This cohort study examines whether the probabilities of postpartum contraception and subsequent deliveries vary according to whether the closest hospital is a Catholic or non-Catholic hospital.
Key Points
Question
Do rates of postpartum contraception and subsequent deliveries differ when the closest hospital is Catholic vs non-Catholic?
Findings
This cohort study included 4 101 443 deliveries in 11 states. Living in a zip code in which the closest hospital was Catholic was associated with a 0.95–percentage point decrease in the probability of surgical sterilization at delivery, a 0.21–percentage point decrease in the probability of surgical sterilization in the year after discharge, and a 0.47–percentage point increase in the probability of subsequent delivery within 3 years.
Meaning
This study finds that living in a zip code in which the closest hospital was Catholic was associated with a modest decrease in the probability of postpartum surgical sterilizations and a modest increase in the probability of subsequent deliveries.
Abstract
Importance
In recent years, the number of Catholic hospitals has grown, raising concerns about access to contraception. The association between living in an area in which the closest hospital is Catholic and the probability of postpartum contraception and subsequent deliveries is unknown.
Objective
To assess whether living in an area in which the closest hospital was Catholic was associated with the probability of postpartum contraception and subsequent deliveries.
Design, Setting, and Participants
This cohort study used data from the Healthcare Cost and Utilization Project’s State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin). Female patients with a delivery from 2016 to 2019 who lived within 20 miles of a nonfederal acute care hospital were included, with patients followed up for 1 to 3 years. Coarsened exact matching was used to match patients based on the county-level percentage of the population affiliated with Catholic churches and urbanicity, and the zip code–level number of hospitals within 5 and 20 miles, median household income, and percentage of the population by race and ethnicity. Data were analyzed from April 2022 to November 2023.
Exposures
Residence in a zip code in which the closest hospital was Catholic.
Main Outcomes and Measures
Probabilities of delivery at a Catholic hospital, surgical sterilization within 1 year of delivery, receipt of long-acting reversible contraception at delivery, and subsequent delivery within 3 years.
Results
The sample consisted of 4 101 443 deliveries (1 301 792 after matching), with 14.5% of patients living in exposed zip codes (ie, where the closest hospital was Catholic). Living in exposed zip codes was associated with a 21.26–percentage point (pp) increase in the probability of delivery at a Catholic hospital (95% CI, 19.50 to 23.02 pp; 237.3% relative to the mean in unexposed zip codes; P < .001). Additionally, comparing exposed vs unexposed zip codes, the probability of surgical sterilization at delivery decreased by 0.95 pp (95% CI, −1.14 to −0.76 pp; P < .001) and the probability of sterilization in the year after discharge further decreased by 0.21 pp (95% CI, −0.29 to −0.13; P < .001). Subsequent deliveries within 3 years increased 0.47 pp (95% CI, −0.03 to 0.97 pp; 2.3% relative to the mean in unexposed zip codes; P = .07).
Conclusions and Relevance
This cohort study finds that living in a zip code in which the closest hospital was Catholic was associated with a modest decrease in the probability of postpartum surgical sterilizations and a modest increase in the probability of subsequent deliveries.
Introduction
From 2001 to 2016, the number of Catholic hospitals in the US grew by 22%, accounting for one-sixth of acute care beds in 2016.1 Previous studies have raised concerns that the growing number of Catholic hospitals in the US may reduce access to postpartum contraception.2,3,4,5 The Ethical and Religious Directives for Catholic Health Care Services prohibits contraceptive services except to treat serious conditions that have no alternative treatment, and the Church Amendment (enacted by Congress in 1973) allows Catholic hospitals to refuse services that conflict with their religious beliefs or moral convictions.6 Despite these restrictions, many women do not know whether their local hospitals are Catholic or whether contraceptive services are offered at Catholic hospitals.7,8,9,10,11
Few studies have examined the association between delivery at Catholic hospitals and postpartum contraception. Prior research found that rates of postpartum tubal ligation decreased in hospitals that changed from non-Catholic to Catholic ownership,12 and that people who delivered in Catholic hospitals were less likely to have received a tubal ligation 2 to 6 months after delivery.13 To our knowledge, no previous peer-reviewed studies have examined how the local availability of Catholic hospitals affects postpartum contraception, or explored the consequences of delivery at Catholic hospitals with longitudinal data for patients more than 6 months post partum.
This study used longitudinal patient data from in-hospital deliveries in 11 states to examine whether the Catholic affiliation of a patient’s nearest hospital was associated with hospital choice for delivery, receipt of contraception (at delivery and 1 year after delivery), and subsequent deliveries (within 3 years). We explored heterogeneity in these measures according to the degree of exposure to Catholic hospitals (based on the distance to the nearest non-Catholic hospital) and by patient age, race and ethnicity, expected payer, and county urbanicity.
Methods
Data and Sample
Data on in-hospital deliveries were obtained from the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases, State Ambulatory Surgery and Services Databases, and State Emergency Department Databases from 2016 to 2019. Data from 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin) were chosen because these states participated in the HCUP databases and reported consistent encrypted patient identifiers for those years.14,15,16 For 9 states (excluding Georgia and Nebraska), it was also possible to analyze 2020 data. These data covered all discharges from inpatient, emergency department, and hospital-affiliated ambulatory surgery departments for nonfederal acute care hospitals in the included states. The Agency for Healthcare Research and Quality’s human protection administrator deemed this cohort study non–human participant research; thus the study was exempt from review and the informed consent requirement according to the Common Rule. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The HCUP data were supplemented with hospital characteristics from the American Hospital Association Annual Survey Database, zip code characteristics from the 2017 US Census Bureau American Community Survey, and county characteristics from the 2010 and 2020 US Religion Census Religious Congregations and Membership Study and the 2022 US Health Resources & Services Administration Area Health Resources File.
The study included deliveries by female patients aged 21 to 55 years. Patients younger than 21 years were excluded because they cannot consent to surgical sterilization in most states. Patients who lived over 20 miles from the nearest hospital were also excluded (3.2% of deliveries). To follow-up patients for at least 1 year after delivery, the study examined deliveries from 2016 to 2019 in 9 states, and from 2016 to 2018 in 2 states.
