Child Health Needs and the Pediatric Critical Care Medicine Workforce: 2020–2040

PCCM subspecialists care for a wide range of seriously and critically ill children in various clinical contexts.⁸  Expert consensus has established criteria for diseases in the domain of PCCM, such as pediatric acute respiratory distress syndrome,⁹  sepsis,¹⁰  and severe traumatic brain injury.¹¹  Estimates of the annual volume of children admitted to a US PICU are limited and have ranged from 230 000 (2001)¹²  to 315 000 (2005).^8,13  Cross-sectional survey data from 2004 indicated that there was a median of 58 admissions per US PICU bed,¹⁴  whereas survey data in 2016 indicated that the number of PICU beds had increased to 5908.¹⁵  This suggests that annual PICU admissions may now exceed 340 000, though the precise number of PICUs is unknown. Thresholds for PICU admission for these and other diseases may vary according to organ support modalities available within or outside the PICU setting. For example, some institutions provide continuous, nebulized albuterol in the non-ICU, whereas other centers confine such therapy to the PICU.¹⁶  Accordingly, there is some variation in the types of patients cared for in different PICUs. Individual PICU case mix also varies according to whether the provided level of care is community, tertiary, or quaternary. Quaternary PICUs are commonly located in urban settings, but serve large catchment areas, and are capable of providing highly advanced, comprehensive care to complex patients.¹⁷  In addition to medical–surgical PICUs in large academic centers, specialized PICUs, such as neurologic or cardiac PICUs, fall under the umbrella of quaternary PICUs. Tertiary PICUs also provide comprehensive services to complex patients but often do not offer some of the specialized services of quaternary units. Community medical center PICUs provide a range of services that can vary substantially.

As in other areas of medicine, patient complexity has increased over time in the PICU as a result of advances in technologies and improved survival.^18–21  A 2012 assessment of the Virtual PICU Systems database revealed that 53% of children admitted to PICUs had complex chronic conditions with an associated increased risk of mortality and prolonged length of stay.²²  A prospective observational study conducted by the Collaborative Pediatric Critical Care Research Network between 2011 and 2013 characterized a random sample of 10 078 children admitted to PICUs in 7 large US children’s hospitals.²³  In that sample, the median age was 3.7 years (interquartile range 0.8–10.8) and 53.8% had government insurance. Reasons for admission included respiratory dysfunction (33.5%), cardiovascular (24.1%), other (22.3%), and neurologic (20.1%) dysfunction. A more recent examination of PICU epidemiology from 2001 to 2019 reported an increase in the prevalence of patients with multiple organ dysfunction (16.4% in 2001–23.5% in 2019), a decline in mortality (2.5%–1.8%), an increase in hospital length of stay by 0.96 days, and a doubling of cost for admissions involving ICU care.²¹ 

In addition, the role of social drivers of health in influencing PICU admissions and outcomes is increasingly recognized. Race, ethnicity, and socioeconomic disparities have been described for cardiac arrest, asthma, trauma, congenital heart disease, and sepsis, as has the underenrollment of minoritized groups in PICU clinical research.^24–26  A cross-sectional study of 773 743 children hospitalized between 2020 and 2021 revealed that children from neighborhoods associated with lower socioeconomic status were more likely to require critical care and not survive hospitalization.²⁷ 

PCCM subspecialists also commonly care for patients outside of the PICU. For example, PCCM subspecialists with additional training in cardiac intensive care or cardiology commonly care for children with congenital or acquired heart disease in specialized PICUs.²⁸  PCCM subspecialists also commonly staff sedation services for inpatient and outpatient procedures in the absence of critical illness. Increasingly, PCCM subspecialists oversee rapid response teams evaluating patients in the nonintensive care inpatient setting.²⁹ 

PCCM arose in response to a need to provide high-intensity organ support previously unavailable outside of the operating theater and was propelled by early advancements in the management of respiratory failure in neonates. The first PICU opened in Göteborg, Sweden, in 1955, and the first US PICUs opened in Washington, District of Columbia, and Philadelphia in the 1960s. In 1970, Peter Holbrook became the first PCCM fellow trainee, entering a training program developed by Peter Safar in Pittsburgh.³⁰  Fellowship programs sprouted up in the following decade at major children’s hospitals across North America. The first American Board of Pediatrics (ABP) certifying examination for PCCM was offered in 1987, which certified 182 subspecialists. The Accreditation Council on Graduate Medical Education accredited the first PCCM training programs in 1990.^1,30 

