Abstract
Background:
We describe variation in postpartum opioid prescribing across a statewide quality collaborative and assess the proportion due to provider and hospital characteristics.
Methods:
We assessed postpartum prescribing data from nulliparous, term, singleton, vertex births between January 2020 and June 2021 included in the clinical registry of a statewide obstetric quality collaborative funded by Blue Cross Blue Shield of Michigan. Data were summarized using descriptive statistics. Mixed-effect logistic regression and linear models adjusted for patient characteristics and assessed provider- and hospital-level predictors of receiving a postpartum opioid prescription and prescription size. Relative contributions of provider and hospital characteristics were assessed using the intraclass correlation coefficient.
Results:
Of 40,589 patients birthing at 68 hospitals, 3.0% (872/29,412) received an opioid prescription after vaginal birth and 87.8% (9,812/11,177) received one after cesarean birth, with high variation across hospitals. In adjusted models, the strongest patient-level predictors of receiving a prescription were cesarean birth (aOR 899.1, 95% CI 752.8–1,066.7) and third-/fourth-degree perineal laceration (aOR 25.7, 95% CI 17.4–37.9). Receiving care from a certified nurse-midwife (aOR 0.63, 95% CI 0.48–0.82) or family medicine physician (aOR 0.60, 95%CI 0.39–0.91) was associated with lower prescribing rates. Hospital-level predictors included receiving care at hospitals with <500 annual births (aOR 4.07, 95% CI 1.61–15.0). A positive safety culture was associated with lower prescribing rates (aOR 0.37, 95% CI 0.15–0.88). Much of the variation in postpartum prescribing was attributable to providers and hospitals (prescription receipt: providers 25.1%, hospitals 12.1%; prescription size: providers 5.4%, hospitals: 52.2%).
Discussion:
Variation in postpartum opioid prescribing after birth is high and driven largely by provider- and hospital-level factors. Opioid stewardship efforts targeted at both the provider and hospital level may be effective for reducing opioid prescribing harms.
INTRODUCTION
Unexplained and unwarranted variation in healthcare delivery has been identified as a target for quality improvement.1,2 This principle has been commonly applied in maternity care to address critical issues such as the cesarean birth rate, use of episiotomy, and planned childbirth prior to 39 weeks gestational age.3–5 Identifying this variation is the first step in increasing adherence to evidence-based practices and improving the quality of care for all birthing patients and their infants.
Postpartum pain management—including opioid prescribing—is another prevalent, critical, but potentially harmful intervention in pregnancy with significant variation in practice. While opioid prescribing following childbirth (whether vaginal or cesarean) is essentially non-existent in European countries, many postpartum patients in the United States will fill an opioid prescription after birth.6,7 Estimates from national claims data demonstrate that 30% of patients after vaginal birth and 75% after cesarean birth fill at least one opioid prescription.7–9 These prescriptions are not without risk. As many as one in 75 postpartum patients who fill an opioid prescription develop new persistent opioid use (persistent opioid fills) in the year after childbirth.7,10 Thus, high-quality pain control that minimizes the potential harms of opioid prescribing after birth is critical for helping postpartum patients successfully navigate childbirth recovery and the transition to parenthood.
Prior work demonstrates significant variation in postpartum opioid prescribing rates and prescription size across regions, hospitals, and providers.11 While these studies serve as an important signal of potentially inappropriate variation in prescribing, they have been performed in insurance claims databases, which may lack nuanced assessment of individual patient factors such as complications at the time of childbirth, comorbidities, or contextual factors like culture of the labor unit. Additionally, questions of opioid use following birth have never been asked in a standardized patient population. Nulliparous, term, singleton, vertex (NTSV) patients are often used for comparisons of quality measures across providers and hospitals, as these patients are at lower risk of maternal complications during labor and delivery.12 Thus, we sought to describe opioid prescribing practices across hospitals and providers for NTSV patients cared for by hospitals participating in a statewide quality collaborative, to identify opportunities for quality improvement initiatives focused on opioid stewardship.
METHODS
Study Design
We performed a retrospective cohort study using detailed registry data obtained by chart abstraction from 68 hospitals participating in a statewide obstetrics quality collaborative focused on reduction of the NTSV cesarean birth rate. The quality collaborative began collecting data on opioid prescribing in 2020; thus, we included all patients who underwent childbirth from January 2020-June 2021. Our institution’s Institutional Review Board reviewed this study and deemed it exempt from formal review and patient consent because the dataset was deidentified. This cohort study followed recommended reporting guidelines as outlined by the STROBE guidelines.
Data Source
The obstetrics registry includes a sample of each hospital’s annual delivery volume of NTSV patients. Hospitals with <3500 annual births submit all NTSV births to the registry. Hospitals with ≥3500 annual births submit the first 100 consecutive NTSV births per month. All data are obtained by trained clinical abstractors and are collected through 42 days after birth.
Population
For this analysis, we identified all NTSV patients with a live birth, aged 12–55. If patients had multiple births during the study period, only the first was included. Patients who had an index hospitalization length of stay of more than seven days, underwent surgery within six weeks postpartum, or were currently receiving treatment for opioid use disorders were excluded, because their postpartum course was considered different from routine care. Patients with a history of substance use without evidence of treatment with medication for opioid use disorder were not excluded.7
Exposure
Childbirth in the population of interest was the primary exposure. We evaluated vaginal birth (including spontaneous and operative birth) and cesarean birth separately, given documented differences in prescribing rates and prescription sizes.7,8 We also included detailed characteristics of birth, including birth complications (e.g., fourth-degree laceration, postpartum hemorrhage) and maternal morbidity procedures (e.g., blood transfusion, hysterectomy, manual removal of the placenta) to provide greater detail on factors that might influence patients’ postpartum recovery.
Outcomes
The two primary outcomes of this study were the individual hospital rates of postpartum opioid prescribing following vaginal birth and cesarean birth calculated after adjusting for patient-, provider-, and hospital-level characteristics using mixed effect regression models. We defined postpartum opioid prescriptions as those prescribed at the time of discharge from the birth hospitalization. Opioid prescription data were abstracted using the same methods as other registry variables. Secondary outcomes were also assessed separately for vaginal and cesarean birth and included average oral morphine equivalents prescribed per patient for each hospital (OMEs, a standardized measure that includes the number of pills prescribed times the strength of the medication) and per hospital. Finally, we assessed the patient-, provider-, and hospital-level factors most associated with opioid prescribing practices.
Covariates
Adjusted models accounted for patient-, provider-, and hospital-level factors known to impact postpartum opioid prescribing, including maternal demographics (e.g., age, race, payer type); maternal comorbidities (e.g., pre-pregnancy diabetes, chronic hypertension); and maternal substance use and abuse (e.g., tobacco, alcohol).13 These factors may influence opioid prescribing through a variety of pathways, including increased complexity of peripartum care, differing experience of pain management, and increased complexity of birth and recovery. Though all conditions potentially associated with increased pain were not available in our dataset (e.g., sickle cell disease), we included opioid use in pregnancy as a proxy for complex pain.7,8,14 Of note, race was included as an important social construct through which pain and opioid prescribing may be influenced by inequitable treatment and racism, as there are clear disparities noted in the obstetric literature on assessment and management of pain in pregnant and postpartum people of color.15 We recognize that race is not a biologic construct, and include it as an exploratory assessment of inequitable treatment.16 Of note, the method of collecting race data varied across hospital settings, with only some institutions using self-reported race. All maternal conditions were obtained from pre-defined flags in the registry data abstracted from the current prenatal record for that birth. We also included the social deprivation index (SDI) to quantify individuals’ socioeconomic environment and examine the relationship between environmental social determinants of health and health outcomes. The index is a composite measure that includes seven demographic characteristics collected in the American Community Survey, including income, education, employment, housing, household characteristics, transportation, and demographics. We matched individuals’ zip codes to zip code tabulation area (ZCTA) SDI measures using a cross-walk to determine zip-code-level SDI.17
Provider-level characteristics were obtained by linking National Provider Identifier records to the Centers for Medicare & Medicaid Services’ National Plan and Provider Enumeration System downloadable file.18 Provider characteristics included gender and specialty [e.g., obstetrician-gynecologist, family medicine physician, advanced practice providers (certified nurse midwives, nurse practitioners). Hospital-level characteristics were obtained from the American Hospital Association Annual Survey database and included obstetric level of care (1, 2, or 3)19; number of annual births; hospital teaching status (sites with at least one Accreditation Council for Graduate Medical Education accredited program)20; and hospital governing body (e.g., not-for-profit, for-profit). We also included the hospital’s score on the patient safety domain of the Labor Culture Survey, a ten-minute survey administered to evaluate the beliefs and attitudes of healthcare professionals and clinicians caring for laboring patients.21 Survey results can be used to guide education and quality improvement efforts within the hospital. Hospitals in the quality collaborative voluntarily completed the Labor Culture Survey in 2021 as part of a pay for performance program. The patient safety subscale comprises eight questions that focus on team communication, patient advocacy, psychological safety to speak up, and hospital leadership support for changes to promote vaginal birth and reduce unnecessary cesareans.21 Each hospital’s score on this domain was used as a proxy for its culture of general engagement in quality improvement activities related to maternity care initiatives.
