Background
Primary cesarean deliveries are a major contributor to the large increase in cesarean delivery rates in the United States over the past 2 decades and are an essential focus for the reduction of related morbidity and costs. Studies have shown that primary cesarean delivery rates among low-risk women in the United States vary 3-fold across hospitals and are not explained by differences in patient case-mix. However, the extent to which maternal vs hospital characteristics contribute to this variation remains poorly understood because previous studies were limited in scope and did not assess the influence of factors such as maternal ethnicity subgroups or prepregnancy obesity.
Objective
We assessed the contribution of individual- and hospital-level risk factors to the hospital variation in primary cesarean delivery rates among low-risk women in Florida.
Study Design
Our population-based retrospective cohort study used Florida’s linked birth certificate and hospital discharge records for the period of 2004-2011. The study population was comprised of 412,192 nulliparous, singleton, vertex, live births with labor at 37-40 weeks gestation in 122 nonmilitary delivery hospitals. Data were analyzed with logistic mixed-effects regression with cesarean delivery as the outcome. This approach provided adjusted risk estimates at an individual and hospital level and the estimated percent of hospital variation statewide that was explained by these factors.
Results
The primary cesarean delivery rate in the study population was 23.9%, with hospital-specific estimates that ranged from 12.8-47.3%. Leading risk factors for cesarean delivery were maternal age ≥35 years (adjusted relative risk, 2.22), prepregnancy obesity (body mass index, ≥30 kg/m 2 ; adjusted relative risk, 1.73), medical risk conditions (adjusted relative risk, 1.72), labor induction (adjusted relative risk, 1.52), and delivery in hospitals located in Miami-Dade County (adjusted relative risk, 1.73). Hospital geographic location was a significant effect modifier for prepregnancy obesity, medical conditions, and labor induction ( P < .05), with a tendency towards lower adjusted relative risks for these factors in Miami-Dade County relative to other Florida regions. Conversely, Miami-Dade County had an increased prevalence of higher-risk ethnic subgroups, such as Cuban or Puerto Rican mothers, and also substantially higher adjusted relative risks that were associated with practice-related factors, such as delivery during weekday hours. Whereas hospital geographic location contributed to 39.6% of the observed variation statewide, the estimated contribution of maternal ethnicity ranged from 1.6-15.7% among Florida regions.
Conclusions
Hospital geographic location contributes to hospital variation in primary cesarean delivery rates among low-risk women in Florida. In contrast to previous studies, our findings suggest that individual level risk factors such as maternal ethnicity also contribute to some of this variation, with differing extent by region. These individual factors likely interact with practice factors and add to the variation. This study was limited by not including maternal Bishop score before induction or obstetrics provider in the analysis. These were not available on the dataset but likely contribute to the variation. Our findings suggest potential issues to consider in quality improvement efforts, such as the need for future qualitative research that focuses on mothers in higher-risk ethnic subgroups and providers in high-rate hospitals, particularly those in Miami-Dade County. These studies may help to identify potential cultural differences in maternal beliefs and expectations for delivery and maternal reasons for differences in obstetrics practices.
Comprising 60% of all cesarean deliveries in the United States, primary cesarean deliveries are a major contributor to the large increase in total cesarean delivery rates over the past 2 decades. Approximately 90% of women who have a primary cesarean delivery are likely to deliver by cesarean again in subsequent pregnancies, which incurs higher costs and progressively higher morbidity risks with each additional cesarean delivery. It is recognized that effective prevention efforts must focus on cesarean deliveries among low-risk women, that is, nulliparous women at term gestation, with a singleton baby in vertex presentation (NTSV).
The few studies that have assessed NTSV cesarean deliveries in the United States have found a 3-fold variation in rates across hospitals, which could not be attributed to differences in patient sociodemographic characteristics or medical diagnoses. These findings suggest hospital differences in quality of care and emphasize the need to better understand the contribution of maternal and hospital characteristics over and above the factors studied thus far. Particularly, risk factors such as maternal prepregnancy obesity and ethnic subgroups were not included in these previous analyses. We explored individual and hospital risk factors for a primary cesarean delivery among low-risk women in Florida using linked birth certificate and hospital discharge data to determine the extent to which they contribute to the hospital variation in cesarean delivery rates. Findings from this study can be used to inform patients, providers, and decision-makers and to assist with planning quality improvement efforts for the reduction of morbidity and costs that are related to cesarean deliveries.