Study Outcomes
For each delivery, we examined hospital choice, receipt of postpartum contraception, and subsequent deliveries. Indicators of hospital choice included delivery at the closest hospital to the patient, which was used to test whether patients avoided the nearest hospital if it was Catholic. Delivery at a Catholic hospital was used to ascertain whether patients who lived closer to a Catholic hospital were more likely to deliver at a Catholic hospital. Contraception outcomes included postpartum surgical sterilization (including tubal ligation) during the delivery admission and after discharge within 1 year of admission for the index delivery (eTable 1 in Supplement 1 lists the medical codes used to identify outcomes). We also examined receipt of postpartum long-acting reversible contraception (LARC) during the delivery admission (LARC insertion after delivery was not examined, as it is primarily performed outside of hospitals). The final outcome was subsequent delivery within 3 years of the reference delivery. Because of the 3-year follow-up, this variable was defined only for deliveries in 2016 or 2017.
Exposure to Catholic Hospitals
Data on the religious affiliation, latitude, and longitude of hospitals were obtained from the American Hospital Association. For latitude and longitude of patient residence, we used the population-weighted centroid of their zip code at the time of delivery. For each delivery, we calculated the patient’s distance to the nearest Catholic and non-Catholic hospital, considering all hospitals with deliveries in that year (this distance measure did not consider roads or obstacles). For each delivery, the patient was considered exposed if the closest hospital to their zip code was Catholic.
Statistical Analysis
The associations of interest were estimated using linear probability models. We regressed each outcome on an indicator for whether the patient’s closest hospital was Catholic and controlled for patient, delivery, and geographic characteristics. For each outcome, we reported the adjusted difference in probability for patients whose closest hospital was Catholic relative to patients whose closest hospital was non-Catholic in percentage points. Patient characteristics included age, insurance type (private insurance or Medicaid), and comorbidities (asthma, chronic obstructive pulmonary disease, morbid obesity, hypertension, and diabetes). Delivery characteristics included gestational age, cesarean delivery, and complications (postpartum hemorrhage, poor fetal growth, chorioamnionitis, preeclampsia, and any severe maternal morbidity). Comorbidities and delivery characteristics were identified using International Statistical Classification of Diseases, Tenth Revision, Clinical Modification diagnosis codes. Zip code characteristics included the number of hospitals within 5 and 20 miles, delivery volumes for the nearest Catholic and non-Catholic hospitals, and additional characteristics from the American Community Survey. These characteristics were median household income, the proportion of the population by race and ethnicity (Hispanic, non-Hispanic Black [hereafter, Black], or non-Hispanic White [hereafter, White]), and the proportion of the population with an income below the federal poverty level. County characteristics included the proportion of county residents affiliated with Catholic churches (interpolated from 2010 and 2020 data), the number of primary care physicians per capita, and urban-rural continuum codes. We also controlled for state-by-year fixed effects. SEs were clustered at the zip code level. Statistical significance was set at P < .05 and calculated using 2-tailed tests.
Unobserved variables may bias linear regression models if they are correlated with the exposure and outcome variables. For example, demand for postpartum contraception could be lower among people who live closer to Catholic hospitals if they prefer to have more children. Preferences and other such variables are not observable, so we used coarsened exact matching to limit comparisons to zip codes with similar characteristics. Matching on more characteristics increased the similarity of comparison zip codes, but also resulted in more unmatched zip codes. Thus, we estimated a basic match that included only a few key characteristics, and also a demographic match that included a more comprehensive set of characteristics and was our preferred specification.
The basic match grouped patients by year of delivery and quartiles of the following variables: the number of hospitals within 5 and 20 miles and the proportion of county residents affiliated with the Catholic church. We excluded groups that did not have at least 1 exposed and 1 unexposed patient, and weighted the number of unexposed patients so that it summed to the number of exposed patients in each group.17 We then estimated linear regressions on the matched data using the calculated weights and controlling for group fixed effects and the previously listed covariates. In addition to the variables from the basic match, the demographic match grouped patients based on whether the patient resided in a metropolitan county and quartiles for the following zip code characteristics: percentage of Black, Hispanic, and White populations, and median household income. Covariate balance after matching was assessed using the absolute standardized mean difference (ASMD), with values less than 0.1 indicating an acceptable balance.18
Heterogeneity was assessed by degree of exposure to Catholic hospitals and patient subsamples. For degree of exposure, zip codes in which the closest hospital was Catholic were further categorized by the distance to the closest non-Catholic hospital: less than 5 miles, 5 to 20 miles, or more than 20 miles. In patient subsamples, we stratified by age, race and ethnicity, and expected payer. Analyses by patient race examined Black, Hispanic, and White patients, with patients of other race or ethnicity excluded due to smaller sample sizes. The HCUP databases rely on patient race and ethnicity as recorded on the discharge record by the hospital, which is ideally based on patient self-report but may be collected by hospital staff based on observation.19,20 For the heterogeneity analyses, we ran separate regressions for each subsample and reported results using the demographic match. The results using other methods were similar.
Data analyses were performed using Stata/MP, version 17 (StataCorp LLC) from April 2022 to November 2023. Findings should be interpreted as exploratory due to the potential for type I error and multiple comparisons.
Results
Descriptive Statistics
The full sample consisted of 4 101 443 deliveries, of which 3 860 214 were included in the basic match and 1 301 792 were included in the demographic match. Patients who lived in exposed zip codes accounted for 596 123 (14.5%) deliveries. In the full sample, there were significant imbalances between exposed patients (ie, those who lived closer to a Catholic hospital based on zip code) and unexposed patients (Table 1). The ASMD was greater than 0.1 for most zip code– and county-level covariates, including the proportion of the population affiliated with Catholic churches, the number of nearby hospitals, and other demographic covariates. After applying the basic match, differences between the exposed and unexposed groups were reduced among many covariates, and after applying the demographic match, the ASMD was below 0.1 for all covariates, indicating balance between the exposed and unexposed groups. Tests for statistically significant differences between means also showed reductions in differences between exposed and unexposed groups after matching (eTable 2 in Supplement 1).
