Child Health and the US Pediatric Subspecialty Workforce: Planning for the Future

All workforce discussions must start with the needs of the patients served. The last 3 decades have witnessed substantive improvements in child health resulting from advances in technology, diagnostic testing, precision medicine, and therapeutic interventions. Advances for rare monogenic disorders, polygenic diseases, and common conditions, such as asthma, are dramatically changing the lives of children with these conditions.¹⁹  Learning collaboratives have vastly improved the dissemination and implementation of evidence-based care for disorders like prematurity, inflammatory bowel disease, cystic fibrosis, and congenital heart disease.²³  These improvements continue to require expertise and collaboration among all members of a patient’s health care team, with a clearly defined structure that incorporates consultation from pediatric subspecialists.

However, changes in the epidemiology of childhood diseases highlight the ongoing need to evaluate the workforce’s capacity to meet children’s health care needs. It is estimated that 2 in 5 children 5 to 17 years of age have at least 1 chronic medical condition.²⁴  Although asthma rates have remained stable, the rates of obesity, hypertension, and type 2 diabetes continue to climb.^25–27  Approximately 1 in 6 US children are now estimated to have a developmental disorder.²⁸  Among adolescents, teen vaping and firearm morbidity and mortality have increased significantly.^29,30  Mental health conditions for children were rising before the COVID-19 pandemic³¹  and have increased alarmingly over the last 3 years, resulting in what the US Surgeon General has called a “mental health pandemic for youth.”³²  With more adolescents and adults self-identifying as sexual- and gender-diverse, the associated rates of attempted suicide were recently reported at 23%, compared with 6% of their heterosexual peers.³³  Children cared for by pediatric subspecialists for acute physical trauma or chronic medical conditions are at a uniquely high risk of coexisting mental health conditions related to factors including their diagnoses, treatment effects, socioeconomic and familial consequences, and reduction in wellbeing and quality of life.^34–37 

Not surprisingly, children’s overall health and the trajectory of any medical condition they acquire may be negatively affected by their exposure to adverse childhood experiences, racism, poverty, environmental pollutants and complications of climate change, and the presence of coexisting medical or developmental disabilities.^19,38–41  The national poverty rate for children 0 to 17 years of age in 2021 was 16.9% (range by state: 8.1% to 27.7%), 4.2 percentage points higher than the national average and almost 7 percentage points higher than for adults >65 years of age (10.3%).⁴²  Sociodemographic data indicate that >50% of US children are from a racial or ethnic minority, and >1 in 4 children are growing up in immigrant families.⁴³  Children from minoritized populations are more likely to live in poverty and have parents who are not securely employed, exacerbating health disparities.⁴¹  In addition, many new therapies, such as the cystic fibrosis transmembrane regulator modulators, have been predominantly tested in non-Hispanic white populations and may be less effective in children from racial or ethnic minorities without additional research, increasing disparities in health outcomes.⁴⁴ 

Geography also impacts child health. The US child population increased from 64.2 million (1990) to 73.6 million (2021) and is anticipated to grow to 78.2 million by 2050. The South and West are responsible for much of that growth, with 3 states (Texas, Florida, California) accounting for one-half of it.⁴⁵  Rural areas, common in those regions, reveal higher rates of children living in families with incomes less than half of the federal poverty level (12.2%) compared with urban areas (9.2%)⁴⁶  and reveal increasing gaps in access to both physical and mental health care.⁴⁷  A significant and increasing percentage of children (>50% of children as of November 2022) are insured through Medicaid or the Children’s Health Insurance Program.⁴⁸  Medicaid coverage varies substantially by state,⁴⁹  and reimbursement rates fall well below those of Medicare and commercial insurance,^50,51  further exacerbating access to care. It is not known how the end of the COVID-19 public health emergency will impact Medicaid coverage rates for children and families; however, some states are revealing a decline, particularly in those that do not prioritize renewals or the continuous enrollment provision added during the COVID-19 pandemic.⁵² 

Pediatric subspecialists are specifically trained to care for children with acute and chronic specialized health needs. Most pediatric subspecialists in the United States first complete a 3-year general pediatrics residency before completing a 3-year subspecialty fellowship. Exceptions are adolescent medicine and pediatric emergency medicine, in which some physicians enter fellowship after a non-pediatric residency, and pediatric hospital medicine, which currently requires a 2-year fellowship. Training program requirements for pediatric subspecialty fellowships are set by the Accreditation Council of Graduate Medical Education (ACGME), which allows for increases in the number of training positions by institutional graduate medical education (GME) leadership with approval by the ACGME Review Committee for Pediatrics. The ABP sets the criteria to determine which individual subspecialty fellowship graduates are eligible to sit for the ABP’s board examinations.