Materials and Methods
We used Florida’s existing linked birth certificate and hospital discharge records, focusing on births that occurred between January 1, 2004, and December 31, 2011. Over the 8-year period, there were 1,634,884 live births to Florida residents who delivered in Florida’s nonmilitary hospitals with maternity service. With the use of social security numbers as the primary linkage variable, in addition to date and facility of birth, infant sex, and mother’s residential zip code, the record linkage captured 91.6% of all live births to Florida mothers in a Florida hospital during these years. The linkage excluded records with missing or invalid maternal social security numbers, which more than likely led to an exclusion of undocumented immigrant population births. Validity studies previously have shown improved accuracy across most clinical variables in linked data compared with the birth certificate or hospital discharge alone.
We initially selected the population that was considered to be at lowest risk for a cesarean delivery: nulliparous women, term (37-41 completed weeks of gestation), singleton, vertex births without malpresentation (n = 561,736). We excluded births from hospitals with <100 births per year (to include only routine delivery hospitals) and births without information for maternal age, race, ethnicity, and prepregnancy body mass index (BMI), which resulted in an initial sample of 495,059 NTSV births. We further excluded births from cesarean deliveries without labor (n = 45,645) and births at ≥41 completed weeks of gestation (n = 37,222), because of the potential differences in risk factors that are related to planning these deliveries. The final analytic sample was comprised of 412,192 NTSV births with labor at 37-40 completed weeks of gestation.
Each birth was categorized as being a cesarean or vaginal delivery based on the presence of a cesarean delivery procedure in either the birth certificate or hospital discharge. Individual and hospital factors were selected for analysis based on the literature and the potential clinical relevance to cesarean deliveries. The following individual level characteristics were included: maternal age, ethnicity (non-Hispanic white, non-Hispanic black or Haitian, Hispanic Mexican, Puerto Rican, Cuban, other Hispanic, other non-Hispanic), level of education, nativity (Florida-born, other US state-born, foreign-born), marital status, insurance status, prepregnancy BMI (kilograms per square meter), gestational age, presence of ≥1 medical risk conditions for delivery, onset of labor, day of week, and time of delivery. Non-Hispanic black and Haitian women were combined in the same ethnic subgroup based on their similar association with primary cesarean delivery after adjustment for significant risk factors. Presence of ≥1 medical risk conditions for delivery was determined based on the presence of any of the following 7 maternal or infant conditions: diabetes mellitus, hypertension, eclampsia, oligohydramnios, placental abruption, chorioamnionitis, and infant birthweight >4000 g. These were selected from an initial list of 11 conditions, which also included maternal infection, polyhydramnios, placenta previa, and cord prolapse based on their respective associations with the risk of cesarean delivery in the study population. Individual inclusion of these medical risk conditions did not improve the regression modeling. Onset of labor was categorized as either spontaneous (vaginal births without an induction, cesarean delivery births without an induction but with a trial or augmentation of labor) or induced (vaginal and cesarean delivery births with induction). Indications for cesarean delivery, such as fetal distress ( Supplemental Table 1 ), were not assessed as risk factors because they may represent subjective and/or ambiguous justifications that were recorded after the procedure.
Hospital-level characteristics were determined based on the facility and year of birth. The following items were ascertained directly from the birth certificate: hospital delivery size (live births/year), percentage of all live births delivered by certified nurse midwives (quartiles), percentage of all live births to Hispanic mothers, urbanicity based on classification from the National Center for Health Statistics, level of neonatal intensive care unit based on licensure of the Florida Agency for Health Care Administration, and ownership. Obstetric residency programs were identified with the list from the Accreditation Council for Graduate Medical Education and contact with accredited programs. Percentage of all live births to mothers with Medicaid insurance was calculated from the hospital discharge records and categorized by quartile. For geographic location, we divided the state into 3 regions from North to South based on the American Congress of Obstetricians and Gynecologists (ACOG) District XII section map. Miami-Dade County, which represents 14% of all births, was classified separately based on unadjusted and adjusted cesarean delivery rates. This resulted in 4 categories: (1) north (ACOG sections 1-2), (2) central (ACOG sections 3-4), (3) south (ACOG sections 5-6, with Miami-Dade County removed), and (4) Miami-Dade County (part of ACOG section 6).
