Women’s Health Service Provider Model Components

Published 2020

In this module:

Women's health refers to the influence of sex and gender on health, wellness, disability, and disease status across the lifespan. Biological, cultural, psychosocial, and socioeconomic factors within environmental and geographic contexts also shape each woman’s health experiences and quality of life. Addressing women’s health from a sex and gender perspective provides an opportunity to manage conditions that may manifest, progress, and/or remit differently in women as compared to men. It also allows other factors that influence use of health care services to be addressed. Health care access and service delivery systems can be coordinated and tailored in ways that improve health outcomes for this population.  

A group of health care professionals, traditionally referred to as “women’s health providers”, delivers obstetrics, gynecology, and other preventive and reproductive care predominantly or solely to women. Women’s health providers may also offer these same health services to individuals who do not self-identify as female. These providers include obstetrics/gynecology physicians (OB-GYN), certified nurse midwives (CNMs), and nurse practitioners (NPs) and physician assistants (PAs) specialized in women’s health. This module focuses on these providers and documents efforts to estimate the portion of family medicine physicians’ time spent on women’s health care provision.

Modeling Supply

Estimating the Current Active Workforce Supply 

The women’s health services workforce includes: 

  • obstetricians, gynecologists, or obstetrician/gynecologists (hereafter referred to as OB-GYNs) 
  • certified nurse midwives and certified midwives (usually referred to as certified nurse midwives or just nurse midwives in this report due to the rarity of certified midwives) 
  • select NPs 
  • select PAs 

To model the supply of these occupations, the HWSM estimates the number and characteristics of current supply, models the number and characteristics of new entrants to supply, models retirement, and models workforce behavior such as hours worked patterns and geographic mobility. The Status Quo scenario models the continuation of current workforce participation patterns and new entrants. A “Graduates Trend” scenario for the rapidly growing occupations of NPs and PAs projects the number of future new entrants instead of using the most recent historical numbers for all future years. 

Physicians: The source for estimating the number, characteristics and geographic location of OB-GYNs is the 2018 American Medical Association (AMA) Physician Masterfile. The analysis is limited to physicians who have completed their residency and whose status is listed as ‘active’. Active is defined by AMA as working 20 or more hours per week in professional activities. The AMA file is known to misclassify older physicians who have retired as active. This can be due to time lags between when physicians retire and when their status is updated in the AMA file. To address this issue, we omitted physicians age 75 and older from the estimate of starting supply. Without making this adjustment, projections of future supply would drop during the first few years as the status of many older physicians changed to ‘retired’ in the simulation model. This is because the age distribution of physicians currently in the AMA file is inconsistent with an age distribution indicated by retirement patterns. 

Identifying OB-GYNs in the AMA Masterfile considered both the first and second recorded specialty. If the physician’s first specialty or second specialty is obstetrics, gynecology, obstetrics and gynecology, or something similar such as maternal and fetal medicine, then the physician is categorized as an OB-GYN. Additionally, in a small number of cases where a physician had a different surgery type as the first recorded specialty and OB-GYN as a second recorded specialty, the physician is categorized as an OB-GYN. This reflects the extensive specialization that is required to practice in the women’s health field and an assumption that physicians are unlikely to obtain this expertise and not utilize it. Physicians with two different surgical specializations are extremely rare in the AMA data. As they do exist, the modeling team has to determine how to categorize them. The vast majority of physicians classified as OB-GYNs have obstetrics and gynecology as their first specialty and no second specialty.  

Estimates by state and by metropolitan/nonmetropolitan location are based on primary office location address in the AMA Masterfile. When that field is unavailable, the estimates are based on the physician’s mailing address. Primary office location is unavailable for less than 5% of the records. 

Nurse midwives: Certified nurse midwives are registered nurses that take additional training to focus in healthcare related to childbirth. The 2018 NSSRN included nurse midwives in the survey sample frame. Unlike NPs, they were clearly identified by their nurse type and did not need to be identified based on their specialty. We also included certified midwives. This group has similar training to nurse midwives but does not have a nursing degree and so does not appear in the NSSRN. To account for certified midwives, we increased the number of midwives in New York by 108. According to the American College of Nurse-Midwives (which also represents certified midwives), all practicing certified midwives are in New York. Discussions with the Center for Health Workforce Studies in Albany, NY, found that there are currently 108 nurse midwives in New York.  