On the basis of ABP data through June 2023, 3689 pediatricians have been board-certified in PCCM, 84.4% (3128) of whom were actively enrolled in Maintenance of Certification.³¹  Thirteen percent of certified PCCM physicians have also been certified in another ABP subspecialty, most commonly cardiology (5.4%), pulmonology (1.9%), and palliative care (1.9%).³²  These figures do not account for dual certification in non-ABP subspecialties, such as anesthesiology, internal medicine, and related subspecialties, or the growing field of clinical informatics. Pediatric intensivists who are board-certified in both PCCM and pediatric cardiology constituted 20% of the pediatric cardiac intensive care workforce per a 2019 survey.²⁸ 

The ABP data on certified subspecialists include individuals who may not be in the workforce because of recent retirement, death, or other factors. Consequently, descriptions of the workforce below limit the sample to certified pediatricians aged ≤70 years. Of 3076 PCCM pediatricians in 2023, 48.6% identified as female and 51.4% as male. The median age was 46 years and 12.8% were aged 61 to 70 years, highlighting that over one-tenth of PCCM subspecialists may retire over the next 1 to 2 decades. Notably, 48.6% of PCCM physicians aged ≤70 years and 63% of PCCM physicians ≤40 years of age are female.³¹  Regarding medical training, 68.9% were American medical graduates (AMGs) with a Doctor of Medicine (MD) degree, 5.5% were AMGs with a Doctor of Osteopathy (DO) degree, 14.1% were international medical graduates (IMGs) with an MD degree, and 11.5% were IMGs with an international degree.³¹  The proportion of PCCM physicians with DO degrees has increased over the past decade. Race and ethnicity estimates from 2018 to 2022 indicate that 12% self-identified as underrepresented in medicine, which includes Black or African American; Hispanic, Latino, or Spanish; American Indian or Native Alaskan; or Native Hawaiian or Pacific Islander origin.³³ 

Data on the work characteristics of current ABP-certified PCCM subspecialists are collected through the ABP’s Maintenance of Certification enrollment surveys.³⁴  From 2018 to 2022, the PCCM response rate was 57.1%, providing data from 1356 eligible subspecialists ≤70 years of age. The majority (92.9%) reported being employed full time, and 61.9% reported working ≥50 hours per week on average over the last 6 months, exclusive of time on call from home. Women were more likely to indicate part-time employment status (9.6%) compared with men (3.4%). Most (78.1%) spent ≥50% of their time in clinical care, whereas 6.1% reported spending ≥50% time on research. About one-third (31.2%) endorsed that their primary work setting was within a medical school or parent university, with most (83.1%) having a faculty appointment. The majority (81.9%) reported a primary work setting within an urban environment. Approximately 40.2% reported that ≥50% of their patients received public insurance.

When the workforce is limited to the United States in 2023, there was an average of 57.8 currently certified PCCM subspecialists per US state (range 0–340), which translates to 4.1 PCCM subspecialists per 100 000 children aged 0 to 17 years (range 0–19.1) (Fig 1).³⁵  There was wide variability in the distribution of PCCM subspecialists within states, with most concentrated in urban settings and few in rural areas. Analyses from 2019 showed the average driving distance to a certified PCCM subspecialist was 19.0 miles; distances ranged from a low of 5.8 miles in New Jersey to a high of 149.4 miles in Wyoming (excluding Alaska, Hawaii, Puerto Rico, and the District of Columbia).³⁶ 

With respect to fellowship training programs, Accreditation Council on Graduate Medical Education data show an increase of 24.2% (from 62 to 77) in the number of accredited US PCCM training programs between academic years 2012 to 2013 and 2021 to 2022. National Resident Matching Program (NRMP) data since 2014 indicate a gradual decline in the number of unmatched PCCM training spots, with 13 open spots in 2014 and none in 2022. However, the ABP count of annual trainees exceeded the number of NRMP positions each year over that time period, resulting in an average final fill rate of 103.8% and indicating that some positions go to trainees outside of the NRMP match.³⁷  The ABP data demonstrate that the number of first-year PCCM fellows increased from 156 in 2012 to 216 in 2022 (+38.5%).³⁸ 

Among the 617 PCCM fellows currently in training levels 1 to 3 in standard, noncombined US fellowship programs during academic year 2022 to 2023, 65.2% identified as female, 34.7% male, and 13% as underrepresented in medicine. Sixty-eight percent were AMGs with an MD degree,12.8% were AMGs with a DO degree, 0.8% were AMGs with an unknown degree, 11.5% were IMGs with an MD degree, and 6.8% were IMGs with an international degree. This distribution of degrees is comparable to the current workforce of PCCM subspecialists.³¹  Fellows commonly take a first position posttraining near their training location; just over half (56%) of trainees stay in the state in which they trained.³⁹ Figure 1 shows PCCM fellowship locations in academic year 2021 to 2022.

