Original Article
107
Variation in Labor Induction over the Days of the Week William A. Grobman, MD, MBA2
1 Division of Neonatology, Department of Pediatrics, Ann & Robert
H. Lurie Children’s Hospital of Chicago, Northwestern University, Chicago, Illinois 2 Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Prentice Women’s Hospital-Northwestern Memorial Hospital, Northwestern University, Chicago, Illinois
Karna Murthy, MD, MSc1
Address for correspondence Gustave H. Falciglia, MD, Division of Neonatology, Department of Pediatrics, Ann & Robert H. Lurie Children’s Hospital of Chicago, Feinberg School of Medicine, Northwestern University, 225E. Chicago Avenue, Box 45, Chicago, IL, 60611 (e-mail: [email protected]).
Am J Perinatol 2015;32:107–112.
Abstract
Keywords
► ► ► ► ►
induction of labor late preterm early term variation epidemiology
Objective To quantify the variation in induction of labor (IOL) over the days of the week for gravid women in the United States. Study Design Women who delivered singletons between 24 and 42 weeks’ gestation were identified using birth certificate data from 2007 to 2010. Women with pregnancyassociated hypertension, fetal anomalies, previous cesarean delivery, or incomplete records were excluded. The primary outcome was IOL. Women were stratified into four gestational age groups: < 34, 34 to 36, 37 to 38, and 39 weeks. Frequencies of IOL were determined according to day of the week. Multivariable logistic regression estimated the association between weekend delivery and IOL, adjusting for maternal characteristics and year of delivery. Results There were 11.6 million eligible women. For each gestational age stratum, the frequency of IOL was increased on weekdays compared with weekends (8.0 vs. 7.4%, 16.5 vs. 13.2%, 25.0 vs. 14.8%, and 33.2 vs. 19.3% at < 34, 34–36, 37–38, and 39 weeks, respectively; p < 0.01 for all). Multivariable analyses demonstrated that weekend IOL was inversely related to gestational age (odds ratios: 0.93, 0.77, 0.52, and 0.48, respectively; p < 0.001 for all). Conclusion The frequency of IOL varies according to the day of the week, with the odds of weekend IOL lowest at greater gestational ages.
Background The frequency of induction of labor (IOL) among women with late preterm (LP: 34 0/7–36 6/7 weeks) and early term (ET: 37 0/7– 38 6/7 weeks) gestations has risen over the last two decades.1–4 Relative to infants born at or after 39 weeks’ gestation, infants born during LP or ET gestations are more likely to experience infant mortality, respiratory difficulties, and admission to the neonatal intensive care unit.3–6 Despite these documented associations, the factors that have contributed to the rise in the frequency of LP and ET inductions remain uncertain.
received April 10, 2014 accepted April 11, 2014 published online May 28, 2014
Although, multiple unambiguous medical indications exist for IOL before 39 weeks’ gestation, others are less established; correspondingly, nonmedical factors may have contributed to the rise in IOL at LP and ET gestations.1,2,7 One nonmedical factor that has been unexplored is the inclination to provide IOL on certain days of the week. Although, Gould et al demonstrated that cesarean delivery varied over the days of the week, whether IOL exhibits similar variability is uncertain.8 Variation in IOL over the days of the week, and particularly variation that is smaller in magnitude at earlier gestations, provides support that a nonmedical influence may
Copyright © 2015 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0034-1376392. ISSN 0735-1631.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Gustave H. Falciglia, MD1
Variation in Labor Induction over the Days of the Week have existed. Thus, we aimed to quantify the variation in IOL in the United States over the days of the week as a function of gestational age.