Table 1. Descriptive Statistics for Deliveries by Patient Proximity to Non-Catholic and Catholic Hospitalsa.
| Characteristic | Mean (SD) % | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Unmatched sample | ASMD | Basic match | ASMD | Demographic match | ASMD | ||||
| Non-Catholic hospital closer (n = 3 505 311) | Catholic hospital closer (n = 596 132) | Non-Catholic hospital closer (n = 3 264 085) | Catholic hospital closer (n = 596 129) | Non-Catholic hospital closer (n = 907 842) | Catholic hospital closer (n = 393 950) | ||||
| Individual level | |||||||||
| Medicaid | 41.6 (49.3) | 38.5 (48.7) | 0.063 | 40.4 (49.1) | 38.5 (48.7) | 0.040 | 35.9 (48.0) | 36.3 (48.1) | 0.008 |
| Private insurance | 53.6 (49.9) | 57.9 (49.4) | 0.087 | 54.7 (49.8) | 57.9 (49.4) | 0.065 | 59.6 (49.1) | 60.1 (49.0) | 0.011 |
| Gestational age, wk | 38.4 (2.2) | 38.4 (2.2) | 0.013 | 38.4 (2.2) | 38.4 (2.2) | 0.007 | 38.3 (2.2) | 38.4 (2.2) | 0.012 |
| Maternal age, y | 30.0 (5.1) | 30.0 (5.0) | 0.000 | 30.1 (5.1) | 30.0 (5.0) | 0.010 | 30.0 (5.0) | 30.1 (5.0) | 0.012 |
| Cesarean delivery | 34.1 (47.4) | 32.2 (46.7) | 0.039 | 34.3 (47.5) | 32.2 (46.7) | 0.043 | 32.9 (47.0) | 32.5 (46.9) | 0.008 |
| Asthma or COPD | 5.1 (21.9) | 5.9 (23.6) | 0.038 | 5.3 (22.3) | 5.9 (23.6) | 0.029 | 5.7 (23.2) | 5.8 (23.4) | 0.003 |
| Postpartum hemorrhage | 3.8 (19.1) | 3.7 (18.9) | 0.005 | 3.8 (19.2) | 3.7 (18.9) | 0.007 | 3.7 (18.8) | 3.6 (18.6) | 0.004 |
| Poor fetal growth | 3.3 (17.8) | 3.3 (17.9) | 0.004 | 3.2 (17.5) | 3.3 (17.9) | 0.008 | 3.2 (17.7) | 3.3 (17.8) | 0.003 |
| Morbid obesity | 3.2 (17.7) | 3.3 (17.9) | 0.006 | 3.2 (17.5) | 3.3 (17.9) | 0.008 | 3.1 (17.2) | 3.2 (17.7) | 0.010 |
| Chorioamnionitis | 2.5 (15.6) | 2.1 (14.3) | 0.028 | 2.4 (15.4) | 2.1 (14.3) | 0.022 | 2.2 (14.8) | 2.1 (14.2) | 0.012 |
| Preeclampsia | 5.0 (21.7) | 4.7 (21.1) | 0.015 | 4.8 (21.4) | 4.7 (21.1) | 0.008 | 4.9 (21.5) | 4.6 (21.0) | 0.011 |
| Hypertension | 2.6 (15.9) | 2.6 (15.9) | 0.001 | 2.5 (15.7) | 2.6 (15.9) | 0.004 | 2.5 (15.6) | 2.6 (15.8) | 0.006 |
| Diabetes | 6.2 (24.2) | 6.0 (23.7) | 0.009 | 6.2 (24.1) | 6.0 (23.7) | 0.007 | 5.6 (23.0) | 5.9 (23.6) | 0.012 |
| Severe maternal morbidityb | 0.8 (8.9) | 0.8 (8.7) | 0.005 | 0.8 (8.9) | 0.8 (8.7) | 0.003 | 0.7 (8.6) | 0.7 (8.5) | 0.000 |
| Zip code and county level | |||||||||
| Black | 15.1 (19.3) | 12.0 (17.9) | 0.165 | 13.0 (17.4) | 12.0 (17.9) | 0.052 | 11.6 (19.1) | 11.5 (18.4) | 0.002 |
| Hispanic | 25.7 (23.7) | 18.8 (20.3) | 0.313 | 26.9 (23.9) | 18.8 (20.3) | 0.366 | 17.9 (20.7) | 18.0 (20.9) | 0.002 |
| White | 64.1 (23.5) | 72.1 (21.0) | 0.362 | 67.1 (21.7) | 72.1 (21.0) | 0.237 | 73.7 (22.2) | 73.5 (21.7) | 0.008 |
| 100% FPL | 14.6 (8.4) | 13.5 (8.1) | 0.126 | 14.0 (8.5) | 13.5 (8.1) | 0.057 | 12.9 (9.3) | 12.8 (8.0) | 0.007 |
| Unemployed | 5.9 (2.8) | 5.3 (2.7) | 0.201 | 5.8 (2.9) | 5.3 (2.7) | 0.181 | 5.5 (3.3) | 5.2 (2.7) | 0.096 |
| Median household income, $ | 67 770 (28 036) | 68 094 (24 131) | 0.012 | 69 166 (28 138) | 68 094 (24 131) | 0.041 | 70 257 (28 455) | 69 675 (25 638) | 0.021 |
| PCP density per 1000 population | 76.7 (27.8) | 78.6 (25.7) | 0.074 | 77.9 (26.2) | 78.6 (25.7) | 0.030 | 77.4 (28.3) | 77.1 (27.1) | 0.011 |
| Metropolitan area | 92.8 (25.9) | 95.0 (21.9) | 0.091 | 94.2 (23.3) | 95.0 (21.9) | 0.032 | 93.7 (24.3) | 93.7 (24.3) | 0.000 |
| Affiliated with Catholic churches | 20.8 (11.2) | 24.5 (11.6) | 0.328 | 24.0 (11.7) | 24.5 (11.6) | 0.045 | 23.4 (11.9) | 23.8 (12.0) | 0.040 |
| No. of hospitals within 5 miles | 2.0 (2.9) | 1.3 (1.4) | 0.293 | 1.4 (1.8) | 1.3 (1.4) | 0.066 | 1.4 (1.7) | 1.4 (1.6) | 0.037 |
| No. of hospitals within 20 miles | 13.1 (33.7) | 9.4 (29.1) | 0.119 | 9.7 (29.6) | 9.4 (29.1) | 0.011 | 10.3 (30.4) | 10.0 (30.0) | 0.012 |
Abbreviations: ASMD, absolute standardized mean difference; COPD, chronic obstructive pulmonary disease; FPL, federal poverty level; PCP, primary care physician.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases, 2016 to 2020 for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin).
Severe maternal morbidity indicates that 1 or more of the following conditions is associated with the patient's discharge record: acute myocardial infarction, aneurysm, acute kidney failure, adult respiratory distress syndrome, amniotic fluid embolism, cardiac arrest or ventricular fibrillation, conversion of cardiac rhythm, disseminated intravascular coagulation, eclampsia, heart failure or cardiac arrest during surgery or procedure, puerperal cerebrovascular disorder, acute heart failure or pulmonary edema, severe anesthesia complication, sepsis, shock, sickle cell disease with crisis, air and thrombotic embolism, hysterectomy, temporary tracheostomy, and the need for ventilation.
