1. Home
  2. Data & Research
  3. Health Workforce Projections
  4. Technical Documentation for HRSA’s Health Workforce Simulation Model
  5. XI. Allied Health & Select Other Occupations Model Components

XI. Allied Health & Select Other Occupations Model Components

Published 2023

In this module:

This module summarizes adaptation of the Health Workforce Simulation Model (HWSM) to project supply and demand for select allied health occupations and several other occupations (e.g., pharmacists, podiatrists, and audiologists). Because of data limitations, which are discussed later, supply projections could only be made for 21 of the 35 occupations included in this component of the workforce modeling effort (Exhibit XI-1).

The research team reaches out to professional associations representing various health professions, with associations representing the modeled health professions contacted approximately every 2-3 years. The team seeks to identify the best available data sources, discuss trends affecting workforce supply and demand, and provide the opportunity for feedback on preliminary findings. The information provided in this technical documentation and in HRSA reports does not necessarily reflect the views of the associations that responded. There may not be clear consensus on all assumptions.

Individuals from the following associations have participated:

  • American Association for Respiratory Care (AARC)
  • American Association of Medical Assistants (AAMA)
  • American Occupational Therapy Association (AOTA)
  • American Optometric Association (AOA)
  • American Physical Therapy Association (APTA)
  • American Podiatric Medical Association (APMA)
  • National Association of Boards of Pharmacy (NABT)
  • National Association of Community Health Workers (NACHW)
  • National Registry of Emergency Medical Technicians (NREMT)
  • Pharmacy Technician Certification Board (PTCB)
  • Society of Nuclear Medicine and Molecular Imaging (SNMMI)

Modeling supply

For the occupations where sufficient data is available to project future supply, we modeled multiple scenarios.

A Status Quo scenario extrapolates current trends in supply determinants:

  • Number and characteristics of current and new providers
  • Hours worked patterns
  • Attrition patterns

Alternative scenarios modeled the sensitivity of future supply to changes in key trends and assumptions. One set of scenarios models a ±10% change in the annual numbers of new graduates entering the workforce. Another set of scenarios reflects if attrition patterns were to change such that providers retire up to two years earlier or delay retirement by up to two years, on average, relative to historical retirement patterns.

A key challenge with modeling supply for allied health occupations is that many occupations have multiple channels for entry to the occupation (including on-the-job-training) so there is a lack of data on the number of new entrants. For many of these occupations (particularly aide occupations), pay is relatively low so there is easy exit from the occupation.

Estimating the base year workforce supply

The primary sources for data on current number and characteristics for the occupations modeled is the pooled 2017-2021 American Community Survey (ACS) supplemented with the Occupational Employment and Wage Statistics (OEWS) data and licensure counts from specialty associations. We used multiple years of ACS data to create a larger sample for each state to produce estimates of current supply by age, sex, and race/ethnicity. Counts for the number of dieticians came from the Academy of Nutrition and Dietetics. This allows for modeling a more precise category than the Standard Occupation Code category (dieticians and nutritionists) in ACS or OEWS. An estimate of the number of podiatrists came from the American Podiatric Medical Association (APMA) and an estimate for the number of optometrists came from the American Optometric Association (AOA), which are based on licensure files. An estimate for the number of radiation therapists came from the American Registry of Radiologic Technologists’ (ARRT) census of March 2022. We used historical 2012-2016 OEWS data to reach separate starting supply numbers for medical and clinical laboratory technologists and medical and clinical laboratory technicians. In more recent OEWS and ACS data, these two occupations are combined.

Modeling new entrants

New entrants were added to the workforce via a “synthetic” cohort. The size of the cohort is based on the number and characteristics of recent graduates in each occupation. The primary source for estimating new entrants in each occupation is the 2021 Integrated Postsecondary Education Data System (IPEDS). Data provided by associations is used instead if such data is available. The number of new entrants for registered dietitians is based on 2021 graduates' data from the Accreditation Council for Education in Nutrition and Dietetics. Physical therapist and physical therapy assistant new entrant’s numbers come from 2021 accreditation data from the American Physical Therapy Association. The number of new entrants for radiation therapists used the American Registry of Radiologic Technologists’ data on radiation therapy exams taken in 2021. Respiratory therapist number of entrants come from 2021 accreditation data from the Commission on Accreditation for Respiratory Care. Finally, the number of new entrants for audiologists and speech-language pathologists is based on the 2020-2021 Communication Sciences and Disorders Education Survey.