Methods Birth certificate data, and detailed descriptions of these data, from the National Vital Statistics System (NVSS) were used to identify gravid women and their infants delivering from 2007 to 2010 in the United States.9 The study was approved by the Ann & Robert H. Lurie Children’s Hospital of Chicago Research Center Institutional Review Board. Women aged 13 to 49 who delivered a nonanomalous singleton infant between 24 and 42 weeks’ gestation were included in the analysis. Women without prenatal care or who delivered outside a hospital were excluded because IOL is a planned event that takes place within a hospital. Women with previous cesarean deliveries or nonvertex fetal presentation were excluded because IOL is less likely in these situations. Women with pregnancy-induced hypertension or premature rupture of membranes were omitted because the timing of onset and severity of these conditions were unknown; thus, it cannot be known whether women with these conditions at any given gestational age received expectant management or underwent IOL at diagnosis. Finally, women whose records had incomplete data that did not allow ascertainment of the primary exposures and outcome (e.g., gestational age at delivery, IOL, and day of the week of delivery) were omitted from the analysis. The primary exposure was the day of the week of delivery, identified using the variable “day of the week” in the NVSS dataset.9 The weekend was defined as “Saturday through Sunday,” while the weekday was defined as “Monday through Friday.” Using the variable “obstetric/clinical gestation estimate” to determine the gestational age at delivery, we stratified the eligible women into four groups: < 34 0/7, 34 0/7 to 36 6/7, 37 0/7 to 38 6/7, and 39 0/7 weeks’ gestations. The primary outcome was IOL and was identified using the variable “induction of labor” from the birth certificates.9 The maternal covariates considered in the analyses were the year of delivery, maternal age, parity, marital status, and maternal diabetes mellitus, and chronic hypertension. Referent to women between 20 and 34 years old, maternal age was categorized as advanced maternal age ( 35 years old) and teen pregnancy ( 19 years old). For each gestational age and weekday strata, the number of deliveries, proportion of deliveries after IOL, and maternal covariates were described. The proportion of deliveries following IOL was calculated by dividing the number of women who underwent labor induction by the number of women who were eligible to receive IOL on a particular day of the week. This proportion was calculated for weekend and weekday, where the denominators were the total number of deliveries on the weekend and weekday, respectively. Significant differences were calculated using student’s t test, or chi-square analyses, as appropriate. Logistic regression was used to estimate the unadjusted and the independent associations between the weekend and delivery after IOL American Journal of Perinatology
Vol. 32
No. 1/2015
Falciglia et al. for women delivering in each gestational age stratum. Equations were adjusted for the listed maternal covariates and year of delivery. Due to the lag between the initiation of IOL and delivery, sensitivity analyses considered different definitions of weekends and weekdays. Thus, “Monday” was considered as a weekend day in one analysis, while omitted in another. IOL is a well-accepted practice postterm. Many hospitals may not schedule IOL on weekends; therefore, we completed an additional sensitivity analysis omitting women delivering after 41 weeks’ gestation. Analyses were performed with STATA v12 (StataCorp, College Station, TX). Given the large sample sizes, a twotailed α ¼ 0.01 was used to define statistical significance.
Results Of the women who delivered in 2007 to 2010 in the United States, 11,639,654 were eligible for analysis (►Table 1). The differences in population characteristics stratified by gestational age and weekend delivery are reported in ►Table 2. Although women’s characteristics differed significantly according to the days of the week on which they delivered, the magnitude of these differences was small. IOL was significantly more frequent at higher gestational ages. Also, although weekday deliveries accounted for a majority of deliveries, the proportion of women delivering on weekdays (78.79%) was greater than that expected by chance (5/7th or 71.42%). This disproportionate number of weekday deliveries is related to greater frequency of delivery after weekday IOL (►Fig. 1). Women delivering 39 weeks’ gestation demonstrated the greatest day-to-day variation, with the magnitude of this variation declining with decreasing gestation. Deliveries due to IOL varied minimally for women delivering < 34 weeks’ gestation. When the days of the week were aggregated into weekends and weekdays, the proportion of deliveries after IOL was increased on weekdays relative to weekends in each gestational age stratum (8.0 vs. 7.4%, 16.5 vs. 13.2%, 25.0 vs. 14.8%, and 33.2 vs. 19.3% at < 34 0/7, 34 0/7 to 36 6/7, 37 0/7 to 38 6/7, and 39 0/7 weeks’ gestational age, respectively; p < 0.01 for all). In multivariable analyses, these differences persisted after adjusting for year of birth, parity, maternal age, the presence of diabetes mellitus, chronic hypertension, and marital status (►Table 3). Sensitivity analyses considering women delivered on Monday as part of the “weekend” group and also omitting women delivering on Mondays from any equation are presented in ►Table 4. The association between weekend delivery and IOL were essentially unchanged. Lastly, excluding women delivering beyond 40 6/7 weeks’ gestation did not change results from the primary analyses (data not shown).
Discussion The frequency of delivery after IOL was greatest on weekdays. The difference in IOL frequency between weekday and weekend was least before 34 weeks’ gestation and greatest at or
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
108
Variation in Labor Induction over the Days of the Week
Falciglia et al.
