Abstract

PURPOSE Currently, 40% of counties in the United States do not have an obstetrician or midwife, and in rural areas the likelihood of childbirth being attended to by a family medicine (FM) physician is increasing. We sought to characterize the effect of the FM presence on unit culture and a key perinatal quality metric in Iowa hospital intrapartum units.

METHODS Using a cross-sectional design, we surveyed Iowa physicians, nurses, and midwives delivering intrapartum care at hospitals participating in a quality improvement initiative to decrease the incidence of cesarean delivery. We linked respondents with their hospital characteristics and outcomes data. The primary outcome was the association between FM physician, obstetrician (OB), or both disciplines’ presence on labor and delivery and hospital low-risk, primary cesarean delivery rate. Unit culture was compared by hospital type (FM-only, OB-only, or Both).

RESULTS A total of 849 clinicians from 39 hospitals completed the survey; 13 FM-only, 11 OB-only, and 15 hospitals with both. FM-only hospitals were all rural, with <1,000 annual births. Among hospitals with <1,000 annual births, births at FM-only hospitals had an adjusted 34.3% lower risk of cesarean delivery (adjusted incident rate ratio = 0.66; 95% CI, 0.52-.0.98) compared with hospitals with both. Nurses endorsed unit norms more supportive of vaginal birth and stronger safety culture at FM-only hospitals (P <.05).

CONCLUSIONS Birthing hospitals staffed exclusively by FM physicians were more likely to have lower cesarean rates and stronger nursing-rated safety culture. Both access and quality of care provide strong arguments for reinforcing the pipeline of FM physicians training in intrapartum care.

Key words:INTRODUCTION

The United States is facing multiple maternal health crises including increasing severe maternal morbidity (SMM) and mortality, racial inequities in SMM, and a shrinking footprint of pregnancy care services in rural areas.1-4 People giving birth in rural areas experience greater rates of SMM despite controlling for sociodemographic and clinical risk factors, and this disparity is projected to increase.5 This disparity is likely driven by lack of access to prenatal and intrapartum care as well as by rural residents being disproportionately affected by public health crises.6 Losing access to hospital units providing pregnancy care in rural areas is associated with adverse birth outcomes.7 Despite this, rural birthing units continue to close, with administrators citing changing local community needs, safety concerns, such as not having enough deliveries to maintain clinical competency among staff and physicians, and financial nonviability.8 To solve these complex issues, it is necessary to look for places where a model of care is succeeding in providing safe, high-quality care for rural areas.

Access to pregnancy care services in the United States has been characterized and mapped by the March of Dimes in their recent report on maternity care deserts.9 The updated March of Dimes report adds analysis and mentions of the role of family medicine (FM) physicians in providing access to pregnancy care in rural areas. Specifically, despite the fact that 40% of counties in the United States do not have an obstetrician or midwife, only 6.5% of counties do not have an FM physician; a minority of FM physicians continue to provide intrapartum care after residency training, potentially diminishing the rural lack of access to pregnancy care.10 Family medicine physicians as providers of intrapartum care have been decreasing almost since the profession’s founding.11 However, hospitals with lower birth volumes in rural areas are more likely to be staffed by FM physicians than obstetricians, often with general surgeons used as back-up for cesarean deliveries.12 This is the current landscape in Iowa, where the Iowa Maternal Quality Care Collaborative supported a statewide quality improvement initiative to promote vaginal birth and decrease cesarean deliveries in an effort to decrease SMM, based on the Alliance for Innovation on Maternal Health patient safety bundle.13

Despite decades of research documenting the high quality of care provided by FM physicians,14-16 controversy continues regarding whether family physicians trained in existing FM residency programs should provide intrapartum obstetric care, given the substantial barriers they face; some of these barriers include being unable to find a practice that offers the opportunity to provide intrapartum obstetric care, malpractice insurance being cost prohibitive, and difficulty obtaining privileges.17,18 This controversy, examined both empirically and philosophically,11 is not new. However, in the last decade, there has been increasing attention to SMM, maternal health disparities, and the role of state-based perinatal quality improvement initiatives as a means to collect hospital performance data and improve care quality to decrease SMM. These movements, while performing critically important work, have tended to focus on engagement with obstetrician-gynecology and nursing professional organizations, and sometimes certified nurse midwives, but have not had substantial engagement or input from FM professional organizations.14 The role of family physicians in contributing to these perinatal quality initiatives has not been investigated. We aimed to use tools developed to measure perinatal care quality in the context of a statewide initiative to decrease cesarean delivery overuse to revisit the question of what effect FM physicians’ presence on labor and delivery might have.

