Providing optimal care regardless of the time or day of birth is an essential goal of perinatal medicine. However, providing optimal care at night and on weekends requires adequate and effective staffing, the availability of diagnostic and therapeutic services, and a team whose performance has not been diminished by fatigue or circadian challenges. Unfortunately, there are many examples of disparities in both maternal and neonatal outcomes by time and day of birth.1–11 For example, in a Swedish study, 12% of early neonatal deaths and in a California study, 9.6% of all neonatal deaths were associated with what was deemed to be decreased quality of care among nighttime births.1,2 In addition to neonatal compromise, clinically significant increases in severe maternal morbidity have also been reported in off-hour births.3 As such, the existence of outcome differentials based on time or day of birth at the institutional, network, or regional level may signal the need for quality improvement interventions. It is important to note that these differentials are not universal. There are several examples of well-staffed, “quaternary,” networks and institutions where there are no time of birth differentials.12–16 These findings support the notion that increased morbidity and mortality during off-hour births may signal clinically significant opportunities to improve perinatal outcomes.
In a recent population-based study that included both vaginal and cesarean births in California, Lyndon and associates found that after careful adjustment for potential differences in case mix, there was a 30% increase in severe maternal morbidity when birth occurred between 11 p.m. and 7 am.3 These findings raised concern that neonatal morbidity may also be increased in off-hour births. To investigate this, we assessed the relationship between severe unexpected neonatal morbidity and time of birth in a cohort of singleton term neonates without evidence of congenital anomalies or preadmission fetal compromise born in California from 2011–2013. We asked the specific questions of whether there was an increase in severe unexpected neonatal morbidity in these uncomplicated births on the afternoon and night shifts and also on weekends, and whether our findings would be seen across all nine perinatal regions.
We conducted a population-based, cross-sectional analysis using data derived from a statewide California linked birth certificate-neonatal discharge and maternal discharge data set17 (with a successful linkage rate of 98.2%) produced by the California Maternal Quality Care Collaborative Data Center for the birth years 2011–2013. Births that occurred in military hospitals, birth centers, or at home were excluded because they do not report hospital discharge data. The analysis addressed potential differences in the rate of severe unexpected neonatal complications by nursing shift and day of birth in a cohort of term singleton neonates without preexisting conditions, growth restriction, birth defects, or other fetal conditions and not exposed to maternal drug use using the National Quality Forum 0716 specifications for severe unexpected neonatal morbidity.18 The “Unexpected Complications in Term Newborns” measure was developed by the California Maternal Quality Care Collaborative,19 and consisted of one or more of the following conditions: neonatal death; transfer to higher level of care; severe birth injuries; and severe neurologic, respiratory, or infectious complications. The International Classification of Diseases, Tenth Revision codes for this measure are presented at https://manual.jointcommission.org/releases/TJC2018B1/MIF0393.html. Conditions were identified using linked data from birth certificates and neonatal hospital discharge diagnosis records (disposition and diagnosis codes).
We conducted unadjusted and risk adjusted analyses to assess the relationship between the risk of severe unexpected neonatal morbidity and the time and day of birth. Unadjusted analyses were performed using a univariate logistic regression model with the binary outcome severe unexpected neonatal morbidity, yes or no. Adjusted analyses were modeled using a hierarchical random effects model where individual hospitals were considered random effects and potential confounders as fixed effects. Potential confounders included maternal age, race and ethnicity, body mass index, parity, education, insurance status, prenatal care, maternal eclampsia and preeclampsia, chronic hypertension, diabetes type 1 or 2, gestational diabetes, neonatal weight greater than 4 kg, and hospital volume. Method of delivery was not included in the risk-adjustment model, because it occurred after admission and was at the discretion of the obstetric health care professional. Although there is no standard definition for the three nursing shifts in California, the most common time frames were: daytime (7 am until 3 pm), evening (3 pm until 11 pm) and nighttime (11 pm until 7 am). For ease of presentation these time periods are noted as (7 am–3 pm), (3–11 pm) and (11 pm–7 am). To obtain a more granular view, time of birth was also analyzed and figuratively presented in 1-hour blocks.
We repeated the assessment by nursing shifts for each of Californians nine major perinatal regions to assess the variability in the risk of severe unexpected neonatal morbidity during off hour's births across California.
Adjusted odds ratios (ORs) with 95% CIs were estimated using the GLIMMIX Procedure (SAS 9.4). To address the potential of multiple comparisons, we used the Benjamini-Hochberg false discovery rate procedure to adjust for P-values. We used an adjusted P<.05 as the threshold to select significant results.
The Stanford University institutional review board approved the study.