There are increasing concerns that the PCCM subspeciality may become oversubscribed, with more PCCM subspecialists relative to available positions. An examination of advertised PCCM job opportunities between 2019 and 2020 indicated a relatively low number of available positions compared with the anticipated number of PCCM fellow graduates seeking employment.⁴⁰  This lower number of job opportunities has coincided with a substantial decrease between 2019 and 2021 in PCCM fellows reporting they agree or strongly agree that there are available jobs in academic centers nationwide (56% agree/strongly agree in 2019 vs 26% in 2021), available jobs in nonacademic centers or private practices (87% vs 64%), available jobs in a desired location (44% vs 29%), and confidence that a job will be available in PCCM (79% vs 51%).⁴¹ 

Although reports suggest compensation has low importance for residents in choosing a pediatric subspecialty career,^42,43  the increasingly high rates of debt may impact decision-making. On the basis of the ABP’s Subspecialty In-Training Exam Survey in 2022, ∼45.8% of current PCCM fellows owe $200 000 or more compared with 39.5% for all pediatric subspecialty fellows (personal communication, ABP, February 20, 2023). PCCM is 1 of 3 ABP subspecialties reported to have a favorable financial return on time invested in additional training when average compensation is compared with that of general pediatrics.⁴⁴ 

The workforce forecasting model, an ABP Foundation-led effort to characterize the future supply of pediatric subspecialists aged ≤70 years in the United States, incorporates data from the ABP and other national resources. Baseline workforce estimates from the model are provided by subspecialty at the national and subnational level from 2020 to 2040. Model estimates represent aggregations of microsimulations at the individual level that use historic pediatric subspecialty data to project future trends. Workforce projections are expressed in either headcount (HC) or clinical workforce equivalent (CWE), which is HC adjusted by the proportion of time spent in direct clinical or consultative care. Each parameter can be viewed as an absolute number or per 100 000 children aged 0 to 18 years. The model also accounts for changes in the child population at the national and subnational level on the basis of the US Census Bureau⁴⁵ ; differences by subspecialty for census regions are discussed in the summary article in this supplement.⁴⁶  In addition to the baseline projection, the model contains alternative scenarios that adjust current conditions by expanding or contracting the trainee pool, increasing or decreasing clinical time, early exit for some or all of the workforce, or 2 scenarios that combine several of these factors together.

All results are accessible in a publicly available, online interactive data visualization tool.⁴⁷ 

Results from the baseline model and several of the scenarios at the national and census region levels are discussed in more detail below. The numbers reported below may differ from those in the previous section because of differences in years (2020 vs 2023), sample selection criteria, and inclusion of self-reported clinical time. Projections at the census division level and 95% confidence intervals are also available (Tables 1–2; online interactive tool).

Under baseline conditions in the United States in 2020, there were an estimated 2397 PPCM physicians in total HC or 1.84 CWE per 100 000 children. This estimate is projected to increase to 4906 HC and 3.28 CWE per 100 000 children by 2040, representing a 105% growth in HC and a 78% growth in CWE. PCCM constitutes 10.3% of the total pediatric subspecialist workforce in 2020 and 12.5% in 2040.

The scenarios vary the projections at the national level (Fig 2). One alternative scenario decreases the fellow training pool by 12.5%; in this scenario, the PCCM clinical workforce growth would decrease from 78% to 71% (−7%) or 3.14 CWE per 100 000 children. Increasing the pool of fellow trainees by 12.5% would result in greater projected growth of the PCCM supply, achieving 3.42 CWE per 100 000 children in 2040 or 86% growth in the clinical workforce (+8%). The scenarios with the most significant effect on the projected supply of PCCM subspecialists are those involving multifactorial changes to the number of intensivists entering and exiting the workforce. A combined 12.5% decrease in the pool of fellow trainees, plus an overall reduction in clinical time, and greater early exit from the field results in a projected smaller increase of the workforce of 59%, or 2.92 CWE per 100 000 children (−21%). This rate of growth is the lowest among the models considered. Conversely, an increase in the number of fellows, an increase of clinical time, and no change in the number of intensivists leaving the field before retirement results in a projected 99% increase in the workforce to 3.66 CWE per 100 000 children (+21%).