Statistical Analysis
We used descriptive statistics to report the rates of postpartum opioid prescribing and prescription size by type of birth. We used multilevel regression models to account for clustering of patients within providers and providers within hospitals. For all models, we serially adjusted for patient-, then provider- and hospital-level characteristics known to affect other quality metrics.22–24 In total, 80.4% (32,650/40,589) of patients in the analysis gave birth in a hospital that completed the Labor Culture Survey. Given the level of missingness on the Labor Culture Survey completion, we performed sensitivity analyses with and without hospitals that completed the survey and results were unchanged. Logistic regression was used to report the adjusted odds of receiving a postpartum opioid prescription. Multilevel linear regression was used for the size of the prescription in OMEs. The intraclass correlation coefficient (ICC), a measure of the relatedness of values within a cluster, was calculated to describe the amount of variation attributable to providers and hospitals after accounting for patient and provider variation.25 Analyses were performed with SAS version 9.4 (SAS Institute Inc., Cary, NC) and Stata SE version 14 (Stata Corp., College Station, TX). Statistical significance was set at a p value of <0.05 with two-sided tests.
RESULTS
Of the 40,589 eligible patients, 29,412 (72.5%) had a vaginal birth and 11,177 (27.5%) had a cesarean birth. The majority of cesarean births were unplanned (9,958/11,177, 89.1%) and the majority of vaginal births were spontaneous, unassisted (26,585/29,412, 90.4%). Compared to patients with a vaginal birth, patients with a cesarean birth were older (cesarean: 27.8 (5.6) years; vaginal 26.0 (5.3) years; p<0.001) and differed in race and insurance coverage. Patients who had a cesarean birth were more likely to have all of the comorbidities assessed and had higher rates of postpartum complications including infant neonatal intensive care unit admission (cesarean: 1,141/11,177, 10.2%; vaginal: 1,361/29,412, 4.6%; p<0.001), postpartum hemorrhage (cesarean: 1,982/11,177, 17.7%; vaginal: 1;093/29,412, 3.7%; p<0.001), and procedures for maternal morbidity, including blood transfusion (cesarean: 409/11,177, 3.7%; vaginal: 351/29,412, 1.2%; p<0.001). The majority of patients (32,650/40,589; 80.4%) gave birth at hospitals that completed the Labor Culture Survey in 2021.
The majority of patients were cared for by female providers (28,094/40,589; 69.2%) who were obstetrician-gynecologists (32,720/40589; 80.6%), followed by advanced practice providers (4,099/40,589; 10.1%). The majority of advanced practice providers were midwives (3,932/4,099; 96.0%), with few nurse practitioners (166/4,099; 4.0%) or physician’s assistants (1/4,099; 0.0%). Provider specialty differed for patients who had cesarean versus vaginal birth. Patients gave birth at hospitals of varying delivery volume, with the majority of births occurring at teaching hospitals (36,452/40,589, 89.8%) and hospitals with not-for-profit status (36,840/40,589, 90.8%). The average labor culture score for the safety subscale was 2.9 (SD 0.24) [Table 1].
Table 1.
Patients by Patient, Provider, and Hospital Characteristics
| Characteristic | All (N=40,589) | Cesarean (n=11,177) | Vaginal (n=29,412) | P-value |
|---|---|---|---|---|
| Patient Characteristics | ||||
| Discharged with an Opioid Prescriptions | 10,684 (26.3) | 9,812 (87.8) | 872 (3.0) | <0.001 |
| Opioid prescription sizea | 98.7 (139.6) | 100.4 (140.0) | 79.6 (133.7) | <0.001 |
| Number of prescriptions received at discharge | 0.392 | |||
| 1 | 10,661 (99.8) | 9,792 (99.8) | 869 (99.7) | |
| 2 | 23 (0.2) | 20 (0.2) | 3 (0.3) | |
| Age, yearsa | 26.5 (5.4) | 27.8 (5.6) | 26.0 (5.3) | <0.001 |
| Race/Ethnicity | <0.001 | |||
| Non-Hispanic White | 27,211 (67.0) | 7,220 (64.6) | 19,991 (68.0) | |
| Non-Hispanic Black | 6,776 (16.7) | 2,080 (18.6) | 4,696 (16.0) | |
| Hispanic | 2,252 (5.5) | 652 (5.8) | 1,600 (5.4) | |
| Asian | 1,576 (3.9) | 483 (4.3) | 1,093 (3.7) | |
| Multiracial | 236 (0.6) | 64 (0.6) | 172 (0.6) | |
| Others | 233 (0.6) | 79 (0.7) | 154 (0.5) | |
| Unknown | 2,305 (5.7) | 599 (5.4) | 1,706 (5.8) | |
| Health Insurance/Payer | <0.001 | |||
| Private | 24,754 (61.0) | 7,034 (62.9) | 17,720 (60.2) | |
| Medicaid | 13,705 (33.8) | 3,609 (32.3) | 10,096 (34.3) | |
| `Other | 2,130 (5.2) | 534 (4.8) | 1,596 (5.4) | |
| Social Deprivation Index (SDI)a | 48.01 (28.94) | 48.82 (28.98) | 47.69 (28.92) | 0.001 |
| Length of Stay after Delivery, daysa | 1.9 (0.7) | 2.4 (0.7) | 1.7 (0.5) | <0.001 |
| Maternal Comorbidities | ||||
| Pre-pregnancy diabetes | 437 (1.1) | 240 (2.1) | 197 (0.7) | <0.001 |
| Gestational diabetes | 3,101 (7.6) | 1,188 (10.6) | 1,913 (6.5) | <0.001 |
| Pre-pregnancy hypertension | 1,369 (3.4) | 610 (5.5) | 759 (2.6) | <0.001 |
| Gestational hypertension | 6,610 (16.3) | 2,340 (20.9) | 4,270 (14.5) | <0.001 |
| Asthma | 6,057 (14.9) | 1,772 (15.9) | 4,285 (14.6) | 0.001 |
| Alcohol use in pregnancy | 1,448 (3.6) | 449 (4.0) | 999 (3.4) | 0.003 |
| Tobacco use before/in pregnancy | 7,542 (18.6) | 2,248 (20.1) | 5,294 (18.0) | <0.001 |
| Marijuana use in pregnancy | 1,902 (13.8) | 516 (14.2) | 1,386 (14.7) | 0.509 |
| Opioid use in pregnancy | 321 (0.8) | 103 (0.9) | 218 (0.7) | <0.001 |
| Delivery Complications | ||||
| Third-degree laceration | 1,293 (3.2) | 0 (0.0) | 1,293 (4.4) | <0.001 |
| Fourth-degree laceration | 221 (0.5) | 1 (0.0) | 220 (0.7) | <0.001 |
| NICU admission | 2,502 (6.2) | 1,141 (10.2) | 1,361 (4.6) | <0.001 |
| Postpartum hemorrhage (EBL/QBL >1000) | 3,075 (7.6) | 1,982 (17.7) | 1,093 (3.7) | <0.001 |
| Maternal Morbidity Procedures | ||||
| Blood products transfusion | 760 (1.9) | 409 (3.7) | 351 (1.2) | <0.001 |
| Conversion of cardiac rhythm | 1 (0.0) | 0 (0.0) | 1 (0.0) | 0.538 |
| Drainage of abscess | 54 (0.1) | 34 (0.3) | 20 (0.1) | <0.001 |
| Drainage of tube placement | 35 (0.1) | 25 (0.2) | 10 (0.0) | <0.001 |
| Evacuation of hematoma | 51 (0.1) | 13 (0.1) | 38 (0.1) | 0.743 |