After exclusions (Fig. 1) the final study cohort included 1,048,957 neonates, of whom 2.0% (20,618) experienced severe unexpected neonatal morbidity. Although there were statistically significant differences in demographic and clinical factors across patients in the three shifts there was no consistent pattern to the differences (Table 1). The rate of severe unexpected neonatal morbidity was lower among births delivered during the day shift (1.8%) and higher on the evening and night shift (2.1% and 2.1%). Births were most frequent during the day shift and decreased on the evening and night shifts. Fewer births occurred on the weekend (Table 2).
Compared with births during the day shift, both the unadjusted and adjusted odds of severe unexpected neonatal morbidity were increased in the neonates of women with uncomplicated births who delivered during the evening and night shifts (Table 2). Compared with births during the day shift, the odds of adjusted severe unexpected neonatal morbidity were increased for births during the evening (adjusted OR 1.10, 95% CI 1.06–1.13) and night shifts (adjusted OR 1.15, 95% CI 1.11–1.19). In our day-of-the-week unadjusted analysis, the odds of severe unexpected neonatal morbidity were significantly increased on both Saturday and Sunday. However, after risk adjustment a significant increased risk of severe unexpected neonatal morbidity was seen only on Sunday (adjusted OR 1.08, 95% CI 1.02–1.14).
Figure 2 shows the hourly course of the adjusted and unadjusted ORs for severe unexpected neonatal morbidity across the three shifts. The above analyses demonstrated that across California as a whole the risk of severe unexpected neonatal morbidity was increased in the evening and night shifts and on Sundays. To examine whether this held true for all of California, we repeated the analysis for each of California's nine major perinatal regions. Only four of the regions had an increase in the odds of severe unexpected neonatal morbidity, two during both the evening and night shifts and two only during the night shift (Table 3). Our analysis also revealed that the increased severe unexpected neonatal morbidity on Sunday was seen only in region E (adjusted OR 1.17, 95% CI 1.06–1.29, Table 3).
In our analysis of severe unexpected morbidity in uncomplicated term singleton neonates born in Californian between 2011–2013, we found that after adjusting for temporal differences in case mix, the odds of severe unexpected neonatal morbidity was increased during the evening and night shifts. There was also an 8% increase in the odds of severe unexpected neonatal morbidity for births on Sunday. Note that the results of the unadjusted and adjusted analyses have important but different utility. The unadjusted ORs are of practical significance because they reflect the observed increase, decrease, or no change in the number of women within a specific risk category who will require delivery care. The adjusted ORs are potentially of etiologic significance because they reflect an estimate of the independent association of a risk factor with increasing, decreasing, or having no change in the odds of the outcome under consideration. Although association does not assure causality, adjusted ORs identify areas of risk that are worthy of further consideration for intervention.
Evaluators of temporal changes in the risk of mortality or morbidity have often approached their task as a search for the validity of this phenomenon.12,14,20,21 However, we believe that the evaluation of temporal differences in risk-adjusted levels of adverse outcomes represents an important quality improvement strategy that asks the question, “Will our patients receive the same quality of care regardless of day of week or time of day.” Although there are many examples in which women who deliver in the evening, at night, and on weekends are at increased risk of complications both in themselves and in their offspring, there are also studies where the risk of perinatal complications did not vary by time of day or day of week. We feel that, rather than invalidating the concept of temporal differentials, these examples, often performed in large quaternary centers such as the members of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units Network,12 members of the NICHD Neonatal Research Network,13 as well as individual hospitals,14,15 demonstrate that it is possible to achieve a continuous level of care, not only during off hours but also during the months when care is being delivered by newly arrived and less-experienced house staff.13 Although temporal metrics may unmask an important opportunity to improve care, the quality-improvement literature suggests that the root causes and their specific interventions will vary across institutions. Each institution must consider the adequacy of their evening, night, and weekend staffing with respect to adequacy of number, experience, and their ability to perform well when capacity is strained by high volume or decreased staff.22–27 Fatigue has also been shown to spawn errors in diagnostic and therapeutic judgment and action.28–32 Therefore, one must consider the possibility of fatigue owing to too-long work shifts or too few staff, especially during periods of high volume. Finally, it is important to consider the availability and quality of key diagnostic, therapeutic and support services during the evening, night, and weekend. Although we believe that consideration of the time of birth should be included in every perinatal morbidity and mortality case review, tools designed to facilitate an institution's systematic assessment and a database recording of potential opportunities for quality improvement are needed.
An important limitation of our study is that temporal evaluations at the regional or very large network level have to be interpreted cautiously. Although risk-adjusted ORs derived from a region or large network may reveal a statistically worrisome signal, the signal may have been generated by only certain hospitals within the region. Indeed, within a region or network most quality metrics, severe unexpected neonatal morbidity included33 vary greatly, there being superior, average, and challenged performers at all levels. Although our all-California analysis found a significant increase in risk-adjusted severe unexpected neonatal morbidity during off hours, when we stratified by perinatal region, the risk of increased severe unexpected neonatal morbidity during evening and night nursing shifts was seen in only four of the nine regions. However, further analysis is also required to explore the likely possibility that even in the regions without night and weekend differences there may also be some challenged hospitals. We are also in the process of assessing the possible differences between the four regions that have the increase in severe unexpected neonatal morbidity and the five that do not. These could include differences in maternal sociodemographic status, patterns of staffing, the use of 24 hour in house hospitalists, as well as the availability of support services such as anesthesia during evenings, nights, and weekends.