Statistical analysis
Given the hierarchical structure of the data, with births nested within hospitals, we examined the risk factors for primary cesarean delivery using logistic mixed-effects regression models with hospital-specific random intercepts. The resulting odds ratios and 95% confidence intervals (CI) were converted to relative risks (RRs) according to the method described by Zhang and Yu ( Supplemental Table 2 ). We attempted to fit log-binomial models, which would provide direct RR estimates but could not accommodate all of our study variables and categories in this type of modeling. The first logistic model predicted the likelihood of a birth at a given hospital being a cesarean delivery, based solely on the overall risk of a cesarean delivery in the entire study population and the specific risk at each hospital. This provided baseline estimates of the variability in cesarean deliveries across hospitals (between-hospital variance) and hospital-specific differences in risk of cesarean delivery (relative to the weighted population average), without adjustment for individual or hospital characteristics. This model was then extended to adjust for individual and hospital level factors. The fully adjusted (multivariable) model retained only those factors that were associated significantly with the risk of cesarean delivery at the 5% level. Regional differences in cesarean delivery risk factors were assessed by rerunning the model that contained all significant individual level risk factors separately by geographic location and by running a model on the entire study population with interaction terms between hospital geographic location and any individual level risk factors for which the RR appeared to vary by region. The extent to which hospital differences in cesarean delivery rates are reflective of the clustering of individual and facility characteristics was assessed by a comparison of the reduction, if any, in the between-hospital variance estimate from the unadjusted model to the adjusted models. The modeling was carried out using the glimmix procedure in SAS software (version 9.4; SAS Institute Inc, Cary, NC). Sensitivity analyses included repeating the logistic mixed-effects modeling on the larger sample of 449,414 NTSV births with labor at 37-41 weeks of gestation and then on the full initial sample of 495,059 NTSV births with and without labor at 37-41 weeks of gestation.
Ethics
Data-use confidentiality agreements with the Florida Department of Health and Agency for Health Care Administration were completed by the study researchers. This study was approved by the institutional review boards at the University of South Florida and the Florida Department of Health.
Results
Among 412,192 births in the study population, 98,472 were from cesarean deliveries. The primary cesarean delivery rate was relatively stable from 2004-2011, with a population average of 23.9% and hospital-specific rates that varied from 12.8-47.3% across 122 facilities. Individual and hospital level characteristics of the study population are summarized in Table 1 . Among cesarean deliveries, there were marked differences across hospital geographic location in the proportion of mothers of Cuban and Hispanic ethnicity other than Mexican or Puerto Rican, foreign-born mothers, deliveries between 8 am and 8 pm , hospitals with level III neonatal intensive care units, hospitals ranked in the highest quartile of Medicaid or Hispanic births, hospitals in the lowest quartile of certified nurse midwife births, and hospitals located in large central metropolitan areas. All of these characteristics were most prevalent in Miami-Dade County.