Association of Outcomes With Closest Hospital Being Catholic
The direction and statistical significance of estimates were similar in the unmatched analysis, the basic match, and the demographic match (Table 2). After applying the demographic match and adjusting for covariates, living in an exposed zip code (ie, where the closest hospital was Catholic) was not associated with a statistically significant difference in the probability of delivery at the closest hospital (ie, 34% of patients delivered at the closest hospital whether or not it was Catholic). However, exposed patients were 21.26 percentage points (pp) (95% CI, 19.50 to 23.02 pp) more likely to deliver at a Catholic hospital, a 237.3% increase relative to the mean for unexposed patients (8.96 pp). The probability of postpartum surgical sterilization for exposed patients vs unexposed patients decreased by 0.95 pp (95% CI, −1.14 to −0.76 pp; −13.8%) at delivery and by 0.21 pp (95% CI, −0.29 to −0.13 pp; −13.1%) after discharge and within 1 year of delivery. We did not find an association with LARC insertion. In the basic match, we found evidence of an increase in the probability of subsequent delivery within 3 years: the 0.47-pp increase (95% CI, −0.03 to 0.97 pp; 2.3%) was significant at the 10% level in the demographic match (P = .07) and at the 5% level (P = .02) in the basic match.
Table 2. Association of Outcomes With Closest Hospital Being Catholica.
| Outcome | Mean for zip codes in which closest hospital is not Catholic, percentage points (SE) | Adjusted difference for zip codes in which closest hospital was Catholic, percentage points (95% CI)b,c | |||||
|---|---|---|---|---|---|---|---|
| Unmatched | P value | Basic match | P value | Demographic match | P value | ||
| Delivery at closest hospital | 34.22 (0.48) | −0.18 (−2.48 to 2.12) | .88 | 0.24 (−1.72 to 2.21) | .81 | 0.75 (−1.09 to 2.59) | .42 |
| Delivery at Catholic hospital | 8.96 (0.28) | 25.47 (23.42 to 27.51) | <.001 | 26.15 (24.12 to 28.18) | <.001 | 21.26 (19.50 to 23.02) | <.001 |
| Surgical sterilization at delivery | 6.88 (0.06) | −1.06 (−1.26 to −0.86) | <.001 | −1.01 (−1.20 to −0.82) | <.001 | −0.95 (−1.14 to −0.76) | <.001 |
| Surgical sterilization after discharge and within 1 y of delivery | 1.60 (0.03) | −0.28 (−0.37 to −0.19) | <.001 | −0.30 (−0.37 to −0.22) | <.001 | −0.21 (−0.29 to −0.13) | <.001 |
| LARC insertion at delivery | 0.93 (0.04) | 0.04 (−0.10 to 0.17) | .57 | −0.03 (−0.13 to 0.08) | .66 | −0.08 (−0.20 to 0.04) | .20 |
| Subsequent delivery within 3 y | 20.16 (0.13) | 0.41 (−0.12 to 0.95) | .13 | 0.54 (0.07 to 1.01) | .02 | 0.47 (−0.03 to 0.97) | .07 |
Abbreviation: LARC, long-acting reversible contraception.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases, 2016 to 2020 for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin).
All regressions other than subsequent deliveries estimated using deliveries that occurred during 2016-2019. The number of deliveries was 4 051 848 for unmatched analyses, 3 813 011 for basic match analyses, and 1 287 358 for demographic match analyses.
Subsequent deliveries estimated using deliveries that occurred during 2016-2017. The number of deliveries was 1 997 864 for unmatched analyses, 1 843 239 for basic match analyses, and 654 715 for demographic match analyses.
The magnitude of these associations increased with the degree of exposure to Catholic hospitals (Table 3). Relative to patients whose closest hospital was non-Catholic, the probability of delivery at a Catholic hospital increased most for exposed patients whose nearest non-Catholic hospital was more than 20 miles away (57.85 pp [95% CI, 53.94 to 61.76 pp]; 657.4%) (subsample means reported in eTable 3 in Supplement 1). We observed smaller increases for exposed patients whose nearest non-Catholic hospital was 5 to 20 miles away (19.34 pp [95% CI, 17.36 to 21.32 pp]; 160.1%) and for exposed patients whose nearest non-Catholic hospital was less than 5 miles away (10.26 pp [95% CI, 8.47 to 12.04 pp]; 64.3%). Similarly, the probability of surgical sterilization at delivery decreased for all exposed patients, with the largest decrease for patients whose nearest non-Catholic hospital was more than 20 miles away (−2.95 pp [95% CI, −3.62 to −2.29 pp]; −32.4%). Decreases in the probability of surgical sterilization after discharge and within 1 year of delivery were statistically significant for patients whose nearest non-Catholic hospital was 5 to 20 miles (−0.16 pp [95% CI, −0.27 to −0.05 pp; −7.5%) and more than 20 miles away (−0.45 pp [95% CI, −0.85 to −0.06 pp; −13.6%). The probability of subsequent delivery within 3 years increased by 2.75 pp (95% CI, 1.24 to 4.27 pp; 12.4%) only for patients whose nearest non-Catholic hospital was more than 20 miles away.
Table 3. Association of Outcomes With Degree of Exposure to Catholic Hospitalsa.
| Outcome | Adjusted difference, percentage points (95% CI)b,c,d | |||||
|---|---|---|---|---|---|---|
| Catholic hospital closer and nearest non-Catholic hospital <5 miles away | P value | Catholic hospital closer and nearest non-Catholic hospital 5-20 miles away | P value | Catholic hospital closer and nearest non-Catholic hospital >20 miles away | P value | |
| Delivery at closest hospital | 0.62 (−2.02 to 3.27) | .64 | 0.15 (−2.73 to 3.03) | .92 | 3.12 (−2.41 to 8.65) | .27 |
| Delivery at Catholic hospital | 10.26 (8.47 to 12.04) | <.001 | 19.34 (17.36 to 21.32) | <.001 | 57.85 (53.94 to 61.76) | <.001 |
| Surgical sterilization at delivery | −0.35 (−0.61 to −0.10) | .006 | −0.93 (−1.20 to −0.66) | <.001 | −2.95 (−3.62 to −2.29) | <.001 |
| Surgical sterilization after discharge and within 1 y of delivery | −0.03 (−0.12 to 0.07) | .60 | −0.16 (−0.27 to −0.05) | .003 | −0.45 (−0.85 to −0.06) | .03 |
| LARC insertion at delivery | 0.08 (−0.08 to 0.23) | .33 | −0.15 (−0.33 to 0.04) | .12 | −0.15 (−0.31 to 0.02) | .08 |
| Subsequent delivery within 3 y | 0.44 (−0.30 to 1.19) | .25 | 0.19 (−0.47 to 0.85) | .58 | 2.75 (1.24 to 4.27) | <.001 |
Abbreviation: LARC, long-acting reversible contraception.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases, 2016 to 2020 for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin).