METHODS

This study was approved by the University of Iowa Institutional Review Board. We used a cross-sectional study design linking a validated survey of Iowa clinicians’ attitudes, beliefs, and perceived unit norms with hospital outcomes data and characteristics. Forty-three of the 56 hospitals with obstetrical units participated in the Iowa Maternal Quality Care Collaborative initiative to promote vaginal birth and decrease cesarean deliveries. All participating facilities were invited to also participate in the Labor Culture Survey (LCS) as part of a larger quality improvement initiative to gain insight into each unit. The LCS is an anonymous validated survey with 6 subscales that measure various aspects of labor and delivery culture associated with cesarean delivery rates. Subscales include overestimation of cesarean birth’s safety, fear of vaginal birth, maternal agency, acceptance of physician oversight, personal support for best practices to decrease cesarean deliveries, and unit norms that support vaginal birth.19 Mean subscale scores have shown significant association with hospital-level nulliparous, term, singleton, vertex (NTSV) cesarean delivery rates in prior studies20 and individual physician-level NTSV cesarean delivery rates.21 Subscales contain varying numbers of items, which use Likert-style responses scored from 1 (Strongly Disagree) to 4 (Strongly Agree) with no neutral option, and are summed and divided by the number of items in the subscale to obtain subscale means.

Clinicians, including nurses, obstetricians, FM physicians, and midwives, providing intrapartum care at participating hospitals were invited to complete the LCS in the first 3 months of the 15-month improvement collaborative, which was launched in May 2021. Each hospital’s local quality improvement champion distributed the survey to eligible clinicians, and study staff provided weekly completion reports aggregated by role to each champion to encourage and track participation. In addition to the validated items, participants self-reported their clinical role, number of years practicing, gender, race/ethnicity, and number of years at their current hospital.

Hospital-level characteristics, including birth volume, nursery acuity level, Iowa-specific maternity care level22 (predates nationally recognized levels of maternal care23), geographic location (rural/urban), proportion of publicly insured patients, proportion of birthing patients with body mass index >30, and proportion of birthing patients aged ≥35 years, were obtained from the Iowa Department of Health and Human Services vital statistics for 2020 (initiative baseline).24 Each hospital’s regional access level to pregnancy care (maternity care desert status) by county was obtained from a publicly available database linked on the March of Dimes website,9 which has been replaced by a digital dashboard.25

Cesarean delivery rate was measured at the hospital level using the National Quality Forum-endorsed quality metric of the NTSV cesarean delivery rate by hospital26 and obtained from the Iowa Department of Health and Human Services vital statistics.24 The NTSV cesarean delivery rate has been designated as a key hospital quality measure to decrease overuse of cesarean delivery because it removes many prelabor risk factors for medically indicated cesarean delivery such as breech presentation, multiple gestation, and history of cesarean delivery.26,27 Healthy People 2030 has set a pragmatic goal NTSV cesarean delivery rate of no more than 23.6% for US hospitals.28

Statistical Methods

We categorized hospitals by the presence of different physician specialties on labor and delivery; specifically, whether FM physicians, obstetricians (OBs), or both provided intrapartum care on the unit (provider presence category). Survey participant and hospital characteristics were compared by provider presence category using the Kruskal-Wallis test for continuous variables and χ2 or Fisher exact test for categorical variables. We examined the association of hospital NTSV cesarean delivery rate with provider presence category using multivariate Poisson regression. We compared all nurse responses on LCS subscales by provider presence category using first the Kruskal-Wallis test followed by pairwise 2-sided multiple comparison analysis using the Dwass-Steel-Critchlow-Fligner method.29 We used partially completed surveys if they completed a preset minimum number of subscales (≥4 of 6 subscales). We calculated subscale scores using the mean of completed items. If any scale had <50% of the items completed, that response was excluded from analysis. By hospital unit, we evaluated for nonresponse bias by performing a sensitivity analysis limiting the nursing sample to nurses from units that had achieved a minimum response rate of 30%. We considered P values <.05 statistically significant. All statistical analysis was performed using SAS version 9.4 (SAS Institute Inc).

RESULTS

A total of 849 clinicians from 39 of the 43 eligible hospitals in Iowa participated in the LCS (Table 1). Sixty-seven percent were nurses, 11.5% OBs, 9.2% FM physicians, and 4.1% certified nurse midwives. A total of 34 hospitals achieved a minimum LCS response rate for their unit nursing staff (mean response rate 57%, range 7%-100%). There were no significant differences in race/ethnicity, gender, years practicing or working in pregnancy care, or years working at their current hospitals by provider presence category. However, hospital characteristics, varied significantly according to provider presence category (Table 2). Hospitals with only FM physicians were all in rural areas with annual delivery volumes of <1,000, had a smaller proportion of patients aged >35 years, and were more likely to have basic nurseries and level 1 maternal care designations. Only 1 FM-only and 1 OB-only hospital were designated teaching hospitals, compared with more than one-half of the Both category. Among the hospitals where FM physicians practice (FM-only and Both), respondents’ characteristics were similar across most demographic characteristics but differed in terms of the number of years practicing independently and number of years working in maternity care, which was significantly greater for FM-only hospitals (Table 3).