Our analysis was based on a 2011–2013 birth cohort. Given California's high level of attention to and rapid changes in perinatal quality,34 the need to use more contemporary data represents a second potential limitation of our study. The analysis of temporal differences at the individual hospital level based on more contemporary data is now underway.
A third limitation is our inability to completely rule out that the observed temporal disadvantages were the result of as-yet unidentified patient risk factors, or that our risk adjustment has been incomplete. For example, we did not include labor method (spontaneous, induced, or no labor) because of concerns of the accuracy of these definitions in the administrative data sets. We also felt that other included risk factors such as parity, hypertension and diabetes would comprise most of the potential confounding. Of note, a sensitivity analysis found that a significant elevated risk-adjusted rate of severe unexpected neonatal morbidity was seen only in the combined analysis of the 629,516 neonates who were delivered in the four regions that had an increased rate of severe unexpected neonatal morbidity during the night or evening shifts and not in the combined analysis of the 419,441 neonates who were delivered in the five regions where the rate of severe unexpected neonatal morbidity was not elevated (data available on request); this suggests that our risk selection and adjustment was at least reasonable.
A fourth limitation of this study is that the temporal differences (or lack of differences) for any outcome may reflect only those decisions and operations that are essential to that outcome.2–4 Our choice of severe unexpected neonatal morbidity in a term singleton vertex birth without detected fetal compromise was chosen, in part, to assess the ability of a perinatal team to recognize and act on the unexpected. The extent to which this reflects a team's ability to effectively care for a known high-risk delivery would require an additional temporal analysis of outcomes in high-risk deliveries. Indeed, the finding that there were time-of-day effects on the lowest risk patients raises concern that there might be similar or even greater effects on higher risk women and neonates when delivery occurs at night or on weekends.
In summary, we propose that assessing temporal differences in adverse outcomes represents an important quality improvement strategy. Reports that fail to identify temporal differences provide an important demonstration that it is possible to maintain one's quality of care regardless of time of day or day of week. This is not the case everywhere in California. Our study demonstrates that in California being born in the evening or at night is associated with an increased risk of severe unexpected neonatal morbidity in term singleton neonates without pre-existing fetal conditions.
1. Stephansson O, Dickman PW, Johansson AL, Kieler H, Cnattingius S. Time of birth and risk of intrapartum and early neonatal death. Epidemiology 2003;14:218–22.
2. Gould JB, Qin C, Chavez G. Time of birth and the risk of neonatal death. Obstet Gynecol 2005;106:352–8.
3. Lyndon A, Lee HC, Gay C, Gilbert WM, Gould JB, Lee KA. Effect of time of birth on maternal morbidity during childbirth hospitalization in California. Am J Obstet Gynecol 2015;213:705.e1–11.
4. Jensen EA, Lorch SA. Association between off-peak hour birth and neonatal morbidity and mortality among very low birth weight infants. J Pediatr 2017;186:41–8.e4.
5. Palmer WL, Bottle A, Aylin P. Association between day of delivery and obstetric outcomes: observational study. BMJ 2015;351:h5774.
6. Peled Y, Melamed N, Chen R, Pardo J, Ben-Shitrit G, Yogev Y. The effect of time of day on outcome of unscheduled cesarean deliveries. J Matern Fetal Neonatal Med 2011;24:1051–4.
7. Gijsen R, Hukkelhoven CW, Schipper CM, Ogbu UC, de Bruin-Kooistra M, Westert GP. Effects of hospital delivery during off-hours on perinatal outcome in several subgroups: a retrospective cohort study. BMC Pregnancy Childbirth 2012;12:92.
8. Kalogiannidis I, Margioula-Siarkou C, Petousis S, Goutzioulis M, Prapas N, Agorastos T. Infant births during the internal night are at increased risk for operative delivery and NICU admission. Arch Gynecol Obstet 2011;284:65–71.
9. Urato AC, Craigo SD, Chelmow D, O'Brien WF. The association between time of birth and fetal injury resulting in death. Am J Obstet Gynecol 2006;195:1521–6.
10. de Graaf JP, Ravelli AC, Visser GH, Hukkelhoven C, Tong WH, Bonsel GJ, et al. Increased adverse perinatal outcome of hospital delivery at night. BJOG 2010;117:1098–107.