Starting in 2023, there are new approaches for calculating the number of new entrants for pharmacy occupations. For pharmacists, data on the sizes of upcoming classes were provided by the American Association of Colleges of Pharmacy. These data show that the number of pharmacists being trained will decrease from 14,223 in 2022 to 11,692 in 2025. For pharmacy technicians, the data source for the number of new graduates is the 2021 IPEDS. However, that number was divided by 0.242 based on a survey provided by the Pharmacy Technician Certification Board showing that 24.2% of pharmacy technicians trained through educational institutions.

Furthermore, the phlebotomist and medical and clinical laboratory technician new entrants' numbers are based on the 2021 IPEDS but also include Steven Wilber’s analysis estimating how many people enter these occupations annually through on-the-job training.

Each new worker is assigned age, sex, and race that reflect the distributions seen in recent years (Exhibit XI‑2). For race and sex, we use the 2021 IPEDS data on graduates by race and sex. To calculate the age distribution of new entrants to the workforce, we compared consecutive years of ACS data. We compared the number of providers of a particular age (e.g., age 29) to the number of providers in the subsequent ACS annual file who were one year older (e.g., age 30). The subsequent ACS file contains information on the workforce from the previous year plus any additions or subtractions to the workforce. In the above example, the difference between the number of providers at age 30 and the number of providers at age 29 in the previous year reflects the net number of new providers entering the workforce at age 29-30. With this information, we estimated the age distribution of new net entrants to the workforce. We used multiple years of data to increase sample size by individual age and occupation. The number of new entrants and their age-sex distribution were assumed to remain constant during the projection period.

Some occupations (usually those titled as aides, technologists, or technicians) have data limitations that preclude projecting future supply. For occupations that do not require licensure or that have few educational or training requirements, there are insufficient data on the number of people newly entering these occupations. Likewise, occupations with easy entry often have high turnover rates. This creates challenges to obtain reliable supply forecasts.

Modeling workforce participation and workers leaving the workforce

Labor force participation rates for health care workers were calculated directly for individuals under age 50 using 2017-2021 ACS data. The analysis used information on whether the person is active in the workforce. Active is defined as working at least one hour per week in their profession. Each simulation year, HWSM model the propensity that a health care worker under age 50 becomes inactive and thus count for 0 FTE supply. However, these individuals remain part of the supply model and are eligible to become active again in the following simulation year.

Health care workers can also leave the workforce of their current occupation permanently by changing careers. Each year, workers under age 50 have a small probability of being removed from the workforce based on observed career change using the Current Population Survey (CPS) Annual Social and Economic Supplement survey data. For audiologists and podiatrists, the sample size was too small for reliable estimates, so physicians were used as a proxy for podiatrists and speech-language pathologists were used as a proxy for audiologists. We also used pharmacy technicians as a proxy for phlebotomists and registered dieticians as a proxy for medical and clinical laboratory technicians. While these workers may join other health care occupations, we assume that they are captured in the new graduate data for that occupation.

For health care workers at age 50 and over, we modeled the probability of retiring, with retirement probability increasing with age. Attrition patterns for each profession were based on 2017-2021 ACS data. These patterns were constructed based on a question asking whether the person is currently employed and whether they were employed last year. Health care workers flagged for attrition drop out of the supply entirely and do not return in later years.

Modeling hours worked

We used Ordinary Least Squares regression to model hours worked patterns for each health occupation. The dependent variable is total hours worked in the previous week. Explanatory variables consisted of age group, sex, race, and ACS survey year (2017-2021).

Estimates and projections consider the changing demographics of the workforce. It also accounts for average hours worked per week differing by age, sex, race, and occupation. Then, the expected number of hours worked by each professional is converted to FTE supply by dividing the total person-hours worked by 40. This creates a uniform standard of 1 FTE as working 40 hours per week regardless of the occupation. It also means that the initial FTE of an occupation can differ from the actual count of persons employed in the occupation.

Modeling state-level supply and migration

States differ in estimated level of supply adequacy for various health occupations. Supply modeling is done at the state level where sufficient data are available, though for some smaller occupations the workforce projections are only reported at the national level. In the ACS and OEWS, some occupations do not have totals reported in every U.S. state. Occupations affected include chiropractors, podiatrists, radiation therapists, audiologists, and occupational therapy assistants.

For occupations with sufficient state-level data, HWSM models movement of health care workers across states. This is accomplished in two steps. First, a logistic regression1 on 2017-2021 ACS data estimates the probability of migrating to another state for the under 50 population. This regression is a function of age group, sex, race, the current state’s population, and a year indicator. Applying the prediction equation, HWSM calculates probabilities for each health care worker and compares probabilities to a random number generated on a uniform distribution to simulate moving to another state. Then, it assigns those who move a new state based on the destination state distribution observed in 2012-2021 ACS data for health care workers who changed residence states. We used more years of ACS data to increase sample size for this distribution because the number of individuals observed moving into each state in the ACS is small.