Table 1.

Survey Responder Characteristics by Hospital of Practice (Hospital Presence of Family Medicine Physicians and/or Obstetricians)

Table 2.

Hospital Characteristics by Family Medicine and Obstetrician Presence on Labor and Delivery

Table 3.

Characteristics of FM Physicians in FM-Only Hospitals vs Hospitals With Both Obstetrician and FM Physicians

To compare within a more homogenous group, and because FM-only hospitals all had <1,000 annual deliveries, we compared the NTSV cesarean delivery rate among hospitals with <1,000 annual deliveries (n = 29) by provider presence category (Table 4). Family medicine-only hospitals showed a significant association with decreased cesarean delivery rates compared with the Both category (Table 5). After adjusting for hospital annual birth volume, geography, proportion of birthing patients with body mass index >30, proportion of birthing patients aged >35 years, and proportion of publicly insured patients, the magnitude of this association increased. Specifically, the NTSV cesarean delivery rate was a relative 34.3% lower in FM-only hospitals (adjusted incident rate ratio = 0.66; 95% CI, 0.52-0.98) compared with the Both category (Table 5).

Table 4.

Hospital Delivery Volumes by Provider Presence Category for Hospitals With <1,000 Deliveries per Year, for 2020

Table 5.

Association of Hospital NTSV Cesarean Birth Rate and Hospital Type (Presence of Both FM & OB, FM-Only, and OB-Only) Delivery Volume <1,000

Differences in culture and individual attitudes were apparent between provider presence categories as well (Table 6). Nurses at FM-only hospitals reported overall unit norms more supportive of vaginal birth compared with nurses at hospitals with Both or OB-only (3.05 vs 2.91 vs 2.93, respectively; P = .007), with both the vaginal birth microculture and safety culture subscales showing significantly greater agreement at FM-only hospitals. Nursing agreement with best practices to support vaginal birth was significantly less at OB-only hospitals compared with Both (3.27 vs 3.37; P = .047). The directionality and significance of these findings were consistent when limiting the sample to nurses practicing at hospitals that met the minimum survey response rate. Physician attitudes differed by provider presence category as well. Family medicine physicians practicing at FM-only hospitals were more likely to endorse unit norms supportive of vaginal birth (3.30 vs 3.04; P <.01 compared to Both) (Supplemental Table 1). Obstetrician attitudes were more supportive of maternal agency in OB-only hospitals compared with Both (2.65 vs 2.34; P = .04); however, OB agreement with best practices to support vaginal birth was significantly lower at OB-only hospitals compared with Both (2.57 vs 2.83; P = .01) (Supplemental Table 2).

Table 6.

Nurse Attitudes in Hospitals With Both FM and OB vs Only OB vs Only FM

Hospital characteristics across maternity care desert categories showed notable differences. Compared with hospitals with full regional access to pregnancy care, those with moderate access were more likely to be FM-only hospitals (13% vs 60%; P = .01), have lower annual delivery volume, be in a rural area, be a nonteaching hospital, have a basic-level nursery, and have a maternal care designation of 1 (Supplemental Table 3).

DISCUSSION

This study of hospitals in Iowa examined a critical shrinking facet of pregnancy care in the United States that is provided by FM physicians in small rural hospitals located in areas with limited access to pregnancy care. These data show that FM physicians are providing the majority of pregnancy care in these small rural hospitals in Iowa; specifically, rural obstetrics care exists in these areas because of family medicine. We found that birthing hospitals staffed exclusively by FM physicians were more likely to have lower adjusted NTSV cesarean delivery rates, stronger nursing-rated patient safety culture, and unit norms more supportive of vaginal birth. These unit culture measures have shown strong associations with lower NTSV cesarean delivery rates in other states and have been associated with a hospital’s ability to change in response to quality improvement initiatives.20,31

Our findings are consistent with studies that have found pregnancy care provided by FM physicians to be of equal quality to that by OBs16,32 and extend the contribution to the quality and safety of pregnancy care provided by FM physicians as a source of strong safety culture.31,33 This contribution to unit culture, which includes teamwork and communication skills, might be a result of FM residency training, which emphasizes relationship building, continuity of care, and shared decision making.34 These skills, whereas certainly not unique to FM physicians, have been considered central aspects of the philosophy of care in FM residency training since its founding35 and in subsequent iterations of the specialty’s focus.36 This emphasis might explain why primary care patients treated by family physicians are more likely to experience patient-centered care compared with other primary care providers.37 In addition, FM training is unique in its encompassing of competency in care of both mother and neonate, namely the maternal-child dyad, an understanding of which has been proposed as a method to prevent failures in communication and cross-disciplinary collaboration, which contribute to perinatal health disparities.38 The fact that our data were collected at baseline, before significant quality improvement work had been undertaken to improve teamwork and patient-centered care, supports the idea that these skills are related to FM training. Incorporating similar training emphases in obstetrics training, particularly in larger teaching hospitals, might improve pregnancy care quality and safety across a broader swath of the United States.39,40