109
Table 1 Composition of the cohort Frequency
US deliveries 2007–2010
Percentage
16,724,105
100.0
Deliver < 24 or > 42 wks’ gestational age
105,437
0.6
Nonsingleton
562,177
3.4
Maternal age
1,802
0.0
Unknown maternal diabetes or chronic hypertension
75,887
0.5
Fetal or neonatal anomalies
125,817
0.8
Unknown initiation of labor
10,818
0.1
Previous cesarean delivery
2,174,437
13.0
Nonvertex presentation
868,355
5.2
Premature rupture of membranes
268,625
1.6
No prenatal care
174,818
1.1
Unknown Hispanic ethnicity
81,414
0.5
Pregnancy-induced hypertension
487,267
2.9
Birth place outside a hospital
147,597
0.9
11,639,654
69.6
Eligible
after 39 weeks’ gestation. The pattern at ET gestations was strikingly similar to the latter. Access to outpatient obstetric care is readily available on the weekdays. Women likely receive more frequent evaluations on weekdays, and, in turn, their obstetrician may find more reasons for IOL on weekdays.10 Yet, why would the difference in frequency of IOL between weekdays and weekends increase with gestational age? There is no evidence to suggest that the reasons for IOL, specifically reasons picked up during outpatient office hours, should increase with gestational age.
This work is important because there is little reason to believe that unambiguous medical indications would vary over the days of the week,11 making the weekend–weekday imbalance in the frequency of IOL that increases with gestational age even more challenging to explain. Taken together, these findings support that the threshold to initiate IOL was likely influenced by nonmedical factors, although the identity and the significance of these factors remain uncertain. This study encompasses all gravid women in the United States between 2007 and 2010, and thus, the reported
Table 2 Characteristics of the eligible population (2007–2010) Gestational age at birth (wks)a
Days of the wka
All
< 34
34–36
37–38
39
Weekday
Weekend
Deliveries (N)
148,437
594,337
3,119,748
7,777,132
9,171,318
2,468,336
11,639,654
Inductions
11,602
93,438
710,744
2,358,430
2,741,367
432,847
3,174,214
Proportion of inductions
7.8%
15.7%
22.8%
30.3%
29.9%
17.5%
27.3%
26.6 6.6
26.8 6.3
27.1 6.1
27.0 6.0
27.0 6.1
26.8 6.0
27.0 6.1
Maternal characteristics Age SD (y) Advanced maternal age
13.9%
13.2%
12.9%
11.9%
12.4%
11.7%
12.3%
Teen pregnancy
15.6%
13.0%
11.1%
11.4%
11.3%
12.2%
11.5%
Multiparous
46.8%
55.3%
59.5%
51.7%
54.5%
52.1%
53.9%
Married
45.7%
53.3%
58.2%
58.8%
58.7%
56.3%
58.2%
Diabetes mellitus
5.7%
6.4%
5.5%
3.3%
4.2%
3.6%
4.1%
Chronic hypertension
4.9%
2.8%
1.6%
0.7%
1.2%
1.0%
1.1%
Abbreviation: SD, standard deviation. a p < 0.001 for all comparisons between gestational age strata and weekend versus weekday. American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Exclusion
Variation in Labor Induction over the Days of the Week
Falciglia et al.
Fig. 1 Proportion of deliveries by induction of labor according to the day of the week. Results are stratified by gestational age at birth.
results represent the national landscape of births and IOL over a recent period. Several limitations of this study should be noted: unmeasured confounding bias may exist in an analysis of secondary data. Moreover, coding error is a concern from this administrative dataset, such that women may have been misclassified as to whether they received IOL. Though the validity of this measure has been challenged, the Centers for Disease Control and Prevention and numerous previous investigators have used these data to examine national trends in births and labor inductions for both infants and mothers.12–15 These datasets remain the standard for epidemiologic analyses of US births. Even if misclassification existed, it was likely to be
random and not systematically related to the day of a woman’s delivery. This study cannot define all of the factors that may account for the change in the frequency in IOL by gestational age or for the origins in the day-to-day differences that were observed. Whether the day-to-day variation in the frequency IOL is due to differences within or between institutions, variation across providers, patient choice (e.g., selections of providers and/or birth hospitals), or the combination of these factors remain uncertain. Most importantly, the degrees to which these reported differences are associated with maternal and perinatal outcomes, including stillbirth, also are not known. Further research
Table 3 Unadjusted and multivariable analysis Variable
Gestational age at delivery (wks) < 34
34–36
37–38
39
0.92 (0.87–0.98)
0.77 (0.75–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
þ Year of delivery
0.92 (0.87–0.98)
0.77 (0.75–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
þ Maternal covariates
0.93 (0.87–0.98)
0.77 (0.76–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
Unadjusted Weekend delivery Adjusted
Note: Analyses are referent to women delivering on weekdays and are adjusted for maternal age, parity, marital status, the presence of maternal chronic hypertension and maternal diabetes mellitus, and year of delivery. Odds ratio and 99% confidence intervals are presented.