Nurse practitioners: The 2018 NSSRN included NPs in the survey sample frame. We estimate the starting supply of nurse practitioners in women’s health using weighted responses from nurse practitioners who listed specialty of gynecology (n=7,582), obstetrics (n=2,754), or labor and delivery (n=1,509). State and by metropolitan/nonmetropolitan location estimates are based on location of the primary nursing position held on December 31, 2017 as reported in the NSSRN restricted file. The sample size is sufficient to estimate NP supply by state. The sample size is insufficient to estimate NP supply by metropolitan/nonmetropolitan location within each state. It is, however, sufficient to estimate the supply of NPs in metropolitan and nonmetropolitan areas at the national level. 

Physician assistants: Estimates of the number, characteristics, and location of PAs in women’s health come from two data sources. They are 2018 data from the National Commission on Certification of Physician Assistants (NCCPA)1 and 2018-2019 data from the American Academy of Physician Assistants (AAPA). NCCPA has data on all certified PAs. NCCPA’s Professional Profile Survey collects data on practice specialty. AAPA survey data combined with membership files provided greater flexibility to categorize PAs by whether they are working in metropolitan or a nonmetropolitan area. They also provide information on the demographics of PAs. Consequently, we used NCCPA data to estimate the total number of active PAs in women’s health. We also used AAPA data to estimate the distribution of women’s health PAs by state, the proportion working in a metropolitan or a nonmetropolitan area, and PA demographics. 

The constructed starting year file used for modeling women’s health PA supply consists of 1,518 records that translate to 1,480 FTEs. The PA age, sex, state, and metropolitan/nonmetropolitan location information reflects the best available data of the characteristics and geography of primary care PA supply in 2018. 

Modeling New Entrants

New entrants are added to the workforce each year via a “synthetic” population of the profession created based on the number and characteristics of recent graduates in each occupation and specialty. Each new clinician is assigned an age and sex that reflect the distribution seen in recent years. Estimates of total annual new physicians, NPs, and PAs and the characteristics of new graduates came from multiple sources.  

Physicians: Published studies note the number of graduates from programs accredited by the Accreditation Council for Graduate Medical Education (ACGME) and the American Osteopathic Association (AOA). These studies show approximately 1,265 physicians completed residencies in obstetrics and/or gynecology in the 2018-2019 school year and entered the workforce.2 3 The number of physicians being trained and entering the workforce is largely constrained by the number of funded residency positions. The supply projections model the continuation of 1,265 new OB-GYNs annually absent legislation and funding to expand graduate medical education pipeline capacity. Estimates of the proportion of new OB-GYNs who are female and the age distribution of new physicians came from analysis of the 2018 AMA Physician Masterfile. This analysis uses data on physicians who completed their training since 2010 for the sex distribution and since 2000 for the age distribution (Exhibit 24).  

Certified nurse midwives: The American Midwifery Certification Board (AMCB) reports the number of certifications received each year for certified nurse midwives as well as certified midwives. In 2018, they reported 642 nurse midwife certifications and 5 certified midwife certifications.4 Estimates of the percentage of new CNMs who are female and the age at completing their CNM program comes from analysis of the NSSRN. These estimates use data for CNMs who graduated since 2010 for the sex distribution and since 2000 for the age distribution (Exhibit 24). 

Nurse practitioners: The American Association of Colleges of Nursing (AACN) reports that 28,700 new NPs completed their academic programs in 2017-2018. This is an increase from the 23,000 new NPs who completed their academic programs in 2015-2016.5 Analysis of HRSA’s 2018 NSSRN suggests that 4.6% of new NPs practice in women’s health. This is equivalent to about 1,316 new women’s health NPs in 2018. 

The size of the NP training pipeline has grown substantially over the past decade. However, a shortfall of nursing program faculty and constraints on funded clinical training slots could slow growth in the number of NPs being trained each year. Estimates of the percentage of new women’s health NPs who are female and the age at completing their NP program comes from analysis of the NSSRN using data. This analysis examined NPs who graduated from an NP program since 2010 for the sex distribution and since 2000 for the age distribution (Exhibit 24).  