We also found confirmation of workforce pipeline insufficiencies in regard to FM physicians filling these critical access points in vulnerable communities with limited access to pregnancy care. Family medicine physicians at FM-only hospitals, which were more likely to be in moderate-access maternity care desert counties, had been practicing independently on average 8 years longer than their counterparts at hospitals with both FM physicians and OBs. Our findings are consistent with studies of recent FM graduates who often espouse the intention to practice pregnancy care but are not able to follow through on this intention. In the most recent survey of graduates from FM residencies, 82% of new FM graduates felt that their residency training prepared them to provide pregnancy care; however, only 11% actually end up providing it.41 The reasons for this leaky pipeline are many but include among other things disinterest, inability to find FM jobs with obstetrics, lifestyle concerns, fear of liability, malpractice insurance costs, and challenges with privileging.41 Credentialing concerns, particularly in light of rural practice where delivery volumes might be low, are significantly greater among FM physicians who end up not providing pregnancy care,42 despite existing evidence that low delivery volume does not affect quality of outcomes in pregnancy care provided by family physicians.43 The present study found evidence of high-quality outcomes in the setting of significantly lower delivery volumes in hospitals staffed exclusively by FM physicians, consistent with prior evidence.43 Iowa, as a state with a larger proportion of rural hospitals staffed only by FM physicians than many others, provides an important window into the communities across the United States living with limited access to pregnancy care. The present study shows that these hospitals provide a model to be replicated and extended.

The present study had limitations given its observational study design, which cannot infer causality. In addition, in the analysis of cesarean delivery rates by hospital, it is possible that individual physicians might be directing their prenatal patients to different hospitals based on risk profiles, which could have decreased NTSV cesarean delivery rates at lower-acuity hospitals. Based on published estimates, this would likely represent a small proportion of rural births and be less common among publicly insured patients,44 which in our sample were more often cared for in hospitals with some FM presence. The population of Iowa clinicians completing the LCS mirrored the state’s population as a whole, which is less racially and ethnically diverse than many states in the United States.45 This might limit the generalizability of our findings in states where rural public health crises intersect with structural racism.

In a 2021 article on FM’s role in pregnancy care,46 Barr asked the question, “What does society need from family medicine?” and concluded that, “comprehensive primary care for women requires a physician who can care for women’s most common health needs, which includes…perinatal health care.” Our findings extend this answer to include the specialty’s foundation in the biopsychosocial model of health. Recent Accreditation Council for Graduate Medical Education (ACGME) changes to FM pregnancy care training might have sent the message that providing pregnancy care is an exception or addition, rather than an expected part of FM graduates’ practice.47 The new competency-based education requirements state that the requirement can be completed, “with or without competence in labor and delivery.” Allowing for this exception is likely to decrease the numbers of FM graduates trained to deliver babies. This was seen after the 2014 ACGME changes allowed programs to tier their training in pregnancy care to specific intensities, after which there was a 22% decrease in deliveries performed by FM residents.48 In effect, the specialty appears to be backing away from pregnancy care at the precise moment that the unique talents of its graduates, in patient- and family centered care, teamwork, and communication, are most needed to solve a growing maternal health crisis.

Acknowledgments

We would like to acknowledge Nancy Weathers for study coordination, Stephanie Trusty for IMQCC coordination within the Iowa Department of Public Health and improvement coaching, Nicole Anderson for IMQCC program coordination, and Jill Hinkle and Amy Dunbar for IMQCC improvement coaching during the cesarean collaborative implementation.

Footnotes

Conflicts of interest: authors report none.

Annals Journal Club selection

Read or post commentaries in response to this article.

Funding support: This work was supported by the Agency for Healthcare Research and Quality (grant No. K08HS028028) and the NorthShore Auxiliary. The Iowa Maternal Quality Care Collaborative (IMQCC) is supported by a Maternal Health Innovation award from the Health Resources and Services Administration (No. U7AMC33719).

Previous presentation: This work was presented at the North American Primary Care Research Group Annual Meeting, Nov 18-22, 2022, Phoenix, Arizona.

Supplemental materials

Received for publication January 18, 2024.Revision received May 31, 2024.Accepted for publication June 4, 2024.© 2024 Annals of Family Medicine, Inc.