Adjusted difference for patients who live in zip codes in which the closest hospital was Catholic, relative to zip codes in which the closest hospital was not Catholic, calculated using the demographic match method.
All regressions other than subsequent deliveries were estimated using deliveries that occurred during 2016-2019. The number of deliveries was 435 365, 757 592, and 250 850 when exposed zip codes were limited to less than 5, 5 to 20, and more than 20 miles away from the nearest non-Catholic hospital, respectively.
Subsequent deliveries estimated using deliveries that occurred during 2016-2017. The number of deliveries was 242 759, 369 962, and 117 776 when exposed zip codes were limited to <5, 5-20, and >20 miles away from the nearest non-Catholic hospital, respectively.
Findings varied by patient age (Table 4). Patients aged 21 to 25 years were more likely (2.31 pp [95% CI, 0.08 to 4.55 pp]; 5.7%) to deliver at their closest hospital if it was Catholic; however, among older patients, we found no significant difference in this probability. Living in an exposed zip code was associated with an increase of 19.46 pp (95% CI, 17.82 to 21.11 pp; 153.7%) in the probability of delivering at a Catholic hospital among patients aged 35 years or older, with larger increases among younger patients. The probability of surgical sterilization at delivery decreased for patients of all ages and was largest in absolute terms among patients aged 35 years or older (−1.67 pp [95% CI, −2.09 to −1.26 pp]; −14.1%), followed by patients aged 26 to 34 years (−1.08 pp [95% CI, −1.28 to −0.89 pp]; −16.3%). Declines in surgical sterilization after discharge were smaller and similar in magnitude for all patients. The probability of subsequent delivery within 3 years increased for all patients but was statistically significant only for patients aged 26 to 34 years (0.64 pp [95% CI, 0.04 to 1.23 pp]; 2.8%).
Table 4. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Patient Agea.
| Outcome | Adjusted difference, percentage points (95% CI)b,c,d | |||||
|---|---|---|---|---|---|---|
| Age 21-25 y | P value | Age 26-34 y | P value | Age ≥35 y | P value | |
| Delivery at closest hospital | 2.31 (0.08 to 4.55) | .04 | −0.14 (−2.01 to 1.74) | .89 | −0.83 (−2.81 to 1.15) | .41 |
| Delivery at Catholic hospital | 22.96 (20.58 to 25.35) | <.001 | 21.39 (19.64 to 23.15) | <.001 | 19.46 (17.82 to 21.11) | <.001 |
| Surgical sterilization at delivery | −0.41 (−0.64 to −0.18) | <.001 | −1.08 (−1.28 to −0.89) | <.001 | −1.67 (−2.09 to −1.26) | <.001 |
| Surgical sterilization after discharge and within 1 y of delivery | −0.27 (−0.42 to −0.11) | .001 | −0.24 (−0.34 to −0.13) | <.001 | −0.21 (−0.38 to −0.04) | .01 |
| LARC insertion at delivery | −0.16 (−0.38 to 0.07) | .17 | −0.11 (−0.22 to 0.00) | .06 | −0.08 (−0.20 to 0.04) | .20 |
| Subsequent delivery within 3 y | 0.27 (−0.6 to 1.13) | .55 | 0.64 (0.04 to 1.23) | .04 | 0.47 (−0.1 to 1.05) | .11 |
Abbreviation: LARC, long-acting reversible contraception.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases, 2016 to 2020 for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin).
Adjusted difference for patients who live in zip codes in which the closest hospital was Catholic, relative to zip codes in which the closest hospital was non-Catholic, calculated using the demographic match method.
All regressions other than subsequent deliveries were estimated using deliveries that occurred during 2016-2019. The number of deliveries was 274 746, 732 933, and 254 817 for patients aged 21 to 25 years, 26 to 34 years, and 35 years or older, respectively
Subsequent deliveries estimated using deliveries that occurred during 2016-2017. The number of deliveries was 144 087, 372 005, and 127 796 for patients aged 21 to 25 years, 26 to 34 years, and 35 years or older, respectively
The results also varied by race and ethnicity (Table 5). Only Hispanic patients were more likely to deliver at the closest hospital if it was Catholic (4.89 pp [95% CI, 1.97 to 7.8]); 7.5%). Living in an exposed zip code was associated with an increase in the probability of delivery at a Catholic hospital by 19.21 pp (95% CI, 15.73 to 22.69 pp; 132.4%) for Hispanic patients, 12.94 pp (95% CI, 10.64 to 15.24 pp; 86.2%) for Black patients, and 23.88 pp (95% CI, 21.87 to 25.89 pp; 194.1%) for White patients. When the closest hospital to patients was Catholic, the probability of surgical sterilization at delivery decreased among patients of all the races and ethnicities studied (Table 5). In contrast, the probability of LARC insertion at delivery decreased only among Black patients, and the probability of surgical sterilization after discharge and within 1 year declined only among White patients. The probability of a subsequent delivery within 3 years increased by 1.27 pp (95% CI, 0.49 to 2.05 pp; 6.6%) for Hispanic patients (P = .001) and by 1.03 pp (95% CI, −0.02 to 2.08 pp; 4.6%) for Black patients (P = .05).