Table 4 Sensitivity analysis Variable
Gestational age at delivery (wk) < 34
34–36
37–38
39
Monday as weekend (Sat–Mon vs. Tues–Fri)
0.88 (0.85–0.93)
0.73 (0.71–0.74)
0.62 (0.61–0.62)
0.62 (0.62–0.62)
Monday omitted (Sat–Sun vs. Tues–Fri)
0.89 (0.85–0.95)
0.73 (0.72–0.75)
0.50 (0.50–0.51)
0.47 (0.47–0.47)
Note: Analyses are referent to women delivering on Tuesday to Friday and are adjusted for maternal age, parity, marital status, the presence of maternal chronic hypertension and maternal diabetes mellitus, and year of delivery. Odds ratio and 99% confidence intervals are presented.
American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
110
Variation in Labor Induction over the Days of the Week
7 Murthy K, Grobman WA, Lee TA, Holl JL. Obstetricians’ rising
8 9
References 1 Murthy K, Holl JL, Lee TA, Grobman WA. National trends and racial
2
3
4 5
6
differences in late preterm induction. Am J Obstet Gynecol 2011; 205(5):458e1–458e7 Murthy K, Grobman WA, Lee TA, Holl JL. Trends in induction of labor at early-term gestation. Am J Obstet Gynecol 2011;204(5): 435e1–435e6 Engle WA, Kominiarek MA. Late preterm infants, early term infants, and timing of elective deliveries. Clin Perinatol 2008; 35(2):325–341, vi ACOG Committee Opinion No 579: Definition of term pregnancy. Obstet Gynecol 2013;122(5):1139–1140 Hibbard JU, Wilkins I, Sun L, et al; Consortium on Safe Labor. Respiratory morbidity in late preterm births. JAMA 2010;304(4): 419–425 Bailit JL, Gregory KD, Reddy UM, et al. Maternal and neonatal outcomes by labor onset type and gestational age. Am J Obstet Gynecol 2010;202(3):245.e1–245.e12
111
10
11
12 13
14
15
liability insurance premiums and inductions at late preterm gestations. Med Care 2009;47(4):425–430 Gould JB, Qin C, Marks AR, Chavez G. Neonatal mortality in weekend vs weekday births. JAMA 2003;289(22):2958–2962 National Center for Health Statistics. National Vital Statistics System, 2007-2010. Available at http://www.cdc.gov/nchs/data_access/Vitalstatsonline.htm. Last accessed: October 09, 2012 Gyamfi-Bannerman C, Fuchs KM, Young OM, Hoffman MK. Nonspontaneous late preterm birth: etiology and outcomes. Am J Obstet Gynecol 2011;205(5):456e1–456e6 ACOG Committee on Practice Bulletins – Obstetrics. ACOG Practice Bulletin No. 107: Induction of labor. Obstet Gynecol 2009;114(2 Pt 1): 386–397 Heron M, Sutton PD, Xu J, Ventura SJ, Strobino DM, Guyer B. Annual summary of vital statistics: 2007. Pediatrics 2010;125(1):4–15 Mathews TJ, Miniño AM, Osterman MJK, Strobino DM, Guyer B. Annual summary of vital statistics: 2008. Pediatrics 2011;127(1): 146–157 Kochanek KD, Kirmeyer SE, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2009. Pediatrics 2012;129(2): 338–348 Hamilton BE, Hoyert DL, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2010-2011. Pediatrics 2013;131(3): 548–558
American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
will need to investigate the underlying variation within and across institutions, as well as among obstetric providers and patients/families that have led to the day-to-day variation in IOL in the United States.
Falciglia et al.