As of June 2023, the total number of pediatric subspecialists ever certified by the ABP was 38 126, of whom 29 359 were actively enrolled in MOC.⁵³  Previous workforce reports have largely focused on pediatric subspecialists as a single entity; however, fundamental differences exist. Pediatric subspecialties may be defined by their patients’ age (eg, neonatal–perinatal medicine), the setting in which they practice (eg, pediatric emergency medicine), or in-depth knowledge of a specific body system (eg, pediatric endocrinology). They may be primarily hospital-based in inpatient units (eg, pediatric critical care medicine), primarily outpatient-based (eg, developmental–behavioral pediatrics), or some combination (eg, pediatric cardiology).

Publications using ABP data have highlighted the variability by subspecialty regarding total numbers, demographics (ie, age, sex, race, and ethnicity), training history (ie, Doctor of Medicine (MD), Doctor of Osteopathy, international degree, American or international medical school location, educational debt), and work characteristics (hours worked, full-time or part-time status, primary work setting, and effort allocation across clinical care, education, research, and administration).^8,11,54,55  Differences across board-certified subspecialists were recently reaffirmed by using available ABP data through June 2023 for subspecialists ≤70 years of age (Tables 1 and 2). Particularly striking are differences across subspecialties in total number; the proportion of women, individuals from underrepresented minority groups, international medical graduates, and full-time or part-time staff; and reported clinical and research effort allocation.

When the workforce is limited to the United States, geographic maldistribution is frequent. Unlike adult subspecialists, most pediatric subspecialists are concentrated in academic medical centers in urban settings, except for neonatologists and emergency medicine subspecialists providing care in hospital-based community programs. Publicly available ABP data dashboards²¹  and Table 1 indicate a marked variation in the ratio of subspecialists to children 0 to 17 years of age and in the average driving distance to a subspecialist, by both state and subspecialty.

Most US pediatric fellowship training programs are in urban areas, closely mirroring the geographic distribution of pediatric subspecialists. The overall number of first-year pediatric subspecialty fellows in fellowship programs has increased from 1145 (2008) to 1692 (2022). However, increases have been limited to selected subspecialties (eg, pediatric critical care medicine) with negligible increases, if any, among some of the smaller subspecialties (eg, pediatric infectious diseases).⁵⁶  Differences in US fellowship programs and the demographic, financial, and training characteristics of US fellows in academic year 2021 to 2022 are summarized in Table 3.

The financing of general GME in the United States is complex. Primary funding comes from the federal government via Medicare GME payments to support direct and indirect medical expenses in large general hospital systems. Fellowship training programs in freestanding children’s hospitals that do not receive Medicare GME dollars are eligible for funding from the Children’s Hospital Graduate Medical Education (CHGME) payment program. However, although Medicare GME dollars are tied to overall Medicare spending and increase every year, CHGME dollars must be reauthorized by Congress annually, are currently capped at $375 million, and are reimbursed at ∼51% of Medicare GME support. Other smaller sources of funding for fellowships include Medicaid GME payments, primary care training programs, health systems, training grants, international stipends, moonlighting opportunities, and philanthropic contributions.⁵⁷  These factors lead to financial instability for fellowship programs and a heightened preference for those programs that generate revenue for hospital systems.^57,58  It is anticipated that financial demands on health systems will continue to rise as training program positions increase because of (1) newly developed subspecialties (eg, hospital medicine, now the third-largest subspecialty) and (2) the recruitment of additional trainees due to changes in work hours, increased clinical demands as more children survive and thrive with chronic medical conditions, and concerns related to burnout and stress within training programs and among practicing subspecialists.

Economic analyses have revealed that adult subspecialists earn close to 25% ($1.2 million) more over their career than pediatric subspecialists.^13,59  Differences are also present among the pediatric subspecialties, although differences vary depending on the data source used.⁶⁰  Two subspecialties (pediatric cardiology and neonatal–perinatal medicine) consistently reveal higher potential than the remaining pediatric subspecialties.^12,15,60,61  These disparities have widened in recent years and compound the already lower lifetime earning potential among pediatric subspecialists compared with their adult counterparts.¹³  Higher lifetime earning potential has been associated with higher mean fellowship fill rates.⁶² 

Medical workforce models are often employed to identify current staffing gaps and predict future staffing and training needs. Most workforce models focus on the supply of physicians per capita, given the complexities of measuring population health needs and market demand.³ 