Table 5. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Patient Race and Ethnicitya.
| Outcome | Adjusted difference in percentage points (95% CI)b,c,d | |||||
|---|---|---|---|---|---|---|
| Black | P value | Hispanic | P value | White | P value | |
| Delivery at closest hospital | −1.03 (−4.29 to 2.22) | .53 | 4.89 (1.97 to 7.80) | .001 | 0.03 (−2.20 to 2.25) | .98 |
| Delivery at Catholic hospital | 12.94 (10.64 to 15.24) | <.001 | 19.21 (15.73 to 22.69) | <.001 | 23.88 (21.87 to 25.89) | <.001 |
| Surgical sterilization at delivery | −0.90 (−1.36 to −0.45) | <.001 | −0.80 (−1.15 to −0.44) | <.001 | −1.16 (−1.39 to −0.93) | <.001 |
| Surgical sterilization after discharge and within 1 y of delivery | 0.06 (−0.18 to 0.29) | .63 | −0.14 (−0.31 to 0.03) | .11 | −0.30 (−0.42 to −0.19) | <.001 |
| LARC insertion at delivery | −0.58 (−1.02 to −0.14) | .01 | 0.02 (−0.14 to 0.17) | .83 | 0.02 (−0.13 to 0.17) | .79 |
| Subsequent delivery within 3 y | 1.03 (−0.02 to 2.08) | .05 | 1.27 (0.49 to 2.05) | .001 | 0.10 (−0.51 to 0.71) | .75 |
Abbreviation: LARC, long-acting reversible contraception.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project State Inpatient Databases, State Emergency Department Databases, and State Ambulatory Surgery and Services Databases, 2016 to 2020 for 11 states (California, Florida, Georgia, Missouri, Nebraska, Nevada, New York, South Carolina, Tennessee, Vermont, and Wisconsin).
Adjusted difference for patients who live in zip codes where closest hospital was Catholic, relative to zip codes where closest hospital was non-Catholic, calculated using demographic match.
All regressions other than subsequent deliveries were estimated using deliveries that occurred during 2016-2019. The number of deliveries was 137 178 for Black patients, 210 486 for Hispanic patients, and 732 036 for White patients.
Subsequent deliveries estimated using deliveries that occurred during 2016-2017. The number of deliveries was 70 864 for Black patients, 112 402 for Hispanic patients, and 367 868 for White patients.
In supplementary analyses, the associations of interest were more pronounced for patients who lived in nonmetropolitan counties relative to patients who lived in metropolitan counties (eTable 4 in Supplement 1). This finding was consistent with the results by the degree of exposure to Catholic hospitals likely because there are fewer hospitals and more limited hospital choice in nonmetropolitan counties. In contrast, results among patients covered by Medicaid were similar to results among patients covered by private insurance (eTable 5 in Supplement 1).
Discussion
This cohort study finds that living in a zip code in which the closest hospital was Catholic was associated with an increased probability of delivery at a Catholic hospital and a decreased probability of postpartum surgical sterilization (both at delivery and within 1 year). We also found evidence of an increase in the probability of subsequent deliveries within 3 years.
The associations between exposed and unexposed patients in the pooled sample were modest in size, with postpartum surgical sterilization decreasing by 0.95 pp (13.8%) at delivery and by 0.21 pp (13.1%) after delivery and within 1 year, and subsequent delivery increasing by 0.47 pp (2.3%). However, not all exposed patients delivered at a Catholic hospital, and we found larger changes among exposed patients whose nearest non-Catholic hospital was more than 20 miles away. For these patients, the probability of delivery at a Catholic hospital increased more than twice as much compared with patients in the pooled sample (adjusted difference, 57.85 pp vs 21.26 pp). Among exposed patients whose nearest non-Catholic hospital was more than 20 miles away, the probability of surgical sterilization decreased by 3.40 pp (27.4%), and the probability of subsequent delivery increased by 2.75 pp (12.4%). Thus, the size of the associations was meaningful for patients with limited hospital choice.
Although the Catholic affiliation of the closest hospital was associated with several outcomes, it was only minimally associated with the probability of delivering at the closest hospital. In the pooled sample, 34% of patients delivered at the closest hospital regardless of whether the hospital was Catholic, and we observed relatively small differences in the subsamples. This finding suggests that many patients may choose what hospital to deliver at based on convenience rather than preferences regarding contraceptive care. Such behavior could be explained by having limited information about the Catholic affiliation of nearby hospitals or whether contraceptive services are offered at Catholic hospitals, which is consistent with prior research.7,8,9,10,11
Decreases in the probability of postpartum surgical sterilization in this study are consistent with an earlier study12 that found when hospitals changed from non-Catholic to Catholic affiliation, tubal ligations decreased by 31%. Similarly, we estimated that the probability of surgical sterilization was 13.8% lower among patients whose closest hospital was Catholic. The previous study12 also found suggestive evidence that the effect size was larger in areas with less hospital competition. This finding is consistent with our findings that differences were larger among patients from nonmetropolitan counties and patients who were further than 20 miles from the nearest non-Catholic hospital. In recent years, many rural hospitals have closed obstetric departments,21,22,23,24 which may have further restricted patient choice and increased the influence of Catholic-affiliated hospitals in nonmetropolitan counties.
Our results by age are in line with studies of contraceptive preferences.25,26,27,28 For example, Fang and Westhoff27 found that tubal ligation rates were highest among women aged 35 years and older. Similarly, we observed that the reduction in postpartum surgical sterilization at delivery was greatest for patients aged 35 years and older (−1.67 pp).
We also found that the probability of a subsequent delivery within 3 years increased by 2.75 pp for patients whose closest hospital was Catholic and whose closest non-Catholic hospital was more than 20 miles away. While the estimated effect size was smaller in magnitude, the direction of our estimate was consistent with previous research that observed that nearly half of patients with an unfulfilled postpartum tubal ligation request became pregnant within 1 year after the index delivery.29 Our findings suggest that the increase in the probability of subsequent deliveries was higher among Black and Hispanic patients. This finding could be explained by barriers to obtaining postpartum contraception, such as lower levels of postpartum access to care.30,31
Given that many women do not know the Catholic affiliation of nearby hospitals and appear to choose the hospital where they deliver based on distance, our findings could be interpreted as a decrease in access to contraception and an increase in unintended subsequent deliveries for patients whose closest hospital is Catholic. Our study suggests that recent increases in the number of Catholic-affiliated hospitals could have important public health and economic consequences for patients who live more than 20 miles from the nearest non-Catholic hospital. Previous research has found that access to contraception is associated with improvements in infant health,32,33 maternal health,34 high school and college completion rates, and labor force participation rates for women and their families.35,36,37,38
Limitations
This study has limitations. First, the results may not be generalizable beyond the 11 included states. Nonetheless, these states span the Southeast, Southwest, West, Midwest, and Northeast regions and include 3 of the 4 most populous US states.