| Exploratory laparotomy | 18 (0.0) | 14 (0.1) | 4 (0.0) | <0.001 |
| Hysterectomy (peripartum/postpartum) | 10 (0.0) | 7 (0.1) | 3 (0.0) | 0.003 |
| Manual removal of the placenta | 721 (1.8) | 344 (3.1) | 377 (1.3) | <0.001 |
| Removal of RPOC with suction device | 181 (0.4) | 22 (0.2) | 159 (0.5) | <0.001 |
| Repair of complex vaginal trauma | 897 (2.2) | 9 (0.1) | 888 (3.0) | <0.001 |
| Repair of dehiscence | 27 (0.1) | 13 (0.1) | 14 (0.0) | 0.017 |
| Temporary tracheostomy | 0 (0.0) | 0 (0.0) | 0 (0.0) | N/A |
| Uterine artery embolization | 4 (0.0) | 2 (0.0) | 2 (0.0) | 0.315 |
| Ventilation | 8 (0.0) | 5 (0.0) | 3 (0.0) | 0.034 |
| Provider Characteristics | ||||
| Sex | 0.234 | |||
| Female | 28,094 (69.2) | 7,676 (68.7) | 20,418 (69.4) | |
| Male | 12,495 (30.8) | 3,501 (31.3) | 8,994 (30.6) | |
| Specialty | <0.001 | |||
| Obstetrician-Gynecologist Physician | 32,720 (80.6) | 9,285 (83.1) | 23,435 (79.7) | |
| Surgeon | 77 (0.2) | 27 (0.2) | 50 (0.2) | |
| General Practice/Family Medicine Physician | 979 (2.4) | 746 (6.7) | 233 (0.8) | |
| Midwife/Nurse Practitioner/Physician Assistant | 4,099 (10.1) | 859 (7.7) | 3,240 (11.0) | |
| Specialist, unspecified | 740 (1.8) | 226 (2.0) | 514 (1.7) | |
| Unknown | 1,974 (4.9) | 547 (4.9) | 1,427 (4.9) | |
| Hospital Characteristics | ||||
| Annual Delivery Volume | <0.001 | |||
| ≤500 | 4,101 (10.1) | 1,107 (9.9) | 2,994 (10.2) | |
| 501–1,000 | 6,152 (15.2) | 1,540 (13.8) | 4,612 (15.7) | |
| 1,001–2,000 | 7,589 (18.7) | 2,107 (18.9) | 5,482 (18.6) | |
| >2,000 | 22,747 (56.0) | 6,423 (57.5) | 16,324 (55.5) | |
| Obstetric Level of Care | <0.001 | |||
| Level 1 | 7,318 (18.0) | 1,875 (16.8) | 5,443 (18.5) | |
| Level 2 | 11,643 (28.7) | 3,199 (28.6) | 8,444 (28.7) | |
| Level 3 | 19,750 (48.7) | 5,577 (49.9) | 14,173 (48.2) | |
| Unknown | 1,878 (4.6) | 526 (4.7) | 526 (1.8) | |
| Teaching Hospital | 36,452 (89.8) | 10,041 (89.8) | 26,411 (89.8) | 0.906 |
| NICU Available | 24,255 (59.8) | 6,773 (60.6) | 17,482 (59.4) | 0.058 |
| Hospital Governing Body | <0.001 | |||
| Not-for-profit | 36,840 (90.8) | 10,150 (90.8) | 26,690 (90.7) | |
| For-profit | 2,503 (6.2) | 737 (6.6) | 1,766 (6.0) | |
| Public/Government-owned | 1,246 (3.1) | 290 (2.6) | 956 (3.3) | |
| Location | <0.001 | |||
| Urban | 37,156 (91.5) | 10,297 (92.1) | 26,859 (91.3) | |
| Suburban | 2,798 (6.9) | 748 (6.7) | 2,050 (7.0) | |
| Rural | 635 (1.6) | 132 (1.2) | 503 (1.7) | |
| Labor Culture Survey Safety Subscalea,b | 2.90 (0.24) | 2.90 (0.24) | 2.90 (0.24) | 0.527 |
Data presented as n (%) unless otherwise noted
Mean (SD)
32,650 out of 40,589 patients were delivered in a hospital where the survey was conducted
NICU=neonatal intensive care unit; EBL=estimated blood loss; QBL=quantification of blood loss; RPOC=retained products of conception
The overall rate of opioid prescribing was 26.3% (10,684/40,589): 87.8% (9,812/11,177) for cesarean birth and 3.0% (872/29,412) for vaginal birth. The median prescription size in OMEs was 90 (IQR 75–112.5), approximately 13 tabs of 5mg oxycodone, and was higher following cesarean birth (cesarean: median 90 (IQR 75–112.5); vaginal: median 60 (IQR 45–90)). The majority of patients received only one prescription (10,661/10,684, 99.8%). The most commonly prescribed opioid was hydrocodone (6,494/10,684, 60.8%) for both vaginal and cesarean birth, followed by oxycodone (3,776/10,684, 35.3%) and codeine (302/10,684, 2.8%).
Unadjusted opioid prescribing practices varied widely between hospitals. Following vaginal birth, hospital prescribing rates varied from 0.0% to 53.6% and median prescription size per hospital varied from 9 OMEs (IQR 9–9) to 150 OMEs (IQR 150–150)—the equivalent of approximately 1.5–20 tabs of 5mg oxycodone. Following cesarean birth, prescribing rates varied from 31.1% to 100.0% and median prescription size for patients who received an opioid prescription varied from 10 OMEs (IQR 9–10) to 495 OMEs (IQR 100–495)—the equivalent of approximately 2–65 tabs of 5mg oxycodone [data not shown].
Variation in opioid prescribing persisted after adjusting for patient-, provider-, and hospital-level variables. Adjusted rates of postpartum prescribing following cesarean birth per hospital ranged from 17.9%−98.7%. Adjusted rates of postpartum prescribing following vaginal birth per hospital ranged from 0.0%−49.2% [Figure 1].
Figure 1.
Rates of Postpartum Opioid Prescribing Following Vaginal and Cesarean Birth at Hospitals Completing the Labor Culture Survey
Several patient-, provider-, and hospital-level predictors of postpartum opioid prescribing rates emerged. Patient demographic factors, including non-Hispanic Black (aOR 1.23 [95% CI 1.02–1.48]) and Asian (aOR 1.48 [95% CI 1.13–1.95]) race were associated with receipt of an opioid prescription, though age, insurance status, and SDI were not. Birth characteristics including cesarean birth (aOR 899.1 [95% CI 752.8–1066.7]), fourth-degree perineal lacerations (aOR 25.7 [95% CI 17.4–37.9]), and some procedures for obstetric morbidity, such as uterine artery embolization (aOR 339.7 [95% CI 5.5–20899.9]) were associated with receipt of a postpartum opioid prescription. On the provider level, receiving care from a certified nurse-midwife (aOR 0.63 [95%CI 0.48–0.82]) or family medicine doctor (aOR 0.60 [95%CI 0.39–0.91]) was associated with a lower rate of opioid prescribing. On the hospital level, giving birth at a teaching hospital (aOR 4.07 [95% CI 1.7–9.9]) and at a hospital with <500 annual births (aOR 4.92 (95% CI 1.61–15.0]) were associated with higher rates of opioid prescribing, while giving birth in a suburban hospital (aOR 0.29 [95%CI 0.13–0.67]) was associated with lower rates of prescribing. Patients giving birth at hospitals with a high score on the safety subscale of the Labor Culture Survey were less likely to receive an opioid prescription (aOR 0.40 [95%CI 0.16–0.97]) [Table 2]. Predictors remained largely unchanged when including all hospitals, not just those with an available Labor Culture Survey safety subscale score—though the hospital-level factors were no longer significant [Appendix 1].
Table 2.
