 1.17, 0.73–1.88 | 1.63, 0.97–2.73 |
Shift |
Day | 1.00 | 1.00 | 1.00 |
Quasi-night | 1.44, 0.85–2.44 | 1.51, 0.89–2.58 | 1.65, 0.93–2.95 |
Night | 1.04, 0.62–1.75 | 1.00, 0.59–1.71 | 1.26, 0.70–2.27 |
Hosmer–Lemeshow test | | P = 0.075 | P = 0.897 |
For those with cesarean section (Table
4), significantly more cases than controls had weekend (OR 1.75, CI 1.09–2.80) or night (OR 2.40, CI 1.40–4.11) deliveries (Model 5). In the multivariable model adding obstetric conditions (Model 6), cases again had significantly more weekend (OR 1.72, CI 1.06–2.81) and night (OR 2.29, CI 1.30–4.02) deliveries. Significantly more cases than controls had a cesarean section in a regional perinatal care center, adjusted for year (Model 5: OR 2.24, CI 1.42–3.53) and obstetric condition (Model 6: OR 2.35, CI 1.47–3.77).
Table 4
Odds ratios for weekend and off-hour effects on cerebral palsy for delivery with emergent caesarian section
Perinatal care centers |
Comprehensive | 1.00 | 1.00 | 1.00 |
Regional | 2.19, 1.40–3.43 | 2.24, 1.42–3.53 | 2.35, 1.47–3.77 |
Day |
Weekday | 1.00 | 1.00 | 1.00 |
Weekend | 1.89, 1.19–2.99 | 1.75, 1.09–2.80 | 1.72, 1.06–2.81 |
Shift |
Day | 1.00 | 1.00 | 1.00 |
Quasi-night | 1.06, 0.61–1.84 | 1.04, 0.59–1.81 | 1.15, 0.64–2.07 |
Night | 2.46, 1.45–4.17 | 2.40, 1.40–4.11 | 2.29, 1.30–4.02 |
Hosmer–Lemeshow test | | P = 0.700 | P = 0.561 |
The Hosmer–Lemeshow test showed a fairly good fit for all models (no significant P values were observed in any models).
Discussion
Using two nationwide perinatal databases, our case-control study in comprehensive and regional perinatal care centers found no weekend or off-hours effects on CP evident in normal spontaneous deliveries. However, we observed a 1.7-fold weekend and 2.3-fold night on CP for delivery with emergency cesarean section without emergency transportation. To the best of our knowledge, this is the first study to investigate the risk for CP associated with delivery during the weekend and off-hour periods.
In Japan, approximately 2500 facilities provide delivery services for approximately one million deliveries per year [
31]. Because pregnant women often prioritize the accessibility and comfort of delivery facilities, deliveries are not intensively performed, and more than half of all deliveries occur in private clinics operated by one or two obstetricians [
31]. Nevertheless, the perinatal mortality rate and maternal mortality rate in Japan in 2012 were 4 per 1000 and 4 per 100,000 births, respectively [
38], the lowest figures worldwide. Thus, the safety standards of prenatal practice in Japan are at least equivalent to those in other developed countries.
The incidence of CP in Japan is estimated to be 2.0 per 1000 births [
39], which is about the same as that in other developed countries [
40]. CP occurrence related to labor management should be limited, but the current study indicates that the risk might be elevated on the weekend and at night even in comprehensive and regional perinatal care centers that offer advanced perinatal care.