Copyright of American Journal of Perinatology is the property of Thieme Medical Publishing Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
107
Variation in Labor Induction over the Days of the Week William A. Grobman, MD, MBA2
1 Division of Neonatology, Department of Pediatrics, Ann & Robert
H. Lurie Children’s Hospital of Chicago, Northwestern University, Chicago, Illinois 2 Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Prentice Women’s Hospital-Northwestern Memorial Hospital, Northwestern University, Chicago, Illinois
Karna Murthy, MD, MSc1
Address for correspondence Gustave H. Falciglia, MD, Division of Neonatology, Department of Pediatrics, Ann & Robert H. Lurie Children’s Hospital of Chicago, Feinberg School of Medicine, Northwestern University, 225E. Chicago Avenue, Box 45, Chicago, IL, 60611 (e-mail: [email protected]).
Am J Perinatol 2015;32:107–112.
Abstract
Keywords
► ► ► ► ►
induction of labor late preterm early term variation epidemiology
Objective To quantify the variation in induction of labor (IOL) over the days of the week for gravid women in the United States. Study Design Women who delivered singletons between 24 and 42 weeks’ gestation were identified using birth certificate data from 2007 to 2010. Women with pregnancyassociated hypertension, fetal anomalies, previous cesarean delivery, or incomplete records were excluded. The primary outcome was IOL. Women were stratified into four gestational age groups: < 34, 34 to 36, 37 to 38, and 39 weeks. Frequencies of IOL were determined according to day of the week. Multivariable logistic regression estimated the association between weekend delivery and IOL, adjusting for maternal characteristics and year of delivery. Results There were 11.6 million eligible women. For each gestational age stratum, the frequency of IOL was increased on weekdays compared with weekends (8.0 vs. 7.4%, 16.5 vs. 13.2%, 25.0 vs. 14.8%, and 33.2 vs. 19.3% at < 34, 34–36, 37–38, and 39 weeks, respectively; p < 0.01 for all). Multivariable analyses demonstrated that weekend IOL was inversely related to gestational age (odds ratios: 0.93, 0.77, 0.52, and 0.48, respectively; p < 0.001 for all). Conclusion The frequency of IOL varies according to the day of the week, with the odds of weekend IOL lowest at greater gestational ages.
Background The frequency of induction of labor (IOL) among women with late preterm (LP: 34 0/7–36 6/7 weeks) and early term (ET: 37 0/7– 38 6/7 weeks) gestations has risen over the last two decades.1–4 Relative to infants born at or after 39 weeks’ gestation, infants born during LP or ET gestations are more likely to experience infant mortality, respiratory difficulties, and admission to the neonatal intensive care unit.3–6 Despite these documented associations, the factors that have contributed to the rise in the frequency of LP and ET inductions remain uncertain.
received April 10, 2014 accepted April 11, 2014 published online May 28, 2014
Although, multiple unambiguous medical indications exist for IOL before 39 weeks’ gestation, others are less established; correspondingly, nonmedical factors may have contributed to the rise in IOL at LP and ET gestations.1,2,7 One nonmedical factor that has been unexplored is the inclination to provide IOL on certain days of the week. Although, Gould et al demonstrated that cesarean delivery varied over the days of the week, whether IOL exhibits similar variability is uncertain.8 Variation in IOL over the days of the week, and particularly variation that is smaller in magnitude at earlier gestations, provides support that a nonmedical influence may
Copyright © 2015 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0034-1376392. ISSN 0735-1631.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Gustave H. Falciglia, MD1
Variation in Labor Induction over the Days of the Week have existed. Thus, we aimed to quantify the variation in IOL in the United States over the days of the week as a function of gestational age.