Adjusted Odds Ratio for Patients Discharged with an Opioid Prescription by Patient, Provider, and Hospital Characteristics among Hospitals Completing the Labor Culture Survey
| Characteristic | Adjusted Odds Ratio | 95% CI | P-value |
|---|---|---|---|
| Patient Characteristics | |||
| Mode of Delivery | |||
| Cesarean | 896.15 | 752.87–1,066.69 | <0.001 |
| Vaginal | Ref | Ref | Ref |
| Age | 1.01 | 1.00–1.02 | 0.137 |
| Race/Ethnicity | |||
| Non-Hispanic White | Ref | Ref | Ref |
| Non-Hispanic Black | 1.23 | 1.02–1.48 | 0.034 |
| Hispanic | 1.11 | 0.87–1.42 | 0.412 |
| Asian | 1.48 | 1.13–1.95 | 0.005 |
| Multiracial | 1.35 | 0.71–2.59 | 0.364 |
| Others | 1.19 | 0.58–2.43 | 0.630 |
| Unknown | 1.17 | 0.91–1.51 | 0.219 |
| Health Insurance/Payer | |||
| Private | Ref | Ref | Ref |
| Medicaid | 0.99 | 0.87–1.14 | 0.931 |
| Other | 1.28 | 0.84–1.95 | 0.244 |
| Social Deprivation Index (SDI) | 1.00 | 1.00–1.00 | 0.961 |
| Length of Stay after Delivery | 1.02 | 0.93–1.12 | 0.641 |
| Maternal Comorbidities | |||
| Asthma | 1.27 | 1.09–1.48 | 0.002 |
| Alcohol use in pregnancy | 1.29 | 0.95–1.75 | 0.108 |
| Tobacco use before/in pregnancy | 1.20 | 1.04–1.39 | 0.013 |
| Opioid use in pregnancy | 0.91 | 0.48–1.72 | 0.78 |
| Delivery Complications | |||
| Third-degree laceration | 6.84 | 5.30 | <0.001 |
| Fourth-degree laceration | 25.68 | 17.38 | <0.001 |
| NICU Admission | 1.13 | 0.90 | 0.308 |
| Postpartum hemorrhage (EBL/QBL >1000) | 1.15 | 0.96 | 0.136 |
| Maternal Morbidity Procedures | |||
| Blood products transfusion | 1.37 | 0.94–1.99 | 0.101 |
| Drainage of abscess | 0.22 | 0.08–0.64 | 0.005 |
| Drainage of tube placement | 2.55 | 0.56–11.51 | 0.224 |
| Exploratory laparotomy | 0.14 | 0.03–0.67 | 0.014 |
| Hysterectomy (peripartum/postpartum) | 3.13 | 0.26–38.25 | 0.372 |
| Manual removal of the placenta | 1.96 | 1.31–2.94 | 0.001 |
| Removal of RPOC with suction device | 1.22 | 0.53–2.79 | 0.635 |
| Repair of complex vaginal trauma | 1.70 | 1.26–2.31 | 0.001 |
| Repair of dehiscence | 0.31 | 0.08–1.13 | 0.076 |
| Uterine artery embolization | 339.72 | 5.52–20,899.91 | 0.006 |
| Ventilation | 1.57 | 0.00–1,377.49 | 0.896 |
| Provider Characteristics | |||
| Sex | |||
| Female | Ref | Ref | Ref |
| Male | 1.06 | 0.88–1.27 | 0.568 |
| Provider Specialty | |||
| Obstetrician-Gynecologist Physician | Ref | Ref | Ref |
| Surgeon | 0.89 | 0.11–6.89 | 0.908 |
| General Practice/Family Medicine Physician | 0.60 | 0.39–0.91 | 0.017 |
| Midwife/Nurse Practitioner/Physician Assistant | 0.63 | 0.48–0.82 | 0.001 |
| Specialist, unspecified | 1.28 | 0.70–2.32 | 0.427 |
| Unknown | 1.09 | 0.72–1.66 | 0.679 |
| Hospital Characteristics | |||
| Teaching Hospital | 4.07 | 1.66–9.93 | 0.005 |
| Hospital Annual Delivery Volume | |||
| ≤500 | 4.92 | 1.61–15.02 | 0.005 |
| 501–1,000 | 0.93 | 0.46–1.86 | 0.832 |
| 1,001–2,000 | Ref | Ref | Ref |
| >2,000 | 1.19 | 0.56–2.53 | 0.659 |
| Obstetric Level of Care | |||
| Level 1 | 0.98 | 0.45–2.09 | 0.949 |
| Level 2 | 1.05 | 0.54–2.04 | 0.888 |
| Level 3 | Ref | Ref | Ref |
| Unknown | 0.37 | 0.11–1.30 | 0.12 |
| Hospital Governing Body | |||
| Not-for-profit | Ref | Ref | Ref |
| For-profit | 2.05 | 0.55–7.72 | 0.286 |
| Public/Government-owned | 0.53 | 0.15–1.83 | 0.316 |
| Hospital Location | |||
| Urban | Ref | Ref | Ref |
| Suburban | 0.29 | 0.13–0.67 | 0.004 |
| Rural | 0.76 | 0.16–3.55 | 0.722 |
| Labor Culture Survey | 0.40 | 0.16–0.97 | 0.042 |
NICU=neonatal intensive care unit; EBL=estimated blood loss; QBL=quantification of blood loss; RPOC=retained products of conception
For patients who received an opioid prescription, those who had a cesarean birth received prescriptions that were 37% larger in OMEs (95% CI 1.32–1.42), while those with fourth-degree lacerations received prescriptions 15% larger in OMEs (95% CI 1.04–1.28) than patients without higher-order lacerations. Hospital score on the safety subscale of the Labor Culture Survey was not associated with opioid prescription size [Table 3]. Predictors remained largely unchanged when including the full population [Appendix 2].
Table 3.