Physician assistants: NCCPA reports 9,287 physician assistants were newly certified in 2018.6 NCCPA surveys new graduates to ascertain the specialty area for PAs who have accepted a position and targeted specialty area for PAs who have not yet accepted a position. Specialty is defined as the specialty of the physician(s) with whom the PA primarily interacts. These survey data suggest that 1.2% of new PAs are practicing in women’s health.7 The number of PAs completing training each year has grown from year to year. This growth trend is projected to continue. Exhibit 24 summarizes the number, age, and sex distribution of new women’s health PA graduates. Characteristics of new graduates are based on analysis of AAPA administrative files.

Modeling Workforce Attrition

HWSM simulates provider retirement probabilities based on provider age, sex, and occupation. Probabilities the provider remains active after age 50 are summarized in Exhibit 25.

Physicians: Attrition probabilities for OB-GYNs are based on self-reported expected retirement age for all surgeons (n=1,116) who participated in AAMC’s 2019 National Sample Survey of Physicians (NSSP). There is an insufficient sample of OB-GYNs to create a retirement pattern using only their responses. To conduct this survey, AAMC contracted with Toluna, an external firm that recruited active physicians from proprietary panels of healthcare professionals. The survey started February 25, 2019 and concluded March 25, 2019 upon reaching the desired quota of 6,000 participants. Survey responses were weighted to be representative of practicing physicians by specialty, age group, sex, and International Medical Graduate status consistent with the 2018 AMA Master File. NSSP findings indicate that physicians intend to retire earlier than projections in previous HRSA reports. We assume that all physicians have retired by age 90. Few remain active past age 75 and many of those older physicians have reduced work hours per week. A 2018 OB-GYN workforce study by Doximity found that OB-GYNs have the highest burn-out rates of all medical specialties except emergency physicians. They also tend to retire earlier than most specialties with a median age of 64.12 Our supply projections for 2030 might be biased high (or low) to the extent that OB-GYNs retire earlier (or later) than the average patterns for surgeons overall. 

Nurse practitioners and certified nurse midwives: Attrition probabilities for women’s health NPs and CNMs are based responses to the following question in HRSA’s 2018 NSSRN:

  • Approximately when do you plan to retire from nursing? 
    • Within a Year 
    • In 1-2 years 
    • In 3-5 years 
    • More than 5 years from now 
    • Undecided 
    • Already retired 

We used the responses “already retired”, “within a year”, and “in 1-2 years” as indications of imminent retirement and created a distribution of NP retirement age for women’s health NPs age 50 to 74. We assume that all NPs have retired by age 75. Because approximately 90% of NPs are female, there is insufficient data to create separate retirement patterns for males. As such, NP supply is modeled using a single retirement pattern. The CNM sample size is insufficient to estimate retirement patterns, so the same retirement patterns as for women’s health NPs are used. 

Physician assistants: Attrition probabilities for women’s health PAs are based on responses to the question in AAPA’s 2015 National Survey of PAs asking whether the PA plans to retire in the next three years. We created a retirement age distribution of PAs age 50 to 74 based on responses to this question. We assume that all PAs have retired by age 75. The sample of older women’s health PAs in this survey is insufficient to obtain reliable estimates of retirement intention by individual age. Retirement patterns used for modeling are based on all PAs—with separate patterns for male and female PAs. 

Modeling Hours Worked

Hours worked in professional activities differ systematically by provider age, sex, and occupation. Changing demographics of the women’s health workforce have implications for future FTE supply. Ordinary Least Squares regression is used to model hours worked patterns using a separate regression for each health occupation. The dependent variable is total hours worked per week. Explanatory variables included provider age group (age <35, 35-44, 45-54, 55-64, 65-74, and 75+ years), sex, and age group by sex interaction term. 

Metropolitan/nonmetropolitan location is included as a dichotomous variable in preliminary analysis. It is omitted from the final regressions as the effect is small and statistically insignificant. Survey data collected in 2018 and 2019 were used to estimate hours worked patterns. 