11. Wu YW, Pham TN, Danielsen B, Towner D, Smith L, Johnston SC. Nighttime delivery and risk of neonatal encephalopathy. Am J Obstet Gynecol 2011;204:37.e1–6.
12. Brookfield KF, O'Malley K, El-Sayed YY, Blumenfeld YJ, Butwick AJ. Does time of delivery influence the risk of neonatal morbidity? Am J Perinatol 2016;33:502–9.
13. Bell EF, Hansen NI, Morriss FH Jr, Stoll BJ, Ambalavanan N, Gould JB, et al. Impact of timing of birth and resident duty-hour restrictions on outcomes for small preterm infants. Pediatrics 2010;126:222–31.
14. Butler K, Ramphul M, Dunney C, Farren M, McSweeney A, McNamara K, et al. A prospective cohort study of the morbidity associated with operative vaginal deliveries performed by day and at night. BMJ Open 2014;4:e006291.
15. Caughey AB, Urato AC, Lee KA, Thiet MP, Washington AE, Laros RK Jr. Time of delivery and neonatal morbidity and mortality. Am J Obstet Gynecol 2008;199:496.e1–5.
16. Gould JB, Qin C, Marks AR, Chavez G. Neonatal mortality in weekend vs weekday births. JAMA 2003;289:2958–62.
17. Herrchen B, Gould JB, Nesbitt TS. Vital statistics linked birth/infant death and hospital discharge record linkage for epidemiological studies. Comput Biomed Res 1997;30:290–305.
18. National Quality Forum. NQF # 0716, unexpected complications in term newborns. 2016. Available at: www.qualityforum.org/qps
. Retrieved June 19, 2020.
19. California Maternity Quality Care Collaborative. Unexpected complications in term newborns. Available at: https://www.cmqcc.org/focus-areas/quality-metrics/unexpected-complications-term-newborns
. Retrieved in 2019.
20. Pauls LA, Johnson-Paben R, McGready J, Murphy JD, Pronovost PJ, Wu CL. The weekend effect in hospitalized patients: a meta-analysis. J Hosp Med 2017;12:760–6.
21. Zhou Y, Li W, Herath C, Xia J, Hu B, Song F, et al. Off-hour admission and mortality risk for 28 specific diseases: a systematic review and meta-analysis of 251 cohorts. J Am Heart Assoc 2016;5:e003102.
22. Snowden JM, Kozhimannil KB, Muoto I, Caughey AB, McConnell KJ. A “busy day” effect on perinatal complications of delivery on weekends: a retrospective cohort study. BMJ Qual Saf 2017;26:e1.
23. Aiken CE, Aiken AR, Scott JG, Brockelsby JC. The influence of hours worked prior to delivery on maternal and neonatal outcomes: a retrospective cohort study. Am J Obstet Gynecol 2016;215:634.e1–7.
24. Snowden JM, Darney BG, Cheng YW, McConnell KJ, Caughey AB. Systems factors in obstetric care: the role of daily obstetric volume. Obstet Gynecol 2013;122:851–7.
25. Snowden JM, Caughey AB. Is there a weekend effect in obstetrics? BMJ 2015;351:h6192.
26. Hamilton KE, Redshaw ME, Tarnow-Mordi W. Nurse staffing in relation to risk-adjusted mortality in neonatal care. Arch Dis Child Fetal Neonatal Ed 2007;92:F99–F103.
27. Needleman J, Buerhaus P, Mattke S, Stewart M, Zelevinsky K. Nurse-staffing levels and the quality of care in hospitals. N Engl J Med 2002;346:1715–22.
28. Johnson AL, Jung L, Song Y, Brown KC, Weaver MT, Richards KC. Sleep deprivation and error in nurses who work the night shift. J Nurs Adm 2014;44:17–22.
29. Smith-Miller CA, Shaw-Kokot J, Curro B, Jones CB. An integrative review: fatigue among nurses in acute care settings. J Nurs Adm 2014;44:487–94.
30. Lockley SW, Barger LK, Ayas NT, Rothschild JM, Czeisler CA, Landrigan CP. Effects of health care provider work hours and sleep deprivation on safety and performance. Joint Comm J Qual Patient Saf 2007;33(11 suppl):7–18.
31. Chua KP, Gordon MB, Sectish T, Landrigan CP. Effects of a night-team system on resident sleep and work hours. Pediatrics 2011;128:1142–7.
32. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and the safety of patients. N Engl J Med 2002;347:1249–55.
33. Sebastiao YV, Womack LS, Lopez Castillo H, Balakrishnan M, Bruder K, Alitz P, et al. Hospital variations in unexpected complications among term newborns. Pediatrics 2017;139:pii: e20162364.
34. Main EK. Reducing maternal mortality and severe maternal morbidity through state-based quality improvement initiatives. Clin Obstet Gynecol 2018:61:319–31.