Multiplicative Effects for Opioid Prescription Size by Patient, Provider, and Hospital Characteristics among Hospitals Completing the Labor Culture Survey
| Characteristic | Multiplicative Effect | 95% CI | P-value |
|---|---|---|---|
| Patient Characteristics | |||
| Mode of Delivery | |||
| Cesarean | 1.37 | 1.33–1.42 | <0.001 |
| Vaginal | Ref | Ref | Ref |
| Age | 1.00 | 1.00–1.00 | 0.894 |
| Race/Ethnicity | |||
| Non-Hispanic White | Ref | Ref | Ref |
| Non-Hispanic Black | 0.97 | 0.93–1.01 | 0.107 |
| Hispanic | 0.99 | 0.97–1.01 | 0.364 |
| Asian | 0.97 | 0.94–1.00 | 0.085 |
| Multiracial | 1.00 | 0.91–1.09 | 0.966 |
| Others | 0.98 | 0.89–1.09 | 0.734 |
| Unknown | 1.02 | 0.99–1.06 | 0.220 |
| Health Insurance/Payer | |||
| Private | Ref | Ref | Ref |
| Medicaid | 0.99 | 0.98–1.01 | 0.506 |
| Other | 1.01 | 0.95–1.07 | 0.821 |
| Social Deprivation Index (SDI) | 1.00 | 1.00–1.00 | 0.660 |
| Length of Stay after Delivery | 1.02 | 1.01–1.03 | 0.002 |
| Maternal Comorbidities | |||
| Asthma | 0.99 | 0.97–1.01 | 0.312 |
| Alcohol use in pregnancy | 1.00 | 0.96–1.04 | 0.872 |
| Tobacco use before/in pregnancy | 0.99 | 0.97–1.01 | 0.299 |
| Opioid use in pregnancy | 1.02 | 0.93–1.11 | 0.698 |
| Delivery Complications | |||
| Third-degree laceration | 1.05 | 0.97–1.13 | 0.204 |
| Fourth-degree laceration | 1.15 | 1.04–1.28 | 0.007 |
| NICU Admission | 0.97 | 0.95–1.00 | 0.065 |
| Postpartum hemorrhage (EBL/QBL >1000) | 1.02 | 1.00–1.04 | 0.088 |
| Maternal Morbidity Procedures | |||
| Blood products transfusion | 1.01 | 0.97–1.05 | 0.705 |
| Drainage of abscess | 1.00 | 0.86–1.16 | 0.998 |
| Drainage of tube placement | 1.19 | 1.00–1.41 | 0.049 |
| Exploratory laparotomy | 1.01 | 0.79–1.30 | 0.924 |
| Hysterectomy (peripartum/postpartum) | 1.41 | 0.94–2.13 | 0.098 |
| Manual removal of the placenta | 1.01 | 0.95–1.08 | 0.756 |
| Removal of retained products of conception with suction device | 1.04 | 0.89–1.20 | 0.650 |
| Repair of complex vaginal trauma | 0.90 | 0.83–0.98 | 0.010 |
| Repair of dehiscence | 0.96 | 0.76–1.20 | 0.698 |
| Uterine artery embolization | 2.26 | 1.29–3.95 | 0.004 |
| Ventilation | 0.59 | 0.38–0.92 | 0.020 |
| Provider Characteristics | |||
| Sex | |||
| Female | Ref | Ref | Ref |
| Male | 1.00 | 0.97–1.02 | 0.901 |
| Provider Specialty | |||
| Obstetrician-Gynecologist Physician | Ref | Ref | Ref |
| Surgeon | 1.14 | 0.86–1.51 | 0.349 |
| General Practice/Family Medicine Physician | 0.95 | 0.89–1.02 | 0.159 |
| Midwife/Nurse Practitioner/Physician Assistant | 0.95 | 0.92–1.00 | 0.031 |
| Specialist, unspecified | 1.02 | 0.94–1.12 | 0.634 |
| Unknown | 1.01 | 0.96–1.06 | 0.649 |
| Hospital Characteristics | |||
| Teaching Hospital | 0.98 | 0.66–1.47 | 0.939 |
| Hospital Annual Delivery Volume | |||
| ≤500 | 1.05 | 0.64–1.72 | 0.857 |
| 501–1,000 | 0.98 | 0.71–1.36 | 0.919 |
| 1,001–2,000 | Ref | Ref | Ref |
| >2,000 | 0.92 | 0.64–1.33 | 0.672 |
| Obstetric Level of Care | |||
| Level 1 | 1.22 | 0.85–1.74 | 0.274 |
| Level 2 | 1.20 | 0.87–1.65 | 0.260 |
| Level 3 | Ref | Ref | Ref |
| Unknown | 1.30 | 0.73–2.34 | 0.371 |
| Hospital Governing Body | |||
| Not-for-profit | Ref | Ref | Ref |
| For-profit | 0.92 | 0.50–1.70 | 0.800 |
| Public/Government-owned | 0.86 | 0.49–1.51 | 0.601 |
| Hospital Location | |||
| Urban | Ref | Ref | Ref |
| Suburban | 1.02 | 0.71–1.45 | 0.934 |
| Rural | 0.81 | 0.44–1.52 | 0.520 |
| Labor Culture Survey | 0.88 | 0.60–1.31 | 0.536 |
NICU=neonatal intensive care unit; EBL=estimated blood loss; QBL=quantification of blood loss
The amount of variation explained was higher at the provider level than at the hospital level in which providers were clustered for opioid prescribing rates (providers: 25.1%; hospitals: 12.1%) and higher at the hospital level than at the provider level for opioid prescription size (providers: 5.4%; hospitals: 52.2%) [Figure 2].
Figure 2.
Proportion of Variation in Outcomes Explained at the Provider and Hospital Level in Hospitals Completing the Labor Culture Survey
DISCUSSION
In this study, we found wide variation in the rate of postpartum opioid prescribing and prescription size for vaginal and cesarean birth across NTSV individuals. A variety of patient-, provider-, and hospital-level factors were associated with opioid prescribing practices. A large proportion of variation in opioid prescribing practices was attributable to providers and hospitals, suggesting an opportunity for policies and procedures to improve opioid stewardship.
Opioid prescribing rates differed in this population when compared to prior national estimates. The opioid prescribing rate following vaginal birth was much lower—3% compared to 30% in other studies.7–9 This difference may reflect national trends in opioid prescribing over time, with a general decline in postpartum prescribing following many procedures.26,27 In contrast, the opioid prescribing rate after cesarean birth was higher than prior estimates—88% compared to 75% in other studies.7–9 This difference may be explained in part by the inclusion of NTSV patients, who are less likely to have a routine scheduled cesarean birth and therefore may have higher rates of complications at the time of cesarean that influence postpartum recovery.28 Interestingly, at some hospitals, all post-cesarean patients received a postpartum opioid prescription—suggesting universal opioid prescribing without tailoring.
Rates of opioid prescribing following childbirth differ significantly from routine practice in European nations, where postpartum prescriptions are rarely provided after either vaginal or cesarean birth.7 While there are multiple potential drivers of differing practice, a likely driver is the overall climate of pain management in the United States, including historical pressure from professional societies, quality improvement groups, advocacy groups, and pharmaceutical companies to treat pain as the “fifth vital sign.”29 These phenomena have influenced the expectations in pain management and practice patterns across many aspects of care, including for birthing people.30 Further work is needed to understand the differences in patient-reported outcomes and long-term outcomes across both settings.
Many of the patient-level factors associated with opioid prescribing in prior studies were also identified in our work, including cesarean birth, higher-order perineal lacerations, tobacco use before and during pregnancy, and additional procedures at the time of childbirth.9,11 Interestingly, non-Hispanic Black race was associated with a higher rate of opioid prescribing. This finding is in contrast to prior work demonstrating that Black and Hispanic postpartum individuals are less likely to receive assessment of their pain, are more likely to have higher pain scores, and are less likely to receive opioid medication in the hospital and at the time of discharge.15,31,32 It is possible that higher prescribing rates for Black individuals represents an increased awareness of disparities in postpartum pain management, provider biases about Black postpartum individuals’ pain needs, or a more complex interplay between patient-, provider-, and hospital-level factors. Importantly, we do not have pain scores and other patient-reported outcomes available in our dataset, which limits our ability to decipher the relationship between opioid prescribing and pain experienced by the patient. Future work is needed to ensure pain management practices are equitable across marginalized groups, and to promote postpartum recovery without increasing the harms of opioid prescribing.
We also found that patients who were cared for by a certified nurse-midwife or family medicine physician during their postpartum course were less likely to receive a postpartum opioid prescription. By promoting the normalcy of postpartum recovery and focusing on a patient-centered approach, midwives may help to improve postpartum pain control and minimize the need for opioid prescriptions.33,34 The midwifery model of care is well-aligned with evidence-based opioid-sparing initiatives that focus on shared decision-making in pain management, minimizing unnecessary interventions, and optimizing non-opioid and non-pharmacologic strategies.35,36 Qualitative work may be helpful in the future for further defining these successful strategies.
Finally, provider- and hospital-level factors accounted for a significant proportion of the variation in postpartum opioid prescribing rates and prescription size. Notably, the Labor Culture Survey subscale assessing patient safety was associated with reduced rates of opioid prescribing across hospitals.21 This domain—reflecting key components of patient safety including team member communication, psychological safety to speak up, and patient advocacy—was used as a proxy measure of hospital engagement in promoting quality care across all members of the maternity care team. While the Labor Culture Survey is specifically intended to measure hospital culture surrounding cesarean overuse, it is possible that an interprofessional, collaborative approach across multiple obstetric team members has spillover effects beyond specific areas of targeted quality improvement due to optimization of communication and promotion of psychologically safe workspaces.37 In particular, a culture that supports team-based care may contribute to better postpartum pain management that maximizes non-opioid options while prioritizing excellent pain control. Focused quality improvement activities and a positive unit culture that promotes optimal team-based care may potentially improve care across multiple domains, from reduction of unnecessary cesarean births to improved opioid stewardship.