Physicians: AAMC’s 2019 NSSP asked 1,092 surgeons how many hours they worked during their last typical week of work (excluding any week with leave). Surgeons were used instead of OB-GYNs due to insufficient sample size for OB-GYNs alone. The survey also asked the percent of time spent on types of activities (patient care, teaching, research, administration, other) as well as weeks worked per year. We found little difference in hours worked patterns for male and female physicians or by physician age in weeks worked per year. Therefore, we used weekly hours worked for modeling.  

Nurse practitioners and certified nurse midwives: HRSA’s 2018 NSSRN asked the 1,069 NPs that we categorized in women’s health and 462 CNMs the number of hours worked in a typical week for their primary nursing position. Women’s health NPs worked an average of 36.0 hours per week while CNMs worked 39.8 hours per week. Statistically significant differences in weekly hours worked existed across groups and by sex. The interaction terms by age group and sex were not statistically significant—likely due to the small sample of male NPs. Furthermore, the number of male CNMs is so small that we had to remove sex and age by sex interactions from the OLS regression for that occupation only. CNMs’ hours worked are determined in the analysis entirely by age. 

Physician assistants: AAPA’s 2019 Salary Survey collected data on hours worked per week for the primary employer for 120 PAs who reported a specialty of Obstetrics and Gynecology. Average hours worked per week were 39.9. There is no statistically significant difference in hours worked by PAs in metropolitan versus nonmetropolitan areas. Hours worked declined starting around age 55. Hours were lower for female PAs compared to male PAs. Reported hours for NPs and PAs were lower than for physicians. Part of this difference might be attributed to the NP and PA surveys only collecting hours worked data for the primary employer.  

Modeling Cross-State Migration

HWSM accounts for annual movement across states of OB-GYNs, CNMs, and NPs and PAs in women’s health. This is accomplished in two steps. First, logistic regression estimates the probability of migrating to any other state for the under age 50 population as a function of age group, sex, race (for NPs and PAs only), the population of the state from which the person moved, and a year indicator.  

To model move probability for OB-GYNs, we used National Plan and Provider Enumeration System (NPPES) data for 2019 and 2020. Using multiple years of NPPES data allowed us to identify physicians who changed states by merging the two years using the physician’s national provider identification (NPI). The NPPES data does not contain the provider’s age, so we merged it with Medicare data that had NPI and graduation year. We also assumed that physicians are age 28 at graduation from medical school. This was the average age at graduation for physicians in the AMA Masterfile. We then compared each person’s move probability to a uniform random number between 0 and 1 to simulate whether a physician moves to a new state each year over the projection horizon. 

In addition to provider age, sex, and occupation as explanatory variables for moving, we included OES data on physician salary (in thousands of dollars) by state. This salary was adjusted for state cost of living using data from the Bureau of Economic Analysis. The rationale is that states with higher pay relative to cost of living will attract more physicians than will states with lower relative pay. State population (in millions) is also included as an explanatory variable. The assumption is that states with large populations have more opportunities for intra-state movements and thus fewer inter-state movements. 

The use of NPPES data to analyze physician migration is a new component of the HWSM in 2020. Future areas of improvement include examining the assumptions used to impute physician age and evaluating the optimal explanatory variables in the logistic regression. 

The ACS contains a question asking respondents what state they lived in one year ago. We used move probabilities for all APRNs to model cross-state movement of women’s health NPs and CNMs. We used move probabilities for all PAs to cross-state movement of women’s health PAs.  

As the second step, the likelihood that each person moving will relocate to a specific state is based on the proportion of people moving to that state as observed in NPPES or ACS data. For example, if NPPES shows 5.8% of OB-GYNs who relocated between 2019 and 2020 ended up in Florida, then in the HWSM each OB-GYN who moves has a 5.8% probability of moving to Florida. 