Supply scenarios

A Status Quo scenario models the supply implications if current patterns continue through the projection period.

Specifically, this scenario holds constant:

  • Number and characteristics of current and new providers
  • Patterns of hours worked
  • Patterns of workers leaving the workforce

Alternative scenarios allow us to gauge the sensitivity of future supply to key trends and assumptions. These scenarios reflect future uncertainty.

These scenarios include:

  • 10% higher annual numbers of new graduates entering the workforce
  • 10% lower annual numbers of new graduates entering the workforce
  • Workers retire two years earlier than current patterns
  • Workers retire two years later than current patterns

Concerns about provider burnout throughout the health care workforce have been prevalent in recent years, including among the allied health workforces.2 3 4 Hence, there is likely higher probability of allied health providers accelerating retirement versus delaying retirement.

Modeling demand

The Demand Modeling Overview module provides an overview of the data sources and approach used to model demand for health care services and providers. Here, we provide additional details specific to modeling demand for the health occupations covered in this analysis.

Demand for workers in many occupation categories are based on rates of patient-clinician encounters across employment settings. This starts with the industry distribution estimated from analysis of ACS mapped to employment settings (Exhibit XI‑3). For occupations that work in provider office, outpatient, and home health settings the growth in demand for health care workers is tied to growth in demand for visits in those respective settings (Exhibit XI‑4). For some occupations total visits is the proxy for demand, while for other occupations a subset of visits is the proxy for demand. For example, demand for chiropractors is driven by the projected number of visits to the offices of chiropractors. Due to a lack of data on visits to opticians, optometrist visits are used as a proxy for demand for both opticians and optometrists. That is, the growth in demand for eye care (as modeled by projected growth in patients seeking optometrist services) is used as a proxy for growth in demand for both optometrists and opticians. Demand for health care workers in hospital inpatient settings is modeled using either total inpatient days or a subset of inpatient days directly tied to the occupation.

Dividing (a) total national health care utilization by (b) total health care workers (plus estimated starting year shortages), by delivery site, provides a national ratio (Exhibit XI‑5). Under the Status Quo scenario, these ratios are assumed to be constant over time. These ratios are not estimates of the number of patient encounters with the health care worker during the course of the year. Rather, these ratios simply relate growth in demand for health care workers to the growth in demand for health care services in the setting in which they work.

Demand for pharmacists, pharmacy technicians, and pharmacy aides is tied to the number of prescriptions written. Using the pooled files of the latest available from the National Ambulatory Medical Care Survey (NAMCS) and the outpatient National Hospital Ambulatory Medical Survey (NHAMCS), we modeled average number of prescriptions written during an office, outpatient, or emergency visit by medical specialty or diagnosis category associated with the visit.5 We pooled 2014-2016 NAMCS files, which report detailed physician specialty, to model prescriptions by visit to a particular specialty. Later years of NAMCS do not include detailed physician specialties, so we applied the distribution of detailed specialties in the pooled 2014-2016 NAMCS data to the overall weighted estimates of the recent 2019 NAMCS. The most recent NHAMCS outpatient public use files available for analyzing prescriptions are 2007-2011. The 2012-2017 NHAMCS outpatient survey data is currently on hold due to data quality issues, and data was not collected for 2018-2020.

When multiplied by the projected number of visits, this provides a projection of future prescriptions written. The assumption is that the ratio of filled to written prescriptions will remain constant over time. Medications prescribed for hospital inpatient is modeled as a ratio of national inpatient days to national number of pharmacists working in hospitals. The projected growth in demand for prescriptions in ambulatory settings is used to project growth in demand for community-based pharmacists and those working in mail-order pharmacies. The projected growth in hospital inpatient days is used to project growth in demand for hospital-based pharmacists.

Emergency medical technicians (EMTs) and paramedics provide a wide range of services—many of which do not show up in medical billing records. The number of ambulance arrivals and transfers between hospitals is used as a proxy for demand for EMTs and paramedics, based on analysis of the 2018-2019 NHAMCS emergency department file and the 2019 NIS which as discharge information indicating if a patient was transferred to another hospital.

The Status Quo scenario starts with the assumption that national demand equals supply plus a current shortfall. The magnitude of the current shortfall is calculated as the estimated shift up in demand caused by COVID-19 becoming endemic. As described in the Demand Modeling Overview module, we estimate an increase in demand for outpatient care and hospital inpatient care associated with COVID-19 becoming endemic. For allied health occupations, the estimated increase varies by occupation but is about 1-2% increase for most allied health occupations. This shift up in demand, despite supply being relatively static, creates a small shortfall that is likely a conservative estimate of shortfall as there is a growing perception of shortfall across many allied health occupations but little data to quantify the shortfall.6 The Status Quo scenario then extrapolates current patterns of care into the future. Predicted probabilities of health care use were applied to project health care service use specific to the settings where these professionals are employed.