Improving the quality of postpartum pain management requires hospital-level interventions that ensure standardization of evidence-based pain management strategies across all providers. Single-institution studies of cesarean birth demonstrate the effectiveness of shared decision-making for discharge opioid prescribing, opioid-sparing protocols, and institutional implementation of electronic health record prescribing support for effectively reducing opioid prescribing while maintaining excellent postpartum pain management.38–41 In particular, tailoring opioid prescribing to individuals’ pain scores and medication use in the 24 hours prior to discharge can significantly reduce prescription size, highlighting an opportunity for improved opioid stewardship.38,42,43 Similar interventions following vaginal birth have been sparse, but may be particularly effective in hospitals with higher rates of opioid prescribing than their peer institutions.44 Standardization of postoperative pain management in other specialties across statewide collaboratives has demonstrated immediate and drastic declines in postoperative opioid prescribing. These effects have “spilled over” to other procedures at target institutions, highlighting the generalizability of these pain management principles.45 Future work, grounded in implementation science, is needed to understand the barriers and facilitators of opioid stewardship practices in the maternity care setting to design strategies that ensure every patient receives high-quality postpartum pain management that reduces their risk of opioid harms.
Our study has several strengths, including the use of detailed registry data and the incorporation of multiple patient-, provider-, and hospital-level factors known to influence opioid prescribing and general quality improvement efforts. We uniquely include only NTSV patients, which allows for more consistent comparisons across hospitals. We also acknowledge important limitations. Our study was completed in a single midwestern state, which may not reflect prescribing practices in other regions of the United States. We restricted our analysis to hospitals that completed the Labor Culture Survey, potentially limiting the generalizability of our findings. However, we performed sensitivity analyses demonstrating that the overarching findings of this work are similar between the group that completed the Labor Culture Survey and the full cohort. While the Labor Culture survey provides an important novel finding about the relationship between hospital culture and opioid prescribing, we cannot evaluate detailed ecological data including provider and organization factors (e.g., selected quality improvement initiatives) in this quantitative analysis. While there were no statewide initiatives to reduce opioid prescribing at the time of childbirth during the study period, we cannot confirm whether individual hospitals engaged in local opioid stewardship efforts. Finally, we are only able to measure whether an opioid was prescribed and the size of the prescription, not how much was consumed or if the patient consumed opioids that were not prescribed. We also cannot assess if individuals used opioid-sparing methods including pharmacologic and non-prescription medications (e.g., acetaminophen and ibuprofen), which may influence an individual’s postpartum pain experience.
Opioid prescribing practices, including rates of opioid prescribing and prescription size, varied widely across NTSV patients giving birth at hospitals participating in a statewide obstetric quality collaborative. Much of this variation is driven by provider- and hospital-level factors, suggesting an opportunity to improve care quality through multilevel initiatives. Future work is needed to identify equitable, patient-centered practices for effectively and equitably improving postpartum pain management while minimizing the risks of opioid prescribing.
Acknowledgments:
The authors wish to acknowledge the Obstetrics Initiative (OBI) for providing access to the data used in this study. Support for OBI is provided by Blue Cross Blue Shield of Michigan and Blue Care Network as part of the BCBSM Value Partnerships program.
Funding:
Dr. Waljee receives funding from the Michigan Department of Health & Human Services (E20221872–00 / AQ), SAMHSA SOR 2/SOR NCE (MDHHS – SAMHSA Subcontract; E20221302–00 / AR). Drs. Waljee and Moniz receive funding from the National Institutes of Health/Food and Drug Administration (U01 FD007803). Dr. Moniz receives funding from the Agency for Healthcare Research and Quality (K08 HS025465). The Obstetrics Initiative is supported by Blue Cross Blue Shield of Michigan.
APPENDIX
Appendix 1.
Adjusted Odds Ratios for Patients Discharged with an Opioid Prescription by Patient, Provider, and Hospital Characteristics for All Hospitalsa
| Characteristic | Adjusted Odds Ratio | 95% CI | P-value |
|---|---|---|---|
| Patient Characteristics | |||
| Mode of Delivery | |||
| Cesarean | 801.57 | 687.84–934.11 | <0.001 |
| Vaginal | Ref | Ref | Ref |
| Age | 1.01 | 1.00–1.02 | 0.103 |
| Race/Ethnicity | |||
| Non-Hispanic White | Ref | Ref | Ref |
| Non-Hispanic Black | 1.18 | 1.00–1.39 | 0.046 |
| Hispanic | 1.09 | 0.88–1.35 | 0.444 |
| Asian | 1.33 | 1.05–1.70 | 0.019 |
| Multiracial | 1.40 | 0.77–2.55 | 0.269 |
| Others | 1.32 | 0.71–2.43 | 0.379 |
| Unknown | 1.12 | 0.90–1.40 | 0.305 |
| Health Insurance/Payer | |||
| Private | Ref | Ref | Ref |
| Medicaid | 1.00 | 0.89–1.13 | 0.967 |
| Other | 1.40 | 1.01–1.92 | 0.042 |
| Social Deprivation Index (SDI) | 1.00 | 1.00–1.00 | 0.974 |
| Length of Stay after Delivery | 1.04 | 0.96–1.12 | 0.370 |
| Maternal Comorbidities | |||
| Asthma | 1.18 | 1.03–1.35 | 0.017 |
| Alcohol use in pregnancy | 1.18 | 0.91–1.54 | 0.220 |
| Tobacco use before/in pregnancy | 1.20 | 1.05–1.37 | 0.006 |
| Opioid use in pregnancy | 1.19 | 0.69–2.05 | 0.525 |
| Delivery Complications | |||
| Third-degree laceration | 6.09 | 4.85–7.64 | <0.001 |
| Fourth-degree laceration | 26.32 | 18.64–37.17 | <0.001 |
| NICU Admission | 1.14 | 0.95–1.38 | 0.162 |
| Postpartum hemorrhage (EBL/QBL >1000) | 1.16 | 0.98–1.37 | 0.079 |
| Maternal Morbidity Procedures | |||
| Blood products transfusion | 1.30 | 0.94–1.79 | 0.111 |
| Drainage of abscess | 0.26 | 0.09–0.72 | 0.010 |
| Drainage of tube placement | 2.35 | 0.54–10.30 | 0.258 |
| Exploratory laparotomy | 0.18 | 0.04–0.79 | 0.023 |
| Hysterectomy (peripartum/postpartum) | 3.74 | 0.38–36.40 | 0.256 |
| Manual removal of the placenta | 1.80 | 1.23–2.65 | 0.003 |
| Removal of RPOC with suction device | 1.19 | 0.57–2.48 | 0.650 |
| Repair of complex vaginal trauma | 1.72 | 1.29–2.29 | <0.001 |
| Repair of dehiscence | 0.30 | 0.08–1.10 | 0.069 |
| Uterine artery embolization | 264.11 | 4.18–16,706.27 | 0.008 |
| Ventilation | 0.31 | 0.01–8.26 | 0.484 |
| Provider Characteristics | |||
| Sex | |||
| Female | Ref | Ref | Ref |
| Male | 1.04 | 0.89–1.21 | 0.647 |
| Provider Specialty | |||
| Obstetrician-Gynecologist Physician | Ref | Ref | Ref |
| Surgeon | 0.96 | 0.16–5.88 | 0.964 |
| General Practice/Family Medicine Physician | 0.69 | 0.48–1.01 | 0.056 |
| Midwife/Nurse Practitioner/Physician Assistant | 0.66 | 0.51–0.84 | 0.001 |
| Specialist, unspecified | 1.41 | 0.82–2.42 | 0.215 |
| Unknown | 1.08 | 0.75–1.56 | 0.663 |
| Hospital Characteristics | |||
| Teaching Hospital | 1.69 | 0.62–4.62 | 0.307 |
| Hospital Annual Delivery Volume | |||
| ≤500 | 1.72 | 0.52–5.70 | 0.373 |
| 501–1000 | 0.88 | 0.38–2.04 | 0.766 |
| 1001–2000 | Ref | Ref | Ref |
| >2000 | 1.09 | 0.45–2.68 | 0.844 |
| OB Level of Care | |||
| Level 1 | 0.85 | 0.33–2.14 | 0.725 |
| Level 2 | 1.04 | 0.46–2.37 | 0.929 |
| Level 3 | Ref | Ref | Ref |
| Unknown | 0.66 | 0.14–3.15 | 0.603 |
| Hospital Governing Body | |||
| Not-for-profit | Ref | Ref | Ref |
| For-profit | 0.74 | 0.25–2.15 | 0.577 |
| Public/Government-owned | 0.63 | 0.13–2.94 | 0.554 |
| Hospital Location | |||
| Urban | Ref | Ref | Ref |
| Suburban | 0.52 | 0.21–1.31 | 0.167 |
| Rural | 1.66 | 0.44–6.23 | 0.450 |
Intraclass correlation coefficient (ICC): Provider-Level: 0.306; Hospital-Level: 0.209
NICU=neonatal intensive care unit; EBL=estimated blood loss; QBL=quantification of blood loss; RPOC=retained products of conception
Appendix 2.