Methods Birth certificate data, and detailed descriptions of these data, from the National Vital Statistics System (NVSS) were used to identify gravid women and their infants delivering from 2007 to 2010 in the United States.9 The study was approved by the Ann & Robert H. Lurie Children’s Hospital of Chicago Research Center Institutional Review Board. Women aged 13 to 49 who delivered a nonanomalous singleton infant between 24 and 42 weeks’ gestation were included in the analysis. Women without prenatal care or who delivered outside a hospital were excluded because IOL is a planned event that takes place within a hospital. Women with previous cesarean deliveries or nonvertex fetal presentation were excluded because IOL is less likely in these situations. Women with pregnancy-induced hypertension or premature rupture of membranes were omitted because the timing of onset and severity of these conditions were unknown; thus, it cannot be known whether women with these conditions at any given gestational age received expectant management or underwent IOL at diagnosis. Finally, women whose records had incomplete data that did not allow ascertainment of the primary exposures and outcome (e.g., gestational age at delivery, IOL, and day of the week of delivery) were omitted from the analysis. The primary exposure was the day of the week of delivery, identified using the variable “day of the week” in the NVSS dataset.9 The weekend was defined as “Saturday through Sunday,” while the weekday was defined as “Monday through Friday.” Using the variable “obstetric/clinical gestation estimate” to determine the gestational age at delivery, we stratified the eligible women into four groups: < 34 0/7, 34 0/7 to 36 6/7, 37 0/7 to 38 6/7, and 39 0/7 weeks’ gestations. The primary outcome was IOL and was identified using the variable “induction of labor” from the birth certificates.9 The maternal covariates considered in the analyses were the year of delivery, maternal age, parity, marital status, and maternal diabetes mellitus, and chronic hypertension. Referent to women between 20 and 34 years old, maternal age was categorized as advanced maternal age ( 35 years old) and teen pregnancy ( 19 years old). For each gestational age and weekday strata, the number of deliveries, proportion of deliveries after IOL, and maternal covariates were described. The proportion of deliveries following IOL was calculated by dividing the number of women who underwent labor induction by the number of women who were eligible to receive IOL on a particular day of the week. This proportion was calculated for weekend and weekday, where the denominators were the total number of deliveries on the weekend and weekday, respectively. Significant differences were calculated using student’s t test, or chi-square analyses, as appropriate. Logistic regression was used to estimate the unadjusted and the independent associations between the weekend and delivery after IOL American Journal of Perinatology
Vol. 32
No. 1/2015
Falciglia et al. for women delivering in each gestational age stratum. Equations were adjusted for the listed maternal covariates and year of delivery. Due to the lag between the initiation of IOL and delivery, sensitivity analyses considered different definitions of weekends and weekdays. Thus, “Monday” was considered as a weekend day in one analysis, while omitted in another. IOL is a well-accepted practice postterm. Many hospitals may not schedule IOL on weekends; therefore, we completed an additional sensitivity analysis omitting women delivering after 41 weeks’ gestation. Analyses were performed with STATA v12 (StataCorp, College Station, TX). Given the large sample sizes, a twotailed α ¼ 0.01 was used to define statistical significance.
Results Of the women who delivered in 2007 to 2010 in the United States, 11,639,654 were eligible for analysis (►Table 1). The differences in population characteristics stratified by gestational age and weekend delivery are reported in ►Table 2. Although women’s characteristics differed significantly according to the days of the week on which they delivered, the magnitude of these differences was small. IOL was significantly more frequent at higher gestational ages. Also, although weekday deliveries accounted for a majority of deliveries, the proportion of women delivering on weekdays (78.79%) was greater than that expected by chance (5/7th or 71.42%). This disproportionate number of weekday deliveries is related to greater frequency of delivery after weekday IOL (►Fig. 1). Women delivering 39 weeks’ gestation demonstrated the greatest day-to-day variation, with the magnitude of this variation declining with decreasing gestation. Deliveries due to IOL varied minimally for women delivering < 34 weeks’ gestation. When the days of the week were aggregated into weekends and weekdays, the proportion of deliveries after IOL was increased on weekdays relative to weekends in each gestational age stratum (8.0 vs. 7.4%, 16.5 vs. 13.2%, 25.0 vs. 14.8%, and 33.2 vs. 19.3% at < 34 0/7, 34 0/7 to 36 6/7, 37 0/7 to 38 6/7, and 39 0/7 weeks’ gestational age, respectively; p < 0.01 for all). In multivariable analyses, these differences persisted after adjusting for year of birth, parity, maternal age, the presence of diabetes mellitus, chronic hypertension, and marital status (►Table 3). Sensitivity analyses considering women delivered on Monday as part of the “weekend” group and also omitting women delivering on Mondays from any equation are presented in ►Table 4. The association between weekend delivery and IOL were essentially unchanged. Lastly, excluding women delivering beyond 40 6/7 weeks’ gestation did not change results from the primary analyses (data not shown).
Discussion The frequency of delivery after IOL was greatest on weekdays. The difference in IOL frequency between weekday and weekend was least before 34 weeks’ gestation and greatest at or
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
108
Variation in Labor Induction over the Days of the Week
Falciglia et al.