Growth in the PCCM workforce (Fig 3) is projected to be most pronounced in the Midwest census region, increasing from 1.93 to 3.7 CWE per 100 000 children (+91%). The lowest growth is projected in the West census region, increasing from 1.75 to 2.79 CWE per 100 000 children (+59%). These growth projections do not parallel predicted population growth patterns; although the model projects an overall increase in supply and incorporates child population growth, geographic areas may not experience the increase in supply equally. For example, the Midwest is projected to experience the largest proportion of PCCM growth by 2040, but this region’s population of children aged 0 to 18 years is only expected to grow 2% by 2040 as compared with the Northeast (4%), the South (19%), and the West (23%).

The impact of the multifactorial decrease in the workforce varies by geography, resulting in projections 18% (West) to 21% (South) lower compared with the baseline scenario. The multifactorial increase in workforce scenario results in projections 18% (West) to 23% (Midwest). Tables 1–2 display the workforce scenarios projections by US Census division for HC and CWE per 100 000 children and provide additional detail.

PCCM is a relatively young subspecialty, though its workforce has already changed in remarkable ways. Training positions for PCCM have increased substantially since the field’s inception and the number has continued to grow. PCCM has been, and appears to remain, a relatively popular subspecialty choice among pediatric residents, with programs consistently filled. Such growth is evident in the workforce model, which projects a greater proportional increase for the PCCM workforce by 2040 compared with the lowest growth subspecialties (adolescent medicine, child abuse pediatrics, developmental–behavioral pediatrics, neonatal–perinatal medicine, nephrology, and pediatric infectious diseases). The current work raises important questions about the future of training pathways for PCCM subspecialists, the environment in which PCCM subspecialists practice, policies governing the PCCM workforce, and areas in need of additional data.

As of 2022, there are 77 PCCM training programs and the number of annual applicants has more than doubled over the past 2 decades. These trends occurred despite no change in the US birth rate, without guidance from a governing agency, and despite a major economic recession.⁴⁸  Cardiac intensive care has emerged as the most common PCCM subspecialty training niche; some workforce projections estimate that the cardiac intensive care job market will be saturated by ∼2027.²⁸  Notably, no formalized pathway exists for training pediatric cardiac intensivists, a matter that has recently been the focus of a dedicated task force.⁴⁹  Training programs for both PCCM and the subspecialty’s niches have emerged organically, largely related to apparent market demands and staffing needs at quaternary centers. The workforce projection models indicate that an alteration of ±12.5% of the pool of PCCM trainees has a relatively small impact on the overall growth of the workforce. Few data are available for expected clinical need or market demand for the broader pool of PCCM subspecialists, nor are rigorous data on anticipated workforce demand available for newly arising niches, such as neurocritical care.⁵⁰ 

Historic growth in the PCCM workforce appears to correlate with growth in PICU beds in the United States and changing staffing of PICUs. The results of a survey of medical directors and nurse managers indicated that between 2001 and 2016, the number of hospitals with PICUs in the United States decreased slightly from 347 to 344 (−0.9%), whereas PICU beds increased from 4135 to 5908 (+43%).¹⁵  A retrospective study of American Hospital Association data indicated that PICU beds increased 16% in the United States between 2008 and 2018, but also suggested a decelerating PICU bed growth trend in the United States.⁵¹  During approximately this same period, there was an increasing prevalence of 24/7 in-house PICU attending coverage.^52–54  In 2012, 35% of surveyed PCCM physicians reported that the number of PCCM physicians in their area was inadequate, and 46% reported the number of PCCM fellows was inadequate,^55,56  though it is not clear that this is true today. In recent years, there has been a growing presence in the PCCM workforce of pediatricians who have not completed subspecialty training (eg, PICU hospitalists) and APPs, though the use of these alternative workers is not well quantified among the entire PCCM workforce.⁵⁷  The scope of the PCCM field has also grown to encompass aspects of patient care outside of the PICU, such as overseeing procedural sedation,⁵⁸  interfacility transport,⁵⁹  overseeing in-hospital emergency response systems, and other administrative responsibilities.