Modeling demand for community health workers (CHWs) is challenging because CHWs work in many different capacities—with approximately 62% working in state or local government programs. Employment opportunities for CHWS are driven largely by government funding, while the need for CHWs is substantial as indicated by disparities in health outcomes by socioeconomic circumstances, by race/ethnicity, and by other dimensions such as geographic location and studies showing the cost-effectiveness of using CHWs.7 8 9 10 For CHWs working in hospital inpatient settings, we use total inpatient days as the proxy for service demand. For CHWs working in ambulatory settings or home health, we use visits for clinical social workers as proxy for demand. For CHWs working in government programs, we use total population as the proxy for demand. Hence, projected growth in demand for CHWs is a weighted average of growth in demand for hospital inpatient days, ambulatory visits, home health visits, and population growth with the weights corresponding to the number of FTEs working in the various employment settings.

Demand scenarios

As with other health occupations, HWSM models demand for allied health workers under two scenarios—a Status Quo scenario and a Reduced Barriers scenario.

The Status Quo scenario models pre-pandemic (2015-2019) national patterns of how people seek health care services plus estimates of demand for acute and Long COVID care as COVID-19 becomes endemic. Care seeking patterns are based on peoples’ characteristics and are projected to continue over the projection period. This scenario captures geographic variation in demographics, health risk factors, disease prevalence, insurance coverage, level of rurality, and household income that can affect demand for health care services. Similarly, the scenario captures population growth and aging—and the associated implications for disease prevalence and other health risk factors—over the projection horizon. This scenario evaluates whether the nation’s future allied health workforce is sufficient to provide at least the base year level of care.

The Reduced Barriers scenario estimates the number of allied health worker FTEs required if populations that historically faced barriers to accessing health care services demonstrated care use patterns comparable to populations perceived to have fewer barriers to accessing care. It models if everyone in the U.S. used health care services at the same rate as a population with the lowest level of barriers to accessing care.

This population is:

  • Non-Hispanic white
  • Living in metropolitan areas
  • With medical insurance

This scenario highlights that national efforts to improve health equity could require a larger allied health workforce as demand for diagnostic services, pharmaceuticals, and other health care services would rise under the modeled scenario.

Date Last Reviewed:
  • 1 Logistic Regression. Published May 14, 2018. Accessed September 23, 2022.
  • 2 Rubin B, Goldfarb R, Satele D, Graham L. Burnout and distress among allied health care professionals in a cardiovascular centre of a quaternary hospital network: a cross-sectional survey. CMAJ Open. 2021;9(1):E29-E37.
  • 3 Prasad K, McLoughlin C, Stillman M, et al. Prevalence and correlates of stress and burnout among U.S. healthcare workers during the COVID-19 pandemic: A national cross-sectional survey study. EClinicalMedicine. 2021;35:100879.
  • 4 Miller AG, Roberts KJ, Smith BJ, et al. Prevalence of Burnout Among Respiratory Therapists Amidst the COVID-19 Pandemic. Respir Care. Published online July 16, 2021:respcare.09283.
  • 5 National Center for Health Statistics. Ambulatory Health Care Data. Published June 28, 2022. Accessed July 22, 2022.
  • 6 AMN Healthcare Services Inc. AMN Healthcare Survey: 85% of Healthcare Facilities Face Shortages of Allied Healthcare Professionals. October 20, 2022. Accessed August 3, 2023.
  • 7 Kangovi S, Mitra N, Norton L, et al. Effect of Community Health Worker Support on Clinical Outcomes of Low-Income Patients Across Primary Care Facilities: A Randomized Clinical Trial. JAMA Internal Medicine. 2018;178(12):1635-1643.
  • 8 Vasan A, Morgan JW, Mitra N, et al. Effects of a standardized community health worker intervention on hospitalization among disadvantaged patients with multiple chronic conditions: A pooled analysis of three clinical trials. Health Services Research. 2020;55(S2):894-901.
  • 9 Hayhoe B, Cowling TE, Pillutla V, Garg P, Majeed A, Harris M. Integrating a nationally scaled workforce of community health workers in primary care: a modelling study. J R Soc Med. 2018;111(12):453-461.
  • 10 Naufal G, Naiser E, Patterson B, Baek J, Carrillo G. A Cost-Effectiveness Analysis of a Community Health Worker Led Asthma Education Program in South Texas. J Asthma Allergy. 2022;15:547-556.