Multiplicative Effects for Opioid Prescription Size by Patient, Provider, and Hospital Characteristics for All Hospitalsa
| Characteristic | Multiplicative Effect | 95% CI | P-value |
|---|---|---|---|
| Patient Characteristics | |||
| Mode of Delivery | |||
| Cesarean | 1.38 | 1.33–1.42 | <0.001 |
| Vaginal | Ref | Ref | Ref |
| Age | 1.00 | 1.00–1.00 | 0.632 |
| Race/Ethnicity | |||
| Non-Hispanic White | Ref | Ref | Ref |
| Non-Hispanic Black | 0.98 | 0.95–1.01 | 0.194 |
| Hispanic | 0.99 | 0.97–1.01 | 0.267 |
| Asian | 0.97 | 0.94–1.00 | 0.050 |
| Multiracial | 1.01 | 0.93–1.11 | 0.778 |
| Others | 1.00 | 0.92–1.09 | 0.980 |
| Unknown | 1.03 | 1.00–1.06 | 0.080 |
| Health Insurance/Payer | |||
| Private | Ref | Ref | Ref |
| Medicaid | 0.98 | 0.97–1.00 | 0.044 |
| Other | 1.00 | 0.96–1.05 | 0.958 |
| Social Deprivation Index (SDI) | 1.00 | 1.00–1.00 | 0.879 |
| Length of Stay after Delivery | 1.01 | 1.00–1.02 | 0.016 |
| Maternal Comorbidities | |||
| Asthma | 0.99 | 0.98–1.01 | 0.495 |
| Alcohol use in pregnancy | 0.99 | 0.96–1.03 | 0.722 |
| Tobacco use before/in pregnancy | 0.99 | 0.97–1.01 | 0.159 |
| Opioid use in pregnancy | 1.16 | 1.08–1.25 | <0.001 |
| Delivery Complications | |||
| Third-degree laceration | 1.02 | 0.95–1.09 | 0.589 |
| Fourth-degree laceration | 1.07 | 0.97–1.17 | 0.164 |
| NICU Admission | 0.98 | 0.96–1.01 | 0.159 |
| Postpartum hemorrhage (EBL/QBL >1000) | 1.01 | 1.00–1.03 | 0.141 |
| Maternal Morbidity Procedures | |||
| Blood products transfusion | 1.00 | 0.96–1.04 | 0.934 |
| Drainage of abscess | 0.98 | 0.85–1.14 | 0.797 |
| Drainage of tube placement | 1.19 | 1.00–1.41 | 0.050 |
| Exploratory laparotomy | 1.03 | 0.82–1.29 | 0.787 |
| Hysterectomy (peripartum/postpartum) | 1.30 | 0.92–1.83 | 0.144 |
| Manual removal of the placenta | 1.00 | 0.94–1.07 | 0.984 |
| Removal of RPOC with suction device | 1.03 | 0.89–1.18 | 0.724 |
| Repair of complex vaginal trauma | 0.93 | 0.86–1.00 | 0.045 |
| Repair of dehiscence | 0.96 | 0.76–1.22 | 0.762 |
| Uterine artery embolization | 1.72 | 1.10–2.71 | 0.019 |
| Ventilation | 0.67 | 0.45–0.99 | 0.042 |
| Provider Characteristics | |||
| Sex | |||
| Female | Ref | Ref | Ref |
| Male | 0.99 | 0.97–1.01 | 0.449 |
| Provider Specialty | |||
| Obstetrician-Gynecologist Physician | Ref | Ref | Ref |
| Surgeon | 1.10 | 0.85–1.43 | 0.448 |
| General Practice/Family Medicine Physician | 0.93 | 0.87–0.99 | 0.016 |
| Midwife/Nurse Practitioner/Physician Assistant | 0.96 | 0.92–1.00 | 0.036 |
| Specialist, unspecified | 1.03 | 0.94–1.11 | 0.554 |
| Unknown | 1.02 | 0.97–1.06 | 0.489 |
| Hospital Characteristics | |||
| Teaching Hospital | 0.96 | 0.69–1.33 | 0.787 |
| Hospital Annual Delivery Volume | |||
| ≤500 | 0.95 | 0.64–1.41 | 0.797 |
| 501–1000 | 0.98 | 0.74–1.30 | 0.878 |
| 1001–2000 | Ref | Ref | Ref |
| >2000 | 0.93 | 0.69–1.26 | 0.647 |
| OB Level of Care | |||
| Level 1 | 1.15 | 0.84–1.58 | 0.369 |
| Level 2 | 1.14 | 0.86–1.51 | 0.351 |
| Level 3 | Ref | Ref | Ref |
| Unknown | 1.32 | 0.78–2.24 | 0.297 |
| Hospital Governing Body | |||
| Not-for-profit | Ref | Ref | Ref |
| For-profit | 0.93 | 0.65–1.33 | 0.689 |
| Public/Government-owned | 0.86 | 0.52–1.44 | 0.570 |
| Hospital Location | |||
| Urban | Ref | Ref | Ref |
| Suburban | 1.05 | 0.78–1.41 | 0.746 |
| Rural | 0.95 | 0.63–1.44 | 0.809 |
Intraclass correlation coefficient (ICC): Provider-Level: 0.053; Hospital-Level: 0.480
NICU=neonatal intensive care unit; EBL=estimated blood loss; QBL=quantification of blood loss; RPOC=retained products of conception
Footnotes
Conflicts of Interest: Dr. Peahl is a paid consultant for Maven Clinic. Dr. Moniz is a paid consultant for RAND and the Society of Family Planning. The remaining authors report no conflicts of interest.
References
- 1.Fisher ES, Bynum JP, Skinner JS. Slowing the growth of health care costs--lessons from regional variation. N Engl J Med. 2009;360(9):849–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Newhouse JP, Garber AM. Geographic variation in health care spending in the United States: insights from an Institute of Medicine report. JAMA. 2013;310(12):1227–1228. [DOI] [PubMed] [Google Scholar]
- 3.Kozhimannil KB, Law MR, Virnig BA. Cesarean delivery rates vary tenfold among US hospitals; reducing variation may address quality and cost issues. Health Aff (Millwood). 2013;32(3):527–535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Friedman AM, Ananth CV, Prendergast E, D’Alton ME, Wright JD. Variation in and factors associated with use of episiotomy. JAMA. 2015;313(2):197–199. [DOI] [PubMed] [Google Scholar]
- 5.Tita AT, Landon MB, Spong CY, et al. Timing of elective repeat cesarean delivery at term and neonatal outcomes. N Engl J Med. 2009;360(2):111–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wong CA, Girard T. Undertreated or overtreated? Opioids for postdelivery analgesia. Br J Anaesth. 2018;121(2):339–342. [DOI] [PubMed] [Google Scholar]
- 7.Peahl AF, Dalton VK, Montgomery JR, Lai YL, Hu HM, Waljee JF. Rates of New Persistent Opioid Use After Vaginal or Cesarean Birth Among US Women. JAMA Netw Open. 2019;2(7):e197863. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bateman BT, Franklin JM, Bykov K, et al. Persistent opioid use following cesarean delivery: patterns and predictors among opioid-naive women. Am J Obstet Gynecol. 2016;215(3):353 e351–353 e318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Osmundson SS, Wiese AD, Min JY, et al. Delivery type, opioid prescribing, and the risk of persistent opioid use after delivery. Am J Obstet Gynecol. 2019;220(4):405–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Peahl AF, Morgan DM, Dalton VK, et al. New persistent opioid use after acute opioid prescribing in pregnancy: a nationwide analysis. Am J Obstet Gynecol. 2020;223(4):566 e561–566 e513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Peahl AF, Morgan DM, Langen ES, et al. Variation in Opioid Prescribing After Vaginal and Cesarean Birth: A Statewide Analysis. Womens Health Issues. 2022;S1049–3867(22)00090–1 [Online ahead of print]. [DOI] [PubMed] [Google Scholar]
- 12.Main EK, Moore D, Farrell B, et al. Is there a useful cesarean birth measure? Assessment of the nulliparous term singleton vertex cesarean birth rate as a tool for obstetric quality improvement. Am J Obstet Gynecol. 2006;194(6):1644–1651; discussion 1651–1642. [DOI] [PubMed] [Google Scholar]
- 13.Brummett CM, Waljee JF, Goesling J, et al. New Persistent Opioid Use After Minor and Major Surgical Procedures in US Adults. JAMA Surg. 2017;152(6):e170504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zhu Y, Huybrechts KF, Desai RJ, et al. Prescription opioid use after vaginal delivery and subsequent persistent opioid use and misuse. Am J Obstet Gynecol MFM. 2021;3(2):100304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Badreldin N, Grobman WA, Yee LM. Racial Disparities in Postpartum Pain Management. Obstet Gynecol. 2019;134(6):1147–1153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Vyas DA, Eisenstein LG, Jones DS. Hidden in Plain Sight - Reconsidering the Use of Race Correction in Clinical Algorithms. N Engl J Med. 2020;383(9):874–882. [DOI] [PubMed] [Google Scholar]
- 17.Robert Graham Center. Social Deprivation Index (SDI). https://www.graham-center.org/maps-data-tools/social-deprivation-index.html. Accessed February 27, 2023.