| Variable | Live births, n | Vaginal delivery births, % | Cesarean delivery births, % | ||||
|---|---|---|---|---|---|---|---|
| Overall | North Florida | Central Florida | South Florida | Miami-Dade County | |||
| Total | 412,192 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 |
| Age, y | |||||||
| <20 | 84,842 | 22.5 | 14.4 | 16.9 | 15.5 | 12.1 | 12.7 |
| 20-29 | 237,352 | 58.1 | 56.0 | 60.7 | 55.8 | 54.5 | 53.5 |
| 30-34 | 62,067 | 13.9 | 18.6 | 14.9 | 18.1 | 20.3 | 21.4 |
| ≥35 | 27,931 | 5.5 | 11.0 | 7.5 | 10.6 | 13.1 | 12.5 |
| Race and ethnicity | |||||||
| Non-Hispanic white | 206,774 | 51.5 | 45.9 | 62.4 | 56.6 | 40.3 | 12.0 |
| Non-Hispanic black or Haitian | 81,033 | 19.4 | 20.6 | 25.2 | 16.7 | 26.8 | 16.3 |
| Hispanic | |||||||
| Mexican | 17,981 | 4.6 | 3.5 | 1.9 | 4.4 | 4.3 | 2.1 |
| Puerto Rican | 21,901 | 5.3 | 5.5 | 1.9 | 8.5 | 4.3 | 3.7 |
| Cuban | 22,923 | 4.7 | 8.4 | 0.7 | 2.1 | 5.5 | 34.9 |
| Other | |||||||
| Hispanic | 40,661 | 9.5 | 11.1 | 2.2 | 5.8 | 13.7 | 29.2 |
| Non-Hispanic | 20,919 | 5.1 | 5.0 | 5.7 | 6.0 | 5.2 | 1.9 |
| Education | |||||||
| Less than high school | 69,978 | 18.0 | 13.9 | 14.6 | 14.4 | 12.9 | 13.0 |
| High school graduate | 125,659 | 30.6 | 30.0 | 31.6 | 28.6 | 28.5 | 33.8 |
| Some college or more | 215,132 | 51.1 | 55.8 | 53.6 | 56.9 | 57.8 | 53.0 |
| Unknown | 1,423 | 0.4 | 0.3 | 0.3 | 0.2 | 0.7 | 0.2 |
| Nativity | |||||||
| Florida | 166,089 | 41.0 | 38.1 | 52.6 | 38.8 | 32.8 | 28.8 |
| Other US state | 138,027 | 33.8 | 32.5 | 36.6 | 40.6 | 31.0 | 11.3 |
| Foreign | 108,076 | 25.2 | 29.5 | 10.8 | 20.6 | 36.1 | 59.9 |
| Married | 195,159 | 46.5 | 50.0 | 49.1 | 50.0 | 52.1 | 48.0 |
| Insurance status | |||||||
| Medicaid | 196,635 | 48.4 | 45.6 | 46.3 | 44.3 | 42.0 | 52.7 |
| Private/other | 202,031 | 48.2 | 51.8 | 51.8 | 53.9 | 53.7 | 44.1 |
| Self-pay | 13,526 | 3.5 | 2.7 | 1.9 | 1.9 | 4.2 | 3.2 |
| Body mass index, kg/m 2 | |||||||
| Underweight (<18.5) | 27,373 | 7.5 | 3.8 | 3.5 | 3.9 | 3.5 | 4.1 |
| Normal (18.5-24.9) | 233,699 | 59.6 | 47.4 | 41.4 | 46.1 | 49.3 | 53.8 |
| Overweight (25-29.9) | 89,120 | 20.6 | 24.9 | 25.1 | 24.5 | 25.2 | 25.0 |
| Obese (≥30) | 62,000 | 12.2 | 24.0 | 29.9 | 25.5 | 22.1 | 17.1 |
| Gestational age, wk | |||||||
| 37 | 37,050 | 9.3 | 8.1 | 8.3 | 8.0 | 8.4 | 7.8 |
| 38 | 89,324 | 22.2 | 20.0 | 19.6 | 19.4 | 20.8 | 21.1 |
| 39 | 149,788 | 36.8 | 35.0 | 36.3 | 34.6 | 34.7 | 35.0 |
| 40 | 136,030 | 31.8 | 36.9 | 35.8 | 38.1 | 36.2 | 36.2 |
| ≥1 Medical risk conditions a | 110,590 | 22.3 | 41.2 | 46.9 | 43.2 | 38.6 | 34.4 |
| Induced labor | 169,774 | 37.0 | 54.6 | 55.6 | 58.4 | 59.8 | 37.4 |
| Delivery on Monday-Friday | 323,520 | 77.8 | 80.8 | 80.9 | 80.3 | 80.2 | 83.1 |
| Delivery time of day | |||||||
| 12:00 -03:59 am | 45,088 | 11.5 | 9.2 | 10.5 | 10.1 | 8.5 | 6.8 |
| 04:00-07:59 am | 41,410 | 10.8 | 7.8 | 8.6 | 8.5 | 7.1 | 6.1 |
| 08:00-11:59 am | 62,172 | 15.3 | 14.5 | 13.3 | 13.6 | 13.8 | 18.5 |
| 12:00-03:59 pm | 91,686 | 22.9 | 20.2 | 17.9 | 19.2 | 20.2 | 24.6 |
| 04:00-07:59 pm | 99,911 | 23.2 | 27.5 | 27.3 | 27.0 | 28.3 | 27.7 |
| 08:00-11:59 pm | 71,925 | 16.4 | 20.9 | 22.4 | 21.6 | 22.0 | 16.4 |
| Hospital size, births/y | |||||||
| <500 | 7,592 | 1.9 | 1.7 | 3.5 | 2.0 | 1.2 | 0.1 |
| 500-999 | 34,711 | 8.7 | 7.5 | 9.7 | 11.6 | 4.3 | 0.0 |