Second, our data captured only contraceptive services obtained at hospital inpatient departments, emergency departments, and hospital-owned ambulatory surgery departments. Outpatient contraceptive services, such as those received at Planned Parenthood, and prescription medicines were not observed. While our data have advantages relative to insurance claims (including capturing all payers and tracking patients over time even after changes in insurance policies), complementary studies with insurance claims or surveys could be useful to capture contraceptive services from all types of clinicians.
Third, unobserved characteristics, such as maternal parity, patient religion, and preferences for contraception, could be associated with whether patients resided closer to a Catholic hospital. Thus, we cannot rule out that our results are explained by patient preferences, such as a preference for less contraception among Catholic patients who happen to live closer to Catholic hospitals. Nevertheless, among exposed patients, the size of the point estimates increased as the distance from the nearest non-Catholic hospital increased. This pattern is consistent with access to contraception driving the results rather than patient preferences. Furthermore, our matching strategy reduced differences across observed covariates, even for variables that were not used for matching. Matching may have also reduced differences in unobserved characteristics, but matching did not materially change the results relative to the unmatched regressions. Thus, biases from remaining differences in unobserved covariates may be small.
Fourth, we did not observe care that patients received in states other than the one in which they delivered, as each state supplies its own encrypted patient identifier to the HCUP. Lastly, as stated earlier, findings should be interpreted as exploratory due to the potential for type I error and multiple comparisons.
Conclusions
This cohort study finds that living in a zip code in which the closest hospital was Catholic was associated with a modest decrease in the probability of postpartum sterilization (both at delivery and within 1 year). We also observed a modest increase in the probability of subsequent delivery within 3 years.
eTable 1. Codes Used to Identify Deliveries, Surgical Sterilization, Long-Acting Reversible Contraception, and Comorbid Conditions
eTable 2. Difference in Means Between Patients Who Lived Closer to Catholic Hospitals Relative to Patients Who Lived Closer to Non-Catholic Hospitals
eTable 3. Means of Outcomes in Zip Codes in Which Closest Hospital Was Non-Catholic
eTable 4. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Urbanicity of Patient’s County
eTable 5. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Primary Expected Payer
Data Sharing Statement
References
- 1.Khaikin C, Uttley L. State oversight of hospital consolidation: inadequate to protect patients’ rights and community access to care. AMA J Ethics. 2016;18(3):272-278. doi: 10.1001/journalofethics.2016.18.3.pfor3-1603 [DOI] [PubMed] [Google Scholar]
- 2.Admon LK, Villavicencio J. Catholic hospitals, patient autonomy, and sexual and reproductive health care in the United States. JAMA Netw Open. 2020;3(1):e1920131. doi: 10.1001/jamanetworkopen.2019.20131 [DOI] [PubMed] [Google Scholar]
- 3.American College of Obstetrics and Gynecology . Access to postpartum sterilization: ACOG committee opinion, number 827. Obstet Gynecol. 2021;137(6):e169-e176. doi: 10.1097/AOG.0000000000004381 [DOI] [PubMed] [Google Scholar]
- 4.Guiahi M. Religious refusals to long-acting reversible contraceptives in Catholic settings: a call for evidence. Am J Obstet Gynecol. 2020;222(4S):S869.e1-S869.e5. [DOI] [PubMed]
- 5.Wolfe ID, Pope TM. Hospital mergers and conscience-based objections—growing threats to access and quality of care. N Engl J Med. 2020;382(15):1388-1389. doi: 10.1056/NEJMp1917047 [DOI] [PubMed] [Google Scholar]
- 6.US Department of Health and Human Services . Church Amendments, 42 U.S.C. § 300a-7. Accessed August 6, 2023. https://www.hhs.gov/sites/default/files/ocr/civilrights/understanding/ConscienceProtect/42usc300a7.pdf
- 7.Guiahi M. Catholic health care and women’s health. Obstet Gynecol. 2018;131(3):534-537. doi: 10.1097/AOG.0000000000002477 [DOI] [PubMed] [Google Scholar]
- 8.Schwandt HM, Sparkle B, Post-Kinney M. Ambiguities in Washington state hospital policies, irrespective of Catholic affiliation, regarding abortion and contraception service provision. Reprod Health. 2018;15(1):178. doi: 10.1186/s12978-018-0621-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Stulberg DB, Guiahi M, Hebert LE, Freedman LR. Women’s expectation of receiving reproductive health care at Catholic and non-Catholic hospitals. Perspect Sex Reprod Health. 2019;51(3):135-142. doi: 10.1363/psrh.12118 [DOI] [PubMed] [Google Scholar]
- 10.Wascher JM, Hebert LE, Freedman LR, Stulberg DB. Do women know whether their hospital is Catholic? results from a national survey. Contraception. 2018;98(6):498-503. doi: 10.1016/j.contraception.2018.05.017 [DOI] [PubMed] [Google Scholar]
- 11.Wong ZJ, Thompson L, Boulware A, et al. What you don’t know can hurt you: patient and provider perspectives on postpartum contraceptive care in Illinois Catholic hospitals. Contraception. 2022;107:62-67. doi: 10.1016/j.contraception.2021.10.004 [DOI] [PubMed] [Google Scholar]
- 12.Hill EL, Slusky DJG, Ginther DK. Reproductive health care in Catholic-owned hospitals. J Health Econ. 2019;65:48-62. doi: 10.1016/j.jhealeco.2019.02.005 [DOI] [PubMed] [Google Scholar]
- 13.Menegay MC, Andridge R, Rivlin K, Gallo MF. Delivery at Catholic hospitals and postpartum contraception use, five US states, 2015-2018. Perspect Sex Reprod Health. 2022;54(1):5-11. doi: 10.1363/psrh.12186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Healthcare Cost and Utilization Project. Overview of the State Emergency Department Databases (SEDD). Agency for Healthcare Research and Quality. Updated September 15, 2021. Accessed December 14, 2023. https://hcup-us.ahrq.gov/seddoverview.jsp
- 15.Healthcare Cost and Utilization Project. Overview of the State Inpatient Databases (SID). Agency for Healthcare Research and Quality. Updated September 15, 2021. Accessed December 14, 2023. http://www.hcup-us.ahrq.gov/sidoverview.jsp