A major goal for the authors of this supplement was to address weaknesses of previous pediatric subspecialty models by (1) examining individual pediatric subspecialties at national and subnational levels, (2) accounting for historic and current demographic, work characteristics, and geographic distribution for each subspecialty, (3) incorporating the projected growth and geographic distribution of pediatric subspecialty fellows, (4) projecting absolute numbers or headcount (HC) and clinically available subspecialists, and (5) accounting for the growth and geographic distribution of US children 0 to 18 years of age, as forecasted by the US Census Bureau.⁶³  Although pediatric subspecialists often see children through young adulthood, we elected to use 18 years of age because that is when children move toward greater legal independence. The baseline model incorporates the number of subspecialists ≤70 years of age in the current workforce, the supply of incoming fellows (held constant at 2019 levels), exits from the workforce, the addition of subspecialists who have worked outside the United States and return to the United States to provide clinical care, and worker diffusion from training to subsequent geographic locations. Projections are provided for HC and clinical workforce equivalent, defined as HC adjusted for the self-reported proportion of time spent in clinical and consultative care, not including time on call when not providing clinical care. This term was used because of ongoing debates regarding how clinical full-time equivalents are calculated.⁶⁴  An online, interactive data visualization tool with analyses at the national, census region, and census division levels permits deliberations regarding potential geographic maldistribution.⁶⁵ 

The model also incorporates possible future scenarios based on emerging trends or potential “wild cards”⁶⁶  that may impact the other 3 domains in Fig 1 (Table 4). Ten scenarios were identified through a literature review and stakeholder engagement and were operationalized for inclusion in the model. Three overarching trends were incorporated into the scenarios and are discussed below, specifically (1) an increase or decrease in the number of fellows, (2) an increase or decrease in effective clinical time, and (3) increased attrition. Limited approaches are currently available to measure child health needs and market demand regarding subspecialty care;⁶⁷  several potential impacts of the environment on these 2 factors are captured in the model’s scenarios.

The supply of fellows is primarily related to career decisions to enter pediatric subspecialty medical training, which are often decided during medical school or the first year of a general pediatrics residency.⁶⁸  Concerns have been raised regarding the overall interest level in pediatrics, particularly considering its markedly lower salaries compared with other specialties, rising debt among medical trainees, and the overall increase in the number of unfilled pediatric residency slots over the last decade.⁶⁹  The number of trainees entering residency could significantly impact supply unless attention is paid to factors affecting career choices from college through fellowship.

An important group of pediatric subspecialties, including adolescent medicine, child abuse pediatrics, developmental–behavioral pediatrics, pediatric endocrinology, pediatric nephrology, and pediatric pulmonary, have struggled to fill their training positions in the National Resident Matching Program. They have averaged 30% to 50% fewer applicants than positions, a trend that has been consistent for the past 10 years. In comparison, programs such as pediatric cardiology, pediatric critical care medicine, and pediatric emergency medicine routinely fill >95% of available positions and may have 20 to 50 more applicants than available positions. Data regarding future earning potential¹⁵  or job availability¹⁶  may alter career decisions for current and future residents. Other factors, such as debt burden, length of subspecialty training, prestige, exposure during medical school and residency, and mentorships collectively may impact the number of fellows in the pediatric subspecialties.

Several scenarios explore differences in clinical time. Available clinical time might decrease because of supply factors, including more subspecialists seeking work–life integration. Market factors (eg, closing of pediatric units and beds for financial reasons) could worsen access for children and place a greater clinical burden on individual subspecialists in tertiary centers, also decreasing available clinical time. Greater child health needs due to environmental factors (eg, increasing poverty, racism, pollution), lack of insurance, pandemics, and global conflicts could also effectively decrease available clinical time by increasing patient needs.

Alternatively, a robust commitment to insurance coverage for children and families and advances in medical technology might decrease health care needs, leading to increased available clinical time. Available clinical time also might increase because of decreased financial support for nonclinical activities (eg, quality improvement, research, medical education) that rely heavily on clinical revenue. The development of new care models (eg, greater management of common subspecialty conditions in pediatric primary care, the implementation of team-based care models with concurrent, robust advanced practice provider training pathways, telehealth) or health care applications of artificial intelligence might positively or negatively affect clinical time. Federal or state agencies might incentivize subspecialty training through loan repayment programs or health system changes to address access issues (eg, wait times) regardless of cost, increasing supply, and available clinical time.

Given the high rates of international medical graduates in some subspecialties and their role in caring for children in underserved areas,⁷⁰  US immigration policies could either increase or decrease clinical care capacity.

Several of the scenarios examine early retirement, early midcareer exits, or career shifts within pediatrics or health care in general and their impact on overall supply. Attrition might be in response to environmental factors like the socioemotional or physical effects of a national crisis (eg, war, climate events, pandemics) on supply and its impacts. Indeed, physician workforce trends due to the COVID-19 pandemic are being closely tracked, given nursing staffing reductions during and after the pandemic.⁷¹  Others have questioned the impact of burnout in the current workforce and among fellows from administrative burdens, perceived lack of support, stigma related to help-seeking, and racism and discrimination in academic medicine.^72–75  An increase in nonclinical professional opportunities (eg, consulting, research) might also lead to attrition from clinical care.