| ≥1000 | 369,889 | 89.4 | 90.8 | 86.7 | 86.4 | 94.6 | 99.9 |
| Hospital level of neonatal intensive care unit | |||||||
| I | 112,090 | 27.6 | 25.8 | 38.2 | 28.2 | 20.4 | 14.8 |
| II | 126,273 | 30.7 | 30.4 | 30.4 | 24.4 | 34.6 | 38.8 |
| III | 173,829 | 41.7 | 43.8 | 31.4 | 47.4 | 45.0 | 46.5 |
| Hospital Medicaid births, % | |||||||
| <35 | 96,198 | 22.9 | 24.7 | 21.7 | 19.5 | 26.1 | 38.0 |
| 35-45 | 105,302 | 25.6 | 25.5 | 14.5 | 35.8 | 28.4 | 8.5 |
| 45-58 | 105,837 | 26.1 | 24.4 | 41.9 | 21.1 | 28.2 | 9.0 |
| ≥59 | 104,855 | 25.5 | 25.4 | 21.9 | 23.6 | 17.3 | 44.6 |
| Hospital certified nurse midwife births, % | |||||||
| <20 | 310,173 | 74.2 | 78.5 | 77.0 | 79.4 | 64.0 | 98.5 |
| 20-29 | 60,365 | 15.2 | 12.8 | 17.6 | 13.1 | 17.2 | 1.1 |
| ≥30 | 41,654 | 10.6 | 8.7 | 5.5 | 7.5 | 18.8 | 0.5 |
| Hospital Hispanic births, % | |||||||
| <11 | 107,653 | 27.0 | 23.2 | 79.1 | 21.4 | 1.8 | 1.0 |
| 11-22 | 106,511 | 26.4 | 24.1 | 20.3 | 28.1 | 33.3 | 5.2 |
| 23-35 | 104,350 | 25.1 | 26.1 | 0.6 | 40.6 | 36.3 | 3.5 |
| ≥36 | 93,678 | 21.5 | 26.7 | 0.0 | 10.0 | 28.6 | 90.4 |
| Hospital urbanicity | |||||||
| Large central metro | 238,160 | 56.3 | 62.4 | 29.0 | 55.5 | 71.4 | 100.0 |
| Large fringe metro | 35,283 | 8.9 | 7.4 | 7.6 | 15.0 | 0.0 | 0.0 |
| Medium metro | 100,694 | 25.2 | 21.8 | 30.5 | 25.2 | 25.5 | 0.0 |
| Small/micro/nonmetro | 38,055 | 9.5 | 8.3 | 32.9 | 4.3 | 3.1 | 0.0 |
| Hospital ownership | |||||||
| Not-for-profit | 249,217 | 61.0 | 58.9 | 61.9 | 74.0 | 37.5 | 50.4 |
| Investor owned | 109,718 | 26.5 | 26.9 | 35.5 | 16.4 | 35.4 | 30.6 |
| Local owned | 53,257 | 12.5 | 14.3 | 2.7 | 9.5 | 27.1 | 19.0 |
| Obstetric residency program | 60,911 | 15.1 | 13.9 | 17.9 | 21.8 | 0.0 | 11.1 |
| Hospital geographic location | |||||||
| North Florida | 17,332 | 22.5 | 17.6 | 100.0 | — | — | — |
| Central Florida | 40,039 | 43.1 | 40.7 | — | 100.0 | — | — |
| South Florida | 24,099 | 23.5 | 24.5 | — | — | 100.0 | — |
| Miami-Dade county | 17,002 | 10.9 | 17.3 | — | — | — | 100.0 |
a Diabetes mellitus, hypertension, eclampsia, oligohydramnios, placental abruption, chorioamnionitis, and infant birthweight >4000 g.
Detailed unadjusted cesarean delivery rates for all individual and hospital level characteristics that were analyzed are provided in Supplemental Table 3 . Across individual level factors, the highest unadjusted cesarean delivery rates were observed in mothers who were ≥35 years old (38.8%), Cuban mothers (35.9%), obese mothers (38.1%), mothers with ≥1 medical risk conditions (36.7%), and mothers who had an induction (31.7%). After full adjustment for all significant risk factors in the multivariable model, the risk of cesarean delivery was twice as high in mothers ≥35 years old compared with mothers <20 years old (adjusted RR [ARR], 2.22; 95% CI, 2.17–2.27), 34% higher in Cuban mothers compared with non-Hispanic white mothers (ARR, 1.34; 95% CI, 1.30–1.38), 73% higher for prepregnancy obesity vs normal weight (ARR, 1.73; 95% CI, 1.71–176), 72% higher for medical risk conditions vs no conditions (ARR, 1.72; 95% CI, 1.70–1.74), and 55% higher for induced vs spontaneous labor (ARR, 1.52; 95% CI, 1.53–1.57). Non-Hispanic black or Haitian, Puerto Rican and other Hispanic ethnicity (excluding Mexican), insurance status other than self-pay, overweight, gestational age of 40 and 39 weeks, and delivery during midmorning, afternoon, and night hours on any day of the week were additional moderate risk factors ( Table 2 ).