- 16.Healthcare Cost and Utilization Project. Overview of the State Ambulatory Surgery and Services Databases (SASD). Agency for Healthcare Research and Quality. Updated September 15, 2021. Accessed December 14, 2023. http://www.hcup-us.ahrq.gov/sasdoverview.jsp
- 17.Iacus SM, King G, Porro G. Causal inference without balance checking: coarsened exact matching. Polit Anal. 2012;20(1):1-24. doi: 10.1093/pan/mpr013 [DOI] [Google Scholar]
- 18.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083-3107. doi: 10.1002/sim.3697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Healthcare Cost and Utilization Project . Race and ethnicity data improvement toolkit. Agency for Healthcare Research and Quality. Accessed August 27, 2023. https://hcup-us.ahrq.gov/datainnovations/raceethnicitytoolkit/home_race.jsp
- 20.Pittman MA, Pierce D, Hasnain-Wynia R. Who, when, and how: the current state of race, ethnicity, and primary language data collection in hospitals. May 1, 2004. Accessed December 14, 2023. https://www.commonwealthfund.org/publications/fund-reports/2004/may/who-when-and-how-current-state-race-ethnicity-and-primary
- 21.Carroll C, Interrante JD, Daw JR, Kozhimannil KB. Association between Medicaid expansion and closure of hospital-based obstetric services. Health Aff (Millwood). 2022;41(4):531-539. doi: 10.1377/hlthaff.2021.01478 [DOI] [PubMed] [Google Scholar]
- 22.Hung P, Kozhimannil KB, Casey MM, Moscovice IS. Why are obstetric units in rural hospitals closing their doors? Health Serv Res. 2016;51(4):1546-1560. doi: 10.1111/1475-6773.12441 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hung P, Kozhimannil K, Henning-Smith C, Casey M. Closure of hospital obstetric services disproportionately affects less-populated rural counties. University of Minnesota Rural Health Research Center Policy Brief; April 2017. Accessed December 14, 2023. https://rhrc.umn.edu/wp-content/uploads/2017/04/UMRHRCOBClosuresPolicyBrief5.2.19.pdf
- 24.Kozhimannil KB, Hung P, Henning-Smith C, Casey MM, Prasad S. Association between loss of hospital-based obstetric services and birth outcomes in rural counties in the United States. JAMA. 2018;319(12):1239-1247. doi: 10.1001/jama.2018.1830 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Daniels K, Abma JC. Current contraceptive status among women aged 15-49: United States, 2015-2017. National Center for Health Statistics Brief no. 327. Accessed December 2018. https://www.cdc.gov/nchs/data/databriefs/db327-h.pdf
- 26.Dehlendorf C, Park SY, Emeremni CA, Comer D, Vincett K, Borrero S. Racial/ethnic disparities in contraceptive use: variation by age and women’s reproductive experiences. Am J Obstet Gynecol. 2014;210(6):526.e1-526.e9. doi: 10.1016/j.ajog.2014.01.037 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fang NZ, Westhoff CL. Update on incidence of inpatient tubal ligation and long-acting reversible contraception in the United States. Am J Obstet Gynecol. 2022;227(3):477.e1-477.e7. doi: 10.1016/j.ajog.2022.05.021 [DOI] [PubMed] [Google Scholar]
- 28.Sheyn D, Arora KS. Changes in rates of inpatient postpartum long-acting reversible contraception and sterilization in the USA, 2012-2016. Matern Child Health J. 2021;25(10):1562-1573. doi: 10.1007/s10995-021-03152-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Thurman AR, Janecek T. One-year follow-up of women with unfulfilled postpartum sterilization requests. Obstet Gynecol. 2010;116(5):1071-1077. doi: 10.1097/AOG.0b013e3181f73eaa [DOI] [PubMed] [Google Scholar]
- 30.Interrante JD, Admon LK, Carroll C, Henning-Smith C, Chastain P, Kozhimannil KB. Association of health insurance, geography, and race and ethnicity with disparities in receipt of recommended postpartum care in the US. JAMA Health Forum. 2022;3(10):e223292. doi: 10.1001/jamahealthforum.2022.3292 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Thiel de Bocanegra H, Braughton M, Bradsberry M, Howell M, Logan J, Schwarz EB. Racial and ethnic disparities in postpartum care and contraception in California’s Medicaid program. Am J Obstet Gynecol. 2017;217(1):47.e1-47.e7. doi: 10.1016/j.ajog.2017.02.040 [DOI] [PubMed] [Google Scholar]
- 32.Flynn J. Can expanding contraceptive access reduce adverse infant health outcomes? SSRN. Published April 6, 2023. Accessed December 14, 2023. doi: 10.2139/ssrn.4408753 [DOI]
- 33.Kost K, Lindberg L. Pregnancy intentions, maternal behaviors, and infant health: investigating relationships with new measures and propensity score analysis. Demography. 2015;52(1):83-111. doi: 10.1007/s13524-014-0359-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hawkins SS, Ghiani M, Harper S, Baum CF, Kaufman JS. Impact of state-level changes on maternal mortality: a population-based, quasi-experimental study. Am J Prev Med. 2020;58(2):165-174. doi: 10.1016/j.amepre.2019.09.012 [DOI] [PubMed] [Google Scholar]
- 35.Bailey MJ. Fifty years of family planning: new evidence on the long-run effects of increasing access to contraception. Brookings Papers Econ Activity. 2013(1):341-409. [DOI] [PMC free article] [PubMed]
- 36.Stevenson AJ, Genadek KR, Yeatman S, Mollborn S, Menken JA. The impact of contraceptive access on high school graduation. Sci Adv. 2021;7(19):eabf6732. doi: 10.1126/sciadv.abf6732 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yeatman S, Flynn JM, Stevenson A, Genadek K, Mollborn S, Menken J. Expanded contraceptive access linked to increase in college completion among women in Colorado. Health Aff (Millwood). 2022;41(12):1754-1762. doi: 10.1377/hlthaff.2022.00066 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Bailey MJ. More power to the pill: the impact of contraceptive freedom on women’s life cycle labor supply. Q J Econ. 2006;121(1):289-320. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Codes Used to Identify Deliveries, Surgical Sterilization, Long-Acting Reversible Contraception, and Comorbid Conditions
eTable 2. Difference in Means Between Patients Who Lived Closer to Catholic Hospitals Relative to Patients Who Lived Closer to Non-Catholic Hospitals
eTable 3. Means of Outcomes in Zip Codes in Which Closest Hospital Was Non-Catholic
eTable 4. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Urbanicity of Patient’s County
eTable 5. Association of Outcomes with Closest Hospital Being Catholic, Stratified by Primary Expected Payer
Data Sharing Statement