The model addresses the limitations of many previous modeling efforts by its use of detailed data on the current and future supply, corrections for child population growth, and attention to geographic distribution and location at the subnational level. Although the model capitalizes on the ABP’s data on certified subspecialists, some decisions were made that may limit the conclusions that can be drawn. For the sake of simplicity, the model assumes the same scope and productivity within a subspecialty, albeit based on ABP data regarding the average hours worked and percentage of clinical time specific to that subspecialty. It includes only ABP-certified pediatric subspecialists; other physicians caring for children (eg, pediatric neurologists, geneticists, allergy/immunologists, child psychiatrists, adult physicians) are not included because data on hours worked and time in pediatric subspecialty clinical care were not consistently available.⁷⁶  The model also excludes advanced practice providers, despite the critical role of team-based care in addressing children’s health care needs. Limited data exist specific to these providers; available data suggest relatively slow growth in the nonphysician workforce interested in pediatric care without substantive policy change.^77–80  The model does not provide estimates at the state or county level because of concerns regarding small numbers for many subspecialties; the reader is referred to the ABP’s state- and county-level data on the counts of subspecialists.⁸¹  The model also focuses only on clinical care and does not address the supply of pediatric subspecialists who lead medical education, quality improvement, research workforce efforts, and scientific discovery critical to improved health outcomes in children. Last, the model does not specifically account for how child health needs unique to a subspecialty may change over time because of environmental factors, such as technological and medical advances, public expectations, state or federal policy decisions, and societal changes, or how changes in market demand could affect pediatric subspecialty care.

This supplement includes an overarching methodology article followed by 15 separate articles focused on the current ABP-certified subspecialties. Each article begins with an overview of the current and anticipated child health needs specific to the subspecialty and a description of the current workforce. The results of the model projecting the US pediatric subspecialty workforce supply ≤70 years of age through 2040 for 14 of the subspecialties are then reviewed, followed by implications across education and training, clinical practice, policy, and future workforce research. Recognizing the uniqueness of the different subspecialties, the authors for each subspecialty were asked to bring their awareness of shifting trends in child health needs in their discipline and their expert perspectives on the future pediatric subspecialty workforce to their deliberations. Given the recent approval of pediatric hospital medicine as a subspecialty and the lack of sufficient historical data for inclusion in the model, this subspecialty’s article focuses solely on extant data.

The 15 subspecialties’ articles are followed by an article examining the diversity of US children in comparison with the pediatric subspecialty workforce. The supplement closes with an epilogue, summarizing core themes across the articles and offering recommendations.⁸²  All articles in the supplement include the term “pathway” as opposed to “pipeline,” acknowledging the lack of a single pathway into pediatric subspecialty care and potential insensitivities to US indigenous peoples.^83,84 

With the proportion of adults >65 years of age in the United States anticipated to surpass children in 2034,⁸⁵  attention to child health will suffer without a more holistic, lifespan-focused approach. Specifically, the United States must adopt an ongoing public commitment to meeting child health needs if our shared national goal is to improve outcomes inclusive of all children, who represent our country’s future workforce.⁸⁶  Efforts are needed to integrate up-to-date data on all 4 domains outlined in Fig 1: (1) child health needs, (2) market demands, (3) the supply of pediatric subspecialists, and (4) trends in environmental factors. These efforts should parallel the creation of services-based workforce models that integrate the contributions of the myriad other members of subspecialty care teams beyond pediatric subspecialists.

The supplement authors hope that this work joins other current national efforts to catalyze a renewed focus on the health of children cared for by pediatric subspecialty care teams, including the recently published NASEM Consensus Study report on the pediatric subspecialty workforce,²⁰  the Association of Medical School Pediatrics Department Chairs Workforce 2025 initiative,⁸⁷  and the current NASEM Consensus Study, “Improving the Health and Wellbeing of Children and Youth Through Health Care System Transformation.”⁸⁸  Together, these efforts challenge pediatricians and the nation at large to proactively and intentionally consider how to best provide for the health care needs of all US children, now and in the future.

The authors thank Virginia A. Moyer and Patience Leino for their editorial support. In addition, we thank the pediatricians who shared their information with the American Board of Pediatrics Foundation and made this supplement possible.

ABP

American Board of Pediatrics

ACGME

Accreditation Council of Graduate Medical Education

CHGME

Children’s Hospital Graduate Medical Education

COVID-19

coronavirus disease 2019

HC

headcount

GME

graduate medical education

MD

Doctor of Medicine

MOC

maintenance of certification

NASEM

National Academies of Sciences, Engineering, and Medicine