| Variable | Cesarean delivery, % | Relative risk (95% confidence interval) a | Adjusted relative risk (95% confidence interval) | Change in between-hospital variance (σ 2 ), % |
|---|---|---|---|---|
| Age, y | 4.0 | |||
| <20 | 16.7 | 1.00 | 1.00 | |
| 20-29 | 23.2 | 1.39 (1.36–1.41) b | 1.32 (1.30–1.34) b | |
| 30-34 | 29.5 | 1.74 (1.71–1.78) b | 1.72 (1.68–1.75) b | |
| ≥35 | 38.8 | 2.29 (2.25–2.34) b | 2.22 (2.17–2.27) b | |
| Race and ethnicity | –0.7 c | |||
| Non-Hispanic white | 21.9 | 1.00 | 1.00 | |
| Non-Hispanic black or Haitian | 25.0 | 1.11 (1.10–1.13) b | 1.21 (1.19–1.23) b | |
| Hispanic | ||||
| Mexican | 19.3 | 0.87 (0.84–0.90) b | 1.01 (0.97–1.04) | |
| Puerto Rican | 24.5 | 1.09 (1.06–1.12) b | 1.20 (1.17–1.23) b | |
| Cuban | 35.9 | 1.30 (1.26–1.33) b | 1.34 (1.31–1.38) b | |
| Other | ||||
| Hispanic | 26.9 | 1.05 (1.03–1.07) d | 1.14 (1.11–1.16) b | |
| Non-Hispanic | 23.7 | 1.06 (1.03–1.09) b | 1.18 (1.15–1.22) b | |
| Married | 25.2 | 1.10 (1.09–1.12) b | 0.95 (0.93–0.96) b | –0.2 c |
| Insurance status | –0.2 c | |||
| Medicaid | 22.8 | 1.19 (1.15–1.24) b | 1.17 (1.13–1.22) b | |
| Private/Other | 25.2 | 1.33 (1.28–1.37) b | 1.16 (1.12–1.21) b | |
| Self-pay | 19.5 | 1.00 | 1.00 | |
| Body mass index, kg/m 2 | 0.2 | |||
| Underweight (<18.5) | 13.7 | 0.69 (0.67–0.71) b | 0.77 (0.74–0.79) b | |
| Normal (18.5-24.9) | 20.0 | 1.00 | 1.00 | |
| Overweight (25-29.9) | 27.5 | 1.40 (1.38–1.42) b | 1.29 (1.27–1.31) b | |
| Obese (≥30) | 38.1 | 1.99 (1.97–2.01) b | 1.73 (1.70–1.75) b | |
| Gestational age, wk | 2.1 | |||
| 37 | 21.5 | 1.00 | 1.00 | |
| 38 | 22.1 | 1.02 (1.00–1.05) | 1.09 (1.06–1.12) b | |
| 39 | 23.0 | 1.07 (1.05–1.10) b | 1.19 (1.16–1.21) b | |
| 40 | 26.7 | 1.27 (1.24–1.29) b | 1.39 (1.36–1.42) b | |
| ≥1 Medical risk conditions e | 36.7 | 1.98 (1.96–2.00) b | 1.72 (1.70–1.74) b | –0.4 c |
| Induced labor | 31.7 | 1.81 (1.79–1.82) b | 1.52 (1.50–1.54) b | 0.1 |
| Delivery on Monday-Friday | 24.6 | 1.12 (1.11–1.14) b | 1.02 (1.00–1.03) d | –1.1 c |
| Delivery time of day | ||||
| 12:00-03:59 am | 20.2 | 1.10 (1.07–1.13) b | 1.05 (1.02–1.08) f | |
| 04:00-07:59 am | 18.5 | 1.00 | 1.00 | |
| 08:00-11:59 am | 22.9 | 1.21 (1.18–1.24) b | 1.22 (1.19–1.25) b | |
| 12:00-03:59 pm | 21.6 | 1.13 (1.10–1.15) b | 1.08 (1.05–1.10) b | |
| 04:00-07:59 pm | 27.1 | 1.43 (1.40–1.46) b | 1.32 (1.28–1.35) b | |
| 08:00-11:59 pm | 28.7 | 1.54 (1.51–1.58) b | 1.43 (1.39–1.46) b | |
| Hospital geographic location | –39.6 c | |||
| North Florida | 19.7 | 1.00 | 1.00 | |
| Central Florida | 22.9 | 1.14 (1.04–1.26) f | 1.16 (1.04–1.28) f | |
| South Florida | 24.6 | 1.21 (1.08–1.34) f | 1.18 (1.05–1.32) f | |
| Miami-Dade County | 33.2 | 1.67 (1.48–1.88) b | 1.73 (1.52–1.96) b |
a Converted from odds ratios using the method Zhang J, Yu KF. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280:1690-1