The most substantial changes in the projected PCCM subspecialist workforce are exhibited by multifactorial changes to the influx and efflux of subspecialists from the clinical environment. Uncertainty about how the practice environment of PCCM will change in the next 2 decades raises questions about future PCCM workforce requirements, particularly whether demand will rise proportionately to the expected increase in PCCM subspecialist supply. APPs are increasingly used in PICUs in the United States to supplement the subspecialist workforce while potentially lowering costs, and initial data suggest such a model might result in similar or possibly improved outcomes compared with physician-only staffing.⁵⁷  Data are needed to characterize how APPs are being incorporated into the PCCM workforce, and how this will, in turn, impact the workforce of PCCM subspecialists. It is also unclear whether the number of PICU beds in the United States will continue to increase, a factor that would strongly influence the need for frontline clinicians. Other ABP subspecialties have noted signs of job market saturation, leading to new fellowship graduates accepting positions not aligned with their personal preferences or career goals or pursuing further training in a subspecialty niche.⁶⁰  Burnout has also long been recognized as prevalent in the field of PCCM.^61,62  Whether increased attention to burnout, moral distress, and moral injury will alter the influx and efflux of PCCM subspecialists from the workforce remains a key question.

Recent American Academy of Pediatrics/Society of Critical Care Medicine guidelines provide recommendations for the provision of pediatric critical care and definitions for types of PICUs,¹⁷  indicating that children admitted to a PICU should be cared for by a pediatric intensivist who is subspecialty fellowship trained. Few other policies provide guidance on PCCM workforce design. There are no policies or guidelines on what constitutes a full-time equivalent, appropriate PCCM subspecialist-to-patient ratio, staffing structures for PCCM subspecialists supervising APPs, or night coverage expectations. Moreover, there are no widely accepted policies in PCCM guiding the duration of leave for life events such as the birth of a child or standard approaches for screening for and treating practice-related mental health conditions among clinicians. These are sorely needed.

This projection of the PCCM subspecialty workforce over the coming decades overcomes a limitation of previous ABP assessments because it considers the workforce supply beyond a “snapshot” in time. Although the present analysis may raise more questions than answers, these findings can support additional work to help avoid the “workforce pendulum” effect, in which a field’s workforce is repeatedly recalibrated in opposing directions.⁶³  Additional data on PCCM training pathways, practice environment, number of PICUs within geographic regions, the growing PCCM APP workforce, and governing policies are needed to better forecast the PCCM workforce supply and demand, while allowing for just-in-time modifications to avoid workforce imbalances.

The United States has achieved tremendous advances in the survival and well-being of children suffering from critical illness. There has been enormous growth and change in the PCCM physician workforce, which is projected to expand more than any other ABP subspecialty over the next 2 decades. This growth does not appear to be matched by indicators of a corresponding increase in PCCM subspecialist demand, because the increase in US PICU beds appears to be slowing, many large programs have completed the transition to 24/7 in-house staffing, and trainees express a lack of optimism regarding job opportunities. Even models with multifactorial changes that should reduce growth in PCCM subspecialist supply still showed an increase in supply over the next 20 years. There is substantial uncertainty regarding whether the volume of pediatricians being trained in PCCM is appropriate to match clinical need and market demand; an oversupply of PCCM subspecialists may be on the horizon. Regulatory bodies that oversee the numbers, size, or locations of training programs should give careful consideration to these workforce issues.⁶⁴  Additional data are needed to better understand the implications of modifications to PCCM education and training pathways, practice environments, and policies that will impact the PCCM workforce.

We thank Emily McCartha, Andrew Knapton, and Adriana R. Gaona for their review of the modeling data presented. We also thank Virginia A. Moyer and Patience Leino for their editorial support. Last, we thank the pediatricians who shared their information with the ABP Foundation and made this supplement possible.

ABP

American Board of Pediatrics

AMG

American medical graduate

APP

advanced practice provider

CWE

clinical workforce equivalent

DO

Doctor of Osteopathy

HC

headcount

IMG

international medical graduate

MD

Doctor of Medicine

NRMP

National Resident Matching Program

PCCM

pediatric critical care medicine