109
Table 1 Composition of the cohort Frequency
US deliveries 2007–2010
Percentage
16,724,105
100.0
Deliver < 24 or > 42 wks’ gestational age
105,437
0.6
Nonsingleton
562,177
3.4
Maternal age
1,802
0.0
Unknown maternal diabetes or chronic hypertension
75,887
0.5
Fetal or neonatal anomalies
125,817
0.8
Unknown initiation of labor
10,818
0.1
Previous cesarean delivery
2,174,437
13.0
Nonvertex presentation
868,355
5.2
Premature rupture of membranes
268,625
1.6
No prenatal care
174,818
1.1
Unknown Hispanic ethnicity
81,414
0.5
Pregnancy-induced hypertension
487,267
2.9
Birth place outside a hospital
147,597
0.9
11,639,654
69.6
Eligible
after 39 weeks’ gestation. The pattern at ET gestations was strikingly similar to the latter. Access to outpatient obstetric care is readily available on the weekdays. Women likely receive more frequent evaluations on weekdays, and, in turn, their obstetrician may find more reasons for IOL on weekdays.10 Yet, why would the difference in frequency of IOL between weekdays and weekends increase with gestational age? There is no evidence to suggest that the reasons for IOL, specifically reasons picked up during outpatient office hours, should increase with gestational age.
This work is important because there is little reason to believe that unambiguous medical indications would vary over the days of the week,11 making the weekend–weekday imbalance in the frequency of IOL that increases with gestational age even more challenging to explain. Taken together, these findings support that the threshold to initiate IOL was likely influenced by nonmedical factors, although the identity and the significance of these factors remain uncertain. This study encompasses all gravid women in the United States between 2007 and 2010, and thus, the reported
Table 2 Characteristics of the eligible population (2007–2010) Gestational age at birth (wks)a
Days of the wka
All
< 34
34–36
37–38
39
Weekday
Weekend
Deliveries (N)
148,437
594,337
3,119,748
7,777,132
9,171,318
2,468,336
11,639,654
Inductions
11,602
93,438
710,744
2,358,430
2,741,367
432,847
3,174,214
Proportion of inductions
7.8%
15.7%
22.8%
30.3%
29.9%
17.5%
27.3%
26.6 6.6
26.8 6.3
27.1 6.1
27.0 6.0
27.0 6.1
26.8 6.0
27.0 6.1
Maternal characteristics Age SD (y) Advanced maternal age
13.9%
13.2%
12.9%
11.9%
12.4%
11.7%
12.3%
Teen pregnancy
15.6%
13.0%
11.1%
11.4%
11.3%
12.2%
11.5%
Multiparous
46.8%
55.3%
59.5%
51.7%
54.5%
52.1%
53.9%
Married
45.7%
53.3%
58.2%
58.8%
58.7%
56.3%
58.2%
Diabetes mellitus
5.7%
6.4%
5.5%
3.3%
4.2%
3.6%
4.1%
Chronic hypertension
4.9%
2.8%
1.6%
0.7%
1.2%
1.0%
1.1%
Abbreviation: SD, standard deviation. a p < 0.001 for all comparisons between gestational age strata and weekend versus weekday. American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Exclusion
Variation in Labor Induction over the Days of the Week
Falciglia et al.
Fig. 1 Proportion of deliveries by induction of labor according to the day of the week. Results are stratified by gestational age at birth.
results represent the national landscape of births and IOL over a recent period. Several limitations of this study should be noted: unmeasured confounding bias may exist in an analysis of secondary data. Moreover, coding error is a concern from this administrative dataset, such that women may have been misclassified as to whether they received IOL. Though the validity of this measure has been challenged, the Centers for Disease Control and Prevention and numerous previous investigators have used these data to examine national trends in births and labor inductions for both infants and mothers.12–15 These datasets remain the standard for epidemiologic analyses of US births. Even if misclassification existed, it was likely to be
random and not systematically related to the day of a woman’s delivery. This study cannot define all of the factors that may account for the change in the frequency in IOL by gestational age or for the origins in the day-to-day differences that were observed. Whether the day-to-day variation in the frequency IOL is due to differences within or between institutions, variation across providers, patient choice (e.g., selections of providers and/or birth hospitals), or the combination of these factors remain uncertain. Most importantly, the degrees to which these reported differences are associated with maternal and perinatal outcomes, including stillbirth, also are not known. Further research
Table 3 Unadjusted and multivariable analysis Variable
Gestational age at delivery (wks) < 34
34–36
37–38
39
0.92 (0.87–0.98)
0.77 (0.75–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
þ Year of delivery
0.92 (0.87–0.98)
0.77 (0.75–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
þ Maternal covariates
0.93 (0.87–0.98)
0.77 (0.76–0.79)
0.52 (0.52–0.53)
0.48 (0.48–0.49)
Unadjusted Weekend delivery Adjusted
Note: Analyses are referent to women delivering on weekdays and are adjusted for maternal age, parity, marital status, the presence of maternal chronic hypertension and maternal diabetes mellitus, and year of delivery. Odds ratio and 99% confidence intervals are presented.