Previous studies on weekend and off-hour effects on obstetrical and perinatal outcomes have proposed several explanations: staff fatigue [
20], a lack of medical personnel [
17,
18], high volumes [
21], and institutional characteristics [
22‐
24]. Other previous studies have reported no difference in patient management on the weekend versus on weekdays for spontaneous labor [
41] nor any increase in adverse maternal and neonatal outcomes from cesarean births at night [
25]. Another study of weekend and off-hour effects on other health outcomes have offered explanations, such as fewer personnel, lower staffing levels, reduced numbers of senior staff, cross-cover of clinical specialties, proportionally more urgent procedures, deficits in continuity of care, and presenting patients often being in a more complex or more critical condition than those who typically present on weekdays [
42]. Our target hospitals, comprehensive and regional perinatal care centers, must meet staff and facility requirements for 24 h operation; however, medical staff is reduced during the night and on weekends, meaning that fewer obstetricians are available to handle serious cases. Weekend and night CP effects during emergency cesarean sections might be the result of multiple simultaneous factors: lower staffing levels, fewer senior staff, cross-cover of clinical specialties, and proportionally more urgent procedures on weekends and at night.
In all models, regional perinatal care centers had a 2.2-fold or higher risk of CP compared to comprehensive perinatal care centers. Generally, the minimum staffing and equipment requirements at regional perinatal care centers are lower than for comprehensive perinatal care centers [
43]. Severe institutional risk conditions induced by deliveries may thus occur more frequently in regional perinatal care centers.
In a previous study, delays in diagnosis and treatment, including maternal transport, intervention, and blood transfusion, were associated with maternal death in Japan [
26]. In fact, such cases encounter several problems in diagnostic procedures, treatment strategies, and inter- and intra-hospital systems [
31]. In 2000, inadequate obstetric services were associated with maternal mortality in Japan. Reducing the number of deliveries in single-obstetrician facilities and establishing regional 24 h inpatient obstetric facilities for high-risk cases might reduce maternal mortality in Japan [
44].
This weekend and holiday problem in regional perinatal care centers may reflect the current condition of obstetric medical care, in particular, the shortage of obstetricians. To tackle this shortage, the Medical Reform Committee of JSOG published its Grand Design 2015 [
45], which recommended that obstetric services be consolidated in secondary and tertiary care hospitals, and regional and comprehensive perinatal care centers, to reach a sufficient number of full-time obstetricians. Consolidation of obstetric facilities impairs access, but a higher caseload and better staffing have the potential advantages of better clinical outcomes and reduced costs [
46]. Metropolitan areas could develop plans to consolidate perinatal medical care; however, rural regions would not be able to achieve similar consolidation. The redistribution of care centers and obstetricians should also be accelerated. Comprehensive perinatal care centers are not distributed throughout the country to provide efficient access to perinatal care. According to the current results, appropriate management and intensive care is needed in regional perinatal care centers during the night and on weekends and holidays. Consolidation progressed during the study period 2009–2011. Further study is needed to fully capture the current situation.
This study has several strengths. First, we included a large number of CP patients from a nationwide database. Statistical power was estimated at 86.0, 62.6, and 97.8%, respectively, for night, weekend, and regional perinatal care center effects for the bivariate analysis of delivery with emergent cesarean section. Second, data were drawn from available medical information and did not include self-administrated questionnaires, attenuating potential recall bias. However, this study also had several limitations. First, to provide grounds for comparison, it focused on cases and controls in comprehensive and regional perinatal care centers without ambulance transportation. Labor in general obstetrics clinics and hospitals with a limited number of obstetricians and midwives might be associated with exaggerated risk during weekend and off-hour delivery or in emergent cases with ambulance transportation. Second, only participants with births at 33 or more weeks were eligible to participate, so night and weekend effects were not studied for births outside of these restrictions. Third, the quality of the dataset was not balanced because the JOCSC data were manually confirmed by JOCSC investigators but the JSOG data were automatically transcribed from electronic patient records. Individual risk behaviors, i.e., alcohol drinking and drug use, were assessed with different questions between databases. No pair of comparable databases exists. Individual risk behaviors occur more often during weekends and at night and may therefore be a cause of weekend and night effects produced in this study [
2]. Pregnant women usually avoid risk behaviors; however, the potentially confounding effects of individual behaviors were not analyzed in the current study. Fourth, the time between amniotic rupture and total delivery was not included in the analysis due to uncertain data recording.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.