c Inclusion of risk factor in multivariable model results in decreased between-hospital hospital variance (σ 2 )
e Medical risk conditions: diabetes mellitus, hypertension, eclampsia, oligohydramnios, placental abruption, chorioamnionitis, and infant birthweight > 4000 g
In separate models by geographic location, there were substantial regional differences in the ARRs for prepregnancy obesity, medical risk conditions, and labor induction, with a general tendency of lower estimates in Miami-Dade County for all of these risk factors. However, the RR of cesarean delivery for mothers of Cuban or Puerto Rican ethnicity had the opposite geographic trend, increasing slightly from North Florida to Miami-Dade County. In addition, the RR for weekdays and daytime hours were considerably higher in Miami-Dade County ( Table 3 ). Note that, at this stage, Central and South Florida were combined into 1 region based on the ARR of cesarean delivery in each of the 2 locations. Geographic location was the only statistically significant hospital risk factor after adjustment for all significant individual level risk factors in the multivariable model. Compared with North Florida hospitals, the adjusted risk of cesarean delivery was 16% higher in Central Florida (ARR, 1.16; 95% CI, 1.04–1.28), 18% higher in South Florida (ARR, 1.18; 95% CI, 1.05–1.32), and 73% higher in Miami-Dade County (ARR, 1.73; 95% CI, 1.52–1.96) ( Table 2 ). In a model with interaction terms, hospital geographic location was a statistically significant modifier of the risk of cesarean delivery for maternal prepregnancy obesity, medical risk conditions, and labor induction ( P < .05).
| Variable | North Florida (29 facilities) | Central and South Florida (80 facilities) | Miami-Dade County (13 facilities) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Cesarean delivery, % | Adjusted relative risk (95% confidence interval) a | Change in between-hospital variance (σ 2 ), % | Cesarean delivery, % | Adjusted relative risk (95% confidence interval) a | Change in between-hospital variance (σ 2 ), % | Cesarean delivery, % | Adjusted relative risk (95% confidence interval) a | Change in hospital variance (σ 2 ), % | |
| Age, y | 8.5 | 5.1 | –0.3 b | ||||||
| <20 | 14.0 | 1.00 | 16.5 | 1.00 | 24.6 | 1.00 | |||
| 20-29 | 19.9 | 1.34 (1.28–1.39) c | 22.7 | 1.32 (1.29–1.35) c | 32.2 | 1.27 (1.22–1.33) c | |||
| 30-34 | 25.0 | 1.77 (1.67–1.87) c | 28.8 | 1.73 (1.68–1.78) c | 37.6 | 1.55 (1.47–1.63) c | |||
| ≥35 | 32.6 | 2.25 (2.11–2.39) c | 38.1 | 2.26 (2.20–2.32) c | 47.0 | 1.91 (1.81–2.01) c | |||
| Race and ethnicity | 4.8 | –1.6 b | –15.7 b | ||||||
| Non-Hispanic white | 19.0 | 1.00 | 22.7 | 1.00 | 28.8 | 1.00 | |||
| Non-Hispanic black or Haitian | 21.6 | 1.25 (1.21–1.30) c | 25.6 | 1.21 (1.19–1.24) c | 29.1 | 1.16 (1.09–1.23) d | |||
| Hispanic | |||||||||
| Mexican | 16.0 | 0.98 (0.88–1.10) | 19.0 | 1.01 (0.97–1.05) | 28.3 | 1.06 (0.95–1.19) | |||
| Puerto Rican | 20.1 | 1.16 (1.04–1.29) d | 24.1 | 1.20 (1.17–1.24) c | 32.9 | 1.23 (1.13–1.34) d | |||
| Cuban | 23.4 | 1.26 (1.06–1.49) e | 28.6 | 1.23 (1.18–1.28) c | 40.2 | 1.38 (1.32–1.45) c | |||
| Other | |||||||||
| Hispanic | 21.2 | 1.22 (1.10–1.34) d | 24.1 | 1.13 (1.10–1.16) c | 31.8 | 1.16 (1.10–1.22) d | |||
| Non-Hispanic | 20.5 | 1.20 (1.13–1.28) c | 24.3 | 1.19 (1.15–1.23) c | 29.8 | 1.10 (0.98–1.23) d | |||
| Married | 21.0 | 0.99 (0.95–1.02) | 0.0 | 24.9 | 0.94 (0.92–0.96) c | –0.4 b | 34.1 | 0.95 (0.92–0.99) d | 0.1 |
| Insurance status | 0.5 | 0.2 | –0.2 b | ||||||
| Medicaid | 18.8 | 1.25 (1.12–1.39) d | 22.0 | 1.16 (1.11–1.21) c | 33.2 | 1.15 (1.06–1.25) d | |||
| Private/other | 20.8 | 1.21 (1.09–1.35) d | 25.3 | 1.17 (1.12–1.23) c | 33.7 | 1.08 (0.99–1.18) | |||
| Self-pay | 15.9 | 1.00 | 18.6 | 1.00 | 27.2 | 1.00 | |||
| Body mass index, kg/m 2 | –2.4 b | 1.1 | 0.0 | ||||||
| Underweight (<18.5) | 15.2 | 0.79 (0.73–0.85) c | 19.5 | 0.77 (0.75–0.80) c | 29.6 | 0.82 (0.77–0.88) c | |||
| Normal (18.5-24.9) | 22.6 | 1.00 | 27.0 | 1.00 | 38.4 | 1.00 | |||
| Overweight (25-29.9) | 33.5 | 1.33 (1.29–1.37) c | 38.3 | 1.25 (1.24–1.27) c | 48.3 | 1.19 (1.15–1.22) c | |||
| Obese (≥30) | 10.6 | 1.75 (1.70–1.80) c | 13.2 | 1.62 (1.60–1.65) c | 21.7 | 1.43 (1.39–1.47) c | |||
| Gestational age, wk | 0.1 | 4.7 | 0.2 | ||||||
| 37 | 18.1 | 1.00 | 21.0 | 1.00 | 30.6 | 1.00 | |||
| 38 | 18.1 | 1.09 (1.03–1.16) d | 21.6 | 1.09 (1.06–1.13) c | 31.2 | 1.06 (0.99–1.12) | |||
| 39 | 18.9 | 1.16 (1.10–1.23) c | 22.7 | 1.21 (1.18–1.25) c | 31.7 | 1.11 (1.05–1.17) d | |||
| 40 | 22.3 | 1.39 (1.32–1.46) c | 26.2 | 1.41 (1.37–1.45) c | 36.9 | 1.29 (1.22–1.36) c | |||
| ≥1 Medical risk conditions f | 31.3 | 1.83 (1.78–1.88) c | 6.9 | 36.3 | 1.68 (1.65–1.70) c | –1.7 b | 51.5 | 1.76 (1.72–1.81) c | –1.6 b |
| Induced labor | 26.3 | 1.47 (1.42–1.51) c | –6.5 b | 32.1 | 1.59 (1.57–1.61) c | –0.2 b | 41.9 | 1.31 (1.27–1.35) c | 10.1 |
| Delivery on Monday-Friday | 20.2 | 1.00 (0.96–1.03) | –1.3 b | 24.1 | 1.00 (0.98–1.02) | –1.0 b | 34.4 | 1.10 (1.06–1.14) d | 4.3 |
| Delivery time of day | |||||||||
| 12:00-03:59 am | 17.5 | 1.03 (0.97–1.11) | 20.3 | 1.05 (1.01–1.09) e | 24.9 | 1.07 (0.98–1.16) | |||
| 04:00-07:59 am | 16.1 | 1.00 | 18.6 | 1.00 | 22.8 | 1.00 | |||
| 08:00-11:59 am | 18.0 | 1.14 (1.07–1.21) d | 21.7 | 1.16 (1.12–1.19) c | 18.5 | 1.56 (1.47–1.66) c | |||
| 12:00-03:59 pm | 16.7 | 0.98 (0.92–1.04) | 20.8 | 1.04 (1.01–1.08) e | 32.6 | 1.36 (1.27–1.44) c | |||
| 04:00-07:59 pm | 22.3 | 1.25 (1.18–1.32) c | 26.7 | 1.29 (1.25–1.32) c | 36.9 | 1.52 (1.43–1.61) c | |||
| 08:00-11:59 pm | 24.7 | 1.39 (1.31–1.47) c | 28.7 | 1.41 (1.37–1.45) c | 36.2 | 1.52 (1.43–1.62) c | |||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