- 18.Centers for Medicare & Medicaid Services. Data Dissemination. https://www.cms.gov/Regulations-and-Guidance/Administrative-Simplification/NationalProvIdentStand/DataDissemination Accessed November 21, 2022.
- 19.American Association of Birth Centers, Association of Women’s Health Obstetric and Neonatal Nurses, American College of Obstetricians and Gynecologists, et al. Obstetric Care Consensus #9: Levels of Maternal Care. Am J Obstet Gynecol. 2019;221(6):B19–B30. [DOI] [PubMed] [Google Scholar]
- 20.Agency for Healthcare Research and Quality. Healthcare Cost & Utilization Project: NIS Description of Data Elements. www.hcup-us.ahrq.gov/db/vars/hosp_locteach/nisnote.jsp. Accessed January 17, 2023.
- 21.White VanGompel E, Perez S, Wang C, Datta A, Cape V, Main E. Measuring labor and delivery unit culture and clinicians’ attitudes toward birth: Revision and validation of the Labor Culture Survey. Birth. 2019;46(2):300–310. [DOI] [PubMed] [Google Scholar]
- 22.Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med. 2009;361(14):1368–1375. [DOI] [PubMed] [Google Scholar]
- 23.Krell RW, Girotti ME, Dimick JB. Extended length of stay after surgery: complications, inefficient practice, or sick patients? JAMA Surg. 2014;149(8):815–820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Leonard PSJ, Crouse DL, Boudreau JG, Gupta N, McDonald JT. Provider volume and maternal complications after Caesarean section: results from a population-based study. BMC Pregnancy Childbirth. 2020;20(1):37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Killip S, Mahfoud Z, Pearce K. What is an intracluster correlation coefficient? Crucial concepts for primary care researchers. Ann Fam Med. 2004;2(3):204–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Agarwal S, Bryan JD, Hu HM, et al. Association of State Opioid Duration Limits With Postoperative Opioid Prescribing. JAMA Netw Open. 2019;2(12):e1918361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Howard R, Alameddine M, Klueh M, et al. Spillover Effect of Evidence-Based Postoperative Opioid Prescribing. J Am Coll Surg. 2018;227(3):374–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Declercq E, Barger M, Cabral HJ, et al. Maternal outcomes associated with planned primary cesarean births compared with planned vaginal births. Obstet Gynecol. 2007;109(3):669–677. [DOI] [PubMed] [Google Scholar]
- 29.Chidgey BA, McGinigle KL, McNaull PP. When a Vital Sign Leads a Country Astray-The Opioid Epidemic. JAMA Surg. 2019;154(11):987–988. [DOI] [PubMed] [Google Scholar]
- 30.Ecker J, Abuhamad A, Hill W, et al. Substance use disorders in pregnancy: clinical, ethical, and research imperatives of the opioid epidemic: a report of a joint workshop of the Society for Maternal-Fetal Medicine, American College of Obstetricians and Gynecologists, and American Society of Addiction Medicine. Am J Obstet Gynecol. 2019;221(1):B5–B28. [DOI] [PubMed] [Google Scholar]
- 31.Wiles A, Korn E, Dinglas C, Bentley B, Rosne J, Rahimi S. Disparities in Post Cesarean Section Pain Management. J Clin Gynecol Obstet. 2022;11(2):27–32. [Google Scholar]
- 32.Johnson JD, Asiodu IV, McKenzie CP, et al. Racial and Ethnic Inequities in Postpartum Pain Evaluation and Management. Obstet Gynecol. 2019;134(6):1155–1162. [DOI] [PubMed] [Google Scholar]
- 33.American College of Nurse-Midwives. Our Philosophy of Care. https://www.midwife.org/our-philosophy-of-care. Accessed January 17, 2023.
- 34.Leziak K, Yee LM, Grobman WA, Badreldin N. Patient Experience with Postpartum Pain Management in the Face of the Opioid Crisis. J Midwifery Womens Health. 2021;66(2):203–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Macones GA, Caughey AB, Wood SL, et al. Guidelines for postoperative care in cesarean delivery: Enhanced Recovery After Surgery (ERAS) Society recommendations (part 3). Am J Obstet Gynecol. 2019;221(3):247 e241–247 e249. [DOI] [PubMed] [Google Scholar]
- 36.Peahl AF, Smith R, Johnson TRB, Morgan DM, Pearlman MD. Better late than never: why obstetricians must implement enhanced recovery after cesarean. Am J Obstet Gynecol. 2019;221(2):117 e111–117 e117. [DOI] [PubMed] [Google Scholar]
- 37.White VanGompel E, Perez S, Datta A, Wang C, Cape V, Main E. Cesarean overuse and the culture of care. Health Serv Res. 2019;54(2):417–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Prabhu M, Dubois H, James K, et al. Implementation of a Quality Improvement Initiative to Decrease Opioid Prescribing After Cesarean Delivery. Obstet Gynecol. 2018;132(3):631–636. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.MacGregor CA, Neerhof M, Sperling MJ, et al. Post-Cesarean Opioid Use after Implementation of Enhanced Recovery after Surgery Protocol. Am J Perinatol. 2021;38(7):637–642. [DOI] [PubMed] [Google Scholar]
- 40.Shinnick JK, Ruhotina M, Has P, et al. Enhanced Recovery after Surgery for Cesarean Delivery Decreases Length of Hospital Stay and Opioid Consumption: A Quality Improvement Initiative. Am J Perinatol. 2020;38(S 01):e215–e223. [DOI] [PubMed] [Google Scholar]
- 41.McCoy JA, Gutman S, Hamm RF, Srinivas SK. The Association between Implementation of an Enhanced Recovery after Cesarean Pathway with Standardized Discharge Prescriptions and Opioid Use and Pain Experience after Cesarean Delivery. Am J Perinatol. 2021;38(13):1341–1347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Prabhu M, McQuaid-Hanson E, Hopp S, et al. A Shared Decision-Making Intervention to Guide Opioid Prescribing After Cesarean Delivery. Obstet Gynecol. 2017;130(1):42–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Osmundson SS, Raymond BL, Kook BT, et al. Individualized Compared With Standard Postdischarge Oxycodone Prescribing After Cesarean Birth: A Randomized Controlled Trial. Obstet Gynecol. 2018;132(3):624–630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Mills JR, Huizinga MM, Robinson SB, et al. Draft Opioid-Prescribing Guidelines for Uncomplicated Normal Spontaneous Vaginal Birth. Obstet Gynecol. 2019;133(1):81–90. [DOI] [PubMed] [Google Scholar]
- 45.Vu JV, Howard RA, Gunaseelan V, Brummett CM, Waljee JF, Englesbe MJ. Statewide Implementation of Postoperative Opioid Prescribing Guidelines. N Engl J Med. 2019;381(7):680–682. [DOI] [PMC free article] [PubMed] [Google Scholar]