When a women’s health provider moves to a specific state, HWSM then tags that provider by the level of rurality of the area in which s/he practices. The provider’s metropolitan/nonmetropolitan designation is based on the current rurality distribution of the workforce in the state. For example, if in a state 90% of women’s health PAs work in a metropolitan location based on data from the 2018 AMA Masterfile, then each physician assistant moving to that state has a 90% probability of working in a metropolitan area. Due to the small amount of data available for OB-GYNs, we have combined the physician migration regressions and state distribution inputs into physician specialty categories. All OB-GYNs have the same probability to relocate and the same distribution of relocation states. These inputs are created from combined data on all surgeons. 

Modeling Demand

The projected demand for women’s health service providers is derived from the common model outlined in Modeling Demand for Health Care Services and Providers.13 HWSM uses provider staffing patterns to project demand by health occupation and specialty based on demand for health care services. Staffing patterns for OB-GYNs are calculated by dividing current national estimates of the workload driver in each work setting (office, outpatient and emergency visits, and inpatient days) by the estimated FTE providers delivering care in that setting. The portion of OB-GYN patient care time spent in each delivery setting comes from a survey of OB-GYNs.14 These ratios are then applied to projections of future demand for services that assumes the Status Quo in terms of care use and delivery patterns (Exhibit 26).  

Growth in demand for PAs in women’s health services is tied to growth in demand for OB-GYNs under the Status Quo scenario that models the continuation of current care use and delivery patterns. This scenario models the PA-to-OB-GYN staffing ratio remaining constant over the projection period and across states.

  • 13Future modeling might explore adding county-level birth rates to prediction equations for women’s health services.
  • 14Rayburn WF. The obstetrician/gynecologist workforce in the United States: Facts, figures, and implications 2011. In: American Congress of Obstetricians and Gynecologists; 2011.

Demand for CNMs and NPs is tied to total patient demand for women’s health services by setting. National estimates of total FTE providers in each care delivery setting are estimated from the NSSRN. The workload measures were divided by FTE supply in 2018 to calculate staffing ratios by care delivery setting (Exhibit 27). The Status Quo demand scenario models a constant visit-to-provider ratio over the projection period and across states. 

Family Physicians as a Source of Women’s Health Services

Primary care physicians including those in the specialties of family medicine, internal medicine, and pediatric medicine, have been studied by NCHWA and reported on previously.16 17 18 Women’s unique health needs extend well beyond just gender-specific, reproductive health services, and primary care physicians play a critical role in closing the many health disparities that women face. However, primary care physicians also deliver many reproductive health services as well. This can include, among other services, the provision of birth control, cervic.al cancer screening, prenatal care, and the management of chronic conditions during pregnancy. However, within the group of primary care providers, those specializing in family medicine are also uniquely trained to deliver obstetrical care. In line with this competency area, and its overlap with the skills of the obstetricians/gynecologists and certified nurse midwives discussed in this report, the role of family medicine physicians in delivering women’s health services was explored further within the context of this report. 

The analysis combines the 2015 and 2016 National Ambulatory Medical Care Survey (NAMCS) data files (2016 is the latest NAMCS file available) to increase sample size. NAMCS is based on a representative sample of physician office visits. For each participating physician practice, information was collected for a random sample of patient visits obtained through record extraction. Among the information collected for each visit are the type of physician seen, the length of time (in minutes) the physician spent with the patient, and up to five diagnosis codes.

We first analyzed ICD-10 diagnosis and procedure codes (2016 NAMCS) and ICD-9 codes (2015 NAMCS) from office visits to OB-GYNs. This was done to identify codes and screening exams associated with the provision of women’s health services in an office setting.19 We then analyzed visits to family medicine physicians and identified visits where women’s health services were provided. Family physicians provide testing for sexually transmitted diseases to both males and females. For purposes of this analysis, we excluded such testing as specifically a women’s health service. Instead, it was classified such testing as a primary care service. 

NAMCS reports the number of minutes the physician spent visiting with the patient. For patients with multiple diagnoses or exams/screening for the visit, there is insufficient information to know what proportion of physician time is spent providing women’s health services. Consequently, the visit time is pro-rated by dividing (a) the sum total of diagnosis codes and exams during the visit that were identified as women’s health services, by (b) the sum total of all diagnosis codes and exams provided during the visit. This approach implicitly assumes that addressing each diagnosis code and each exam takes approximately the same amount of time. If a 20-minute visit had one diagnosis code, and that code was for women’s health, then 100% of the visit time was assumed for women’s health services. If there were two diagnosis/procedure codes and one was for women’s health services, then we assumed 50% (10 minutes) of the visit time was to provide women’s health services. 