Table 4 Sensitivity analysis Variable
Gestational age at delivery (wk) < 34
34–36
37–38
39
Monday as weekend (Sat–Mon vs. Tues–Fri)
0.88 (0.85–0.93)
0.73 (0.71–0.74)
0.62 (0.61–0.62)
0.62 (0.62–0.62)
Monday omitted (Sat–Sun vs. Tues–Fri)
0.89 (0.85–0.95)
0.73 (0.72–0.75)
0.50 (0.50–0.51)
0.47 (0.47–0.47)
Note: Analyses are referent to women delivering on Tuesday to Friday and are adjusted for maternal age, parity, marital status, the presence of maternal chronic hypertension and maternal diabetes mellitus, and year of delivery. Odds ratio and 99% confidence intervals are presented.
American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
110
Variation in Labor Induction over the Days of the Week
7 Murthy K, Grobman WA, Lee TA, Holl JL. Obstetricians’ rising
8 9
References 1 Murthy K, Holl JL, Lee TA, Grobman WA. National trends and racial
2
3
4 5
6
differences in late preterm induction. Am J Obstet Gynecol 2011; 205(5):458e1–458e7 Murthy K, Grobman WA, Lee TA, Holl JL. Trends in induction of labor at early-term gestation. Am J Obstet Gynecol 2011;204(5): 435e1–435e6 Engle WA, Kominiarek MA. Late preterm infants, early term infants, and timing of elective deliveries. Clin Perinatol 2008; 35(2):325–341, vi ACOG Committee Opinion No 579: Definition of term pregnancy. Obstet Gynecol 2013;122(5):1139–1140 Hibbard JU, Wilkins I, Sun L, et al; Consortium on Safe Labor. Respiratory morbidity in late preterm births. JAMA 2010;304(4): 419–425 Bailit JL, Gregory KD, Reddy UM, et al. Maternal and neonatal outcomes by labor onset type and gestational age. Am J Obstet Gynecol 2010;202(3):245.e1–245.e12
111
10
11
12 13
14
15
liability insurance premiums and inductions at late preterm gestations. Med Care 2009;47(4):425–430 Gould JB, Qin C, Marks AR, Chavez G. Neonatal mortality in weekend vs weekday births. JAMA 2003;289(22):2958–2962 National Center for Health Statistics. National Vital Statistics System, 2007-2010. Available at http://www.cdc.gov/nchs/data_access/Vitalstatsonline.htm. Last accessed: October 09, 2012 Gyamfi-Bannerman C, Fuchs KM, Young OM, Hoffman MK. Nonspontaneous late preterm birth: etiology and outcomes. Am J Obstet Gynecol 2011;205(5):456e1–456e6 ACOG Committee on Practice Bulletins – Obstetrics. ACOG Practice Bulletin No. 107: Induction of labor. Obstet Gynecol 2009;114(2 Pt 1): 386–397 Heron M, Sutton PD, Xu J, Ventura SJ, Strobino DM, Guyer B. Annual summary of vital statistics: 2007. Pediatrics 2010;125(1):4–15 Mathews TJ, Miniño AM, Osterman MJK, Strobino DM, Guyer B. Annual summary of vital statistics: 2008. Pediatrics 2011;127(1): 146–157 Kochanek KD, Kirmeyer SE, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2009. Pediatrics 2012;129(2): 338–348 Hamilton BE, Hoyert DL, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2010-2011. Pediatrics 2013;131(3): 548–558
American Journal of Perinatology
Vol. 32
No. 1/2015
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
will need to investigate the underlying variation within and across institutions, as well as among obstetric providers and patients/families that have led to the day-to-day variation in IOL in the United States.
Falciglia et al.
Copyright of American Journal of Perinatology is the property of Thieme Medical Publishing Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