Summing physician time across all patients and using NAMCS sample weights, we calculate that 3.9% of family physician time in direct patient care was spent providing women’s health services. For visits where the patient was a women or adolescent girl age 13 or older, 7% of the visit time was spent providing women’s health services, on average. This includes 6.6% of time spent providing services to women and adolescent girls in metropolitan areas and 9.4% of time during visits in nonmetropolitan areas. 

We calculate that 4,060 FTE family physicians provided women’s health services. This number consists of 3,280 FTEs serving women in metropolitan areas and 780 FTEs in nonmetropolitan areas. The 4,060 figure was calculated using the 3.9% estimate and multiplying it by the 105,400 FTE family physicians identified using the 2018 AMA Masterfile, 

Between 2018 and 2030, demand for OB-GYNs is projected to increase by 5.1% in metropolitan areas and decrease by 6.3% in nonmetropolitan areas. These projections are associated with changing demographics. In metropolitan areas, the number of women and adolescent girls age 13 or older is projected to grow by 13.2 million (11%) between 2018 and 2030; the population in nonmetropolitan areas is projected to decline by 445,000 (2%) during this same period. If we look at the HWSM age group that is most closely aligned with childbirth (women age 18-34 age group), the population in metropolitan areas grows by 2% during this period while the population in nonmetropolitan areas declines by 9%. Applying the 5.1% increase in metropolitan areas and the 6.3% decline in nonmetropolitan areas to the FTE family physician numbers in 2018, we calculate that by 2030 this would be equivalent to 730 FTE family physicians providing women’s health services in nonmetropolitan areas and 3,450 FTEs in metropolitan areas. Given the level of demand for obstetrical care in nonmetropolitan areas and the larger role for family medicine physicians in delivering it, training on obstetrics and for the management of other complex women’s health issues will remain an important component for family medicine residency training programs in the future.20

  • 16 Health Resources & Services Administration. State-Level Projections of Supply and Demand for Primary Care Practitioners: 2013-2025 (PDF - 993 KB). Department of Health and Human Services; 2016. Accessed June 5, 2020.
  • 17Health Resources & Services Administration. Health Workforce Projections: General Pediatricians (PDF - 192 KB). Department of Health and Human Services; 2017. Accessed June 5, 2020.
  • 18Health Resources & Services Administration. National and Regional Projections of Supply and Demand for Primary Care Practitioners: 2013-2025 (PDF - 295 KB). Department of Health and Human Services; 2016. Accessed March 4, 2021.
  • 19ICD-10 codes identified as those for women’s health services are C50-C58 (malignant neoplasms of breast and female genital organs), D24-D28 (benign neoplasms of breast, ovary, female genital organs), E28 (ovarian dysfunction), K62 (diseases of anus and rectum), N39 (urinary system), N60-N64 (diseases of the breast), N70-N98 (noninflammatory diseases of female genital tract / inflammatory diseases of female and pelvic organs), O00-O99 (pregnancy, termination of pregnancy, pregnancy observation, childbirth), R30-R39 (urinary system), Z124 (cervical exam), Z123 (breast screen), and Z30-Z39 (health service related to reproduction). ICD-9 codes are 180-183 (neoplasms of cervix, placenta, uterus, ovary), 217-221 (benign neoplasms of breast, uterus, ovary, other female organs), 630-679 (pregnancy and childbirth), and 760-779 (maternal causes of perinatal morbidity and mortality). Exams/screenings listed as women’s health services are breast exam, pelvic exam, Pap smear, pregnancy test, mammogram, fetal monitoring, and family planning.
  • 20Eden AR, Peterson LE. Challenges Faced by Family Physicians Providing Advanced Maternity Care. Maternal and Child Health Journal. 2018;22(6):932-940. doi:10.1007/s10995-018-2469-2.
Date Last Reviewed: