The Health Workforce Simulation Model (HWSM) is an integrated microsimulation model that estimates the current and future supply of and demand for health care workers by occupation, geographic area, and year. Demand projections also are modeled by employment setting. HWSM is designed to produce national and state-level workforce projections. Starting in 2019, HWSM models demand at the county level so that projections can be aggregated to report demand by metropolitan versus nonmetropolitan location. HWSM models the implications of changing demographics on health workforce supply and demand, as well as trends and policies affecting care use and delivery.
The purpose of workforce modeling is to quantify the implications of trends affecting health workforce supply and demand, and whether full-time equivalent (FTE)1 supply will be adequate to meet demand. The gap between supply and demand is often referred to as a shortage if demand exceeds supply, or as a surplus (or excess capacity) if supply exceeds demand. Such information promotes efficient allocation of resources regarding the number of health workers to train and whether jobseekers should enter a particular health occupation or specialty.
Workforce demand is defined as the number of health workers required to provide a level of services that will be utilized given patient health-seeking behavior and ability and willingness to pay for health care services. Training more health workers than required (i.e., excess capacity) can have detrimental consequences for providers seeking fulfillment in their career. Training too few health workers (i.e., shortage) reduces access to care—especially for historically underserved and vulnerable populations—and contributes to burnout among existing healthcare workers. As discussed later, demand is different from need. Demand reflects the level of care that people are likely to use in the absence of supply constraints while need is a clinical definition.
Starting year supply is estimated based on the number of people active in the workforce, which consists of people working and people actively seeking employment. It reflects estimates of hours worked to calculate FTEs. Estimates of active supply and FTE supply generally are lower than estimates of licensed supply (for occupations requiring a license) or number of trained workers (for health occupations that do not require a license.) This is because some individuals who are licensed and trained choose not to participate in the labor force. HWSM models the number of individuals trained each year who enter the workforce, so the supply projections are estimates of total number of people trained to provide services. Projections of total people trained might exceed total employment for an occupation.
HWSM uses a microsimulation approach to supply modeling, meaning that individual health workers are modeled with data obtained from:
- Associations like
- American Medical Association Masterfile
- American Dental Association Masterfile
- National surveys with a representative sample of health workers like
- American Community Survey
- Health Resources and Services Administration’s [HRSA] National Sample Survey of Registered Nurses
- State licensure files as available
For supply modeling, HWSM simulates the current workforce and labor force participation decisions to project how supply will evolve over time. The projections reflect estimates and assumptions of the annual number and characteristics of newly-trained workers entering a given occupation. They also include prediction equations that describe workforce attrition probabilities and weekly number of hours worked.
While the nuances of modeling differ for individual health occupations and medical specialties, the basic framework used within HWSM remains the same and consists of three components:
- The model for supply of health professionals
- The model for demand for health care services
- The staffing ratios that convert demand for services to demand for health care workers (Exhibit 1)
To project the number and characteristics of future health care workers and service users, HWSM simulated individual-level data based on predicted probabilities estimated from the current or base year data. Depending on the predicted probabilities, individual records were simulated to age forward. The aged individual-level records were then aggregated to obtain the projections by geographic area. On the service use side, the current utilization rates by individual characteristics were applied to projected populations at the national and state levels.
Modeling demand starts with constructing a database that contains characteristics for each person in a representative sample of the population in each county and state over time. Prediction equations model the expected demand for health care services based on each person’s characteristics like:
- health risk factors including smoking and obesity
- presence of diseases such as diabetes and cardiovascular disease, among others
- economic considerations including whether the person has medical insurance and level of household income
Demand for health care services is then used to model health workforce demand.
- 1For modeling, we measure both supply and demand in terms of full-time equivalents (FTEs). Unless otherwise specified throughout this report, demand is used synonymous with “FTE demand” and supply is used synonymous with “FTE supply.” An FTE has been defined as working 40 hours per week since the year 2017. Prior studies used average weekly hours worked within a profession to define an FTE, so the definition varied by profession. Hence, estimates of FTE supply will differ from other supply metrics such as licensed supply or active supply, or estimates that use a different definition for FTE.
Exhibit 1: HRSA’s Health Workforce Simulation Model
This diagram shows how various inputs are synthesized into projections of supply and demand. It gives a high-level view of the model. The individual components are explained in more detail in this module.
For supply, the model uses the starting year number of people active in the workforce. It then adds new entrants, like recent graduates and professionals moving into the field, and subtracts those exiting, like retirees and professionals moving out of the field. The result is the number of full-time equivalent workers at the end of that year. This value becomes the starting value for the next year, and the process continues until it reaches the final year of the model.
Labor market factors like wages and unemployment will impact new entrants and attrition. For example, workers may opt to exit the field if wages in other fields increase and make working in those fields more attractive.
For demand, we examine the population characteristics of the U.S. These include the demographic and socioeconomic characteristics of the current and future population as well as where they live. The model also uses data on the health conditions and health behavior of the population. These data are used to construct healthcare utilization patterns, which are then used to project the demand for specific healthcare services by provider type and setting.
The final demand step is to convert the demand for specific services into a demand for healthcare workers by using staffing ratios. For example, a projected number of x teeth cleanings demanded by the population would necessitate y number of dental hygienists to perform.
HWSM models future supply and demand under different scenarios reflecting trends and assumptions about key supply and demand determinants. All scenarios reflect changing demographics. For example, supply accounts for aging of the health workforce and differences between men and women in labor force participation. Demand accounts for population growth and aging. Additional supply scenarios model the sensitivity of projections to trends in early or delayed retirement, and training more or fewer health workers compared to current levels. Additional demand scenarios reflect estimates of how patient health care use might change if barriers to accessing care were removed, or how demand and/or supply might change as a result of developments and trends in our evolving healthcare delivery system.
Demand for health workers is based on projections of the level of health care services that patients will use and how the staffing is configured to deliver care. The "Status Quo" demand projections extrapolate current national health care use and delivery patterns by personal characteristics to the state and county level2 projected populations into the future. These hypothetical future populations’ demographics, disease prevalence, economic factors, and other health risk factors reflect the expected changes in these factors across geographic areas and over time. Therefore, demand estimates for each state reflect what demand would be for the population in that state if each person used the national average level of care for a like person. A like person is one with the same demographics, same health risk factors, same presence of disease, same rurality3 4 of residence, and same household income and insurance status.
Extrapolating current national patterns of care use and delivery does not imply that current patterns of care use and delivery are optimal or even efficient. This scenario simply reflects the realities of the current health care system and economic considerations including:
- medical technology
- health policy
- insurance coverage
- prices for health care services
- reimbursement rates to providers
- cultural norms
- other factors that affect care use and delivery
Status Quo projections of future demand indicate the number and occupation/specialty mix of health workers that would be required if we continue to use and deliver care according to current patterns. Alternative scenarios quantify future demand for health workers if care use and delivery patterns change. Comparison of Status Quo to alternative demand scenarios provides insights on the contribution of population growth and aging to future demand for health workers. This can be compared to the contribution of other factors that might change how care is used or delivered (for example, if dental schools started training more dentists or population health programs improved their patients’ overall health).
Current health care use and delivery patterns reflect the current supply of health workers. Hence, for many occupations modeled using HWSM, FTE national demand equals national supply in the base year. This approach is common across health workforce models.5 In HWSM, there are several occupation groups where national demand is calculated to be higher than national supply in the base year—primary care physicians, psychiatrists, and general surgeons. For these occupations, there is great concern of national shortfalls. There is also external evidence of substantial under-supply in many geographic areas as corroborated by the HRSA’s efforts to define health professional shortage areas (HPSAs).6 7 8 9 Demand projections for primary care physicians and psychiatrists use estimates of the number of providers required to remove HPSA designations as the starting year shortfall. The general surgeon shortfall national estimate models assumptions of adequate supply in urban areas but undersupply in rural areas. The specifics of the shortfall assumptions and data sources are described later in sections of this report.
For many occupations (e.g., psychiatrists, psychologists, and other types of behavioral healthcare providers), there is concern about substantial unmet need. That is, patients have a clinical need for care but do not receive the level of care that either they or a health provider considers appropriate. Unmet need can occur for many reasons including:
- supply access barriers where there is no provider within a reasonable distance to the patient
- financial access barriers where either the patient does not have the ability to pay for services or where providers do not accept the patient’s insurance
- patients choosing not to seek care because of stigma, cultural norms, they do not feel care is warranted, they are too busy, or other reasons
For some occupations, demand for health workers is modeled under a scenario that reflects how demand might change if access barriers were removed. Past studies have sometimes referred to this as an "unmet needs" scenario or as a "health care utilization equity" scenario depending on the assumptions and methods used. The rationale for this scenario is the desire to build a workforce to meet the future needs of a health care system that delivers on national goals to improve access to affordable, high-quality care.
- 2For estimation by metropolitan/nonmetropolitan designation in each state, county-level estimates were generated for oral health professions and for general surgeons in 2019 and for primary care providers, women’s health services providers, and behavioral healthcare workers in 2020.
- 3Before 2019, this was metropolitan statistical area (MSA) versus non-MSA; after 2019 this was stratified by the six classifications in NCHS’s Urban Rural Classification System.1 In this classification scheme, counties are classified (from highest to lowest population density) as “Large central metro,” “Large fringe metro” (which NCHS notes is a proxy for suburban), “Medium metro”, “Small metro”, “Micropolitan”, and “Non-core”.
- 4Centers for Disease Control and Prevention. 2013 NCHS Urban-Rural Classification Scheme for Counties.
- 5Ono T, Lafortune G, Schoenstein M. Health Workforce Planning in OECD Countries. Published online 2013.
- 6Bureau of Health Workforce, Health Resources and Services Administration. Designated Health Professional Shortage Areas Statistics: Third Quarter of Fiscal Year 2019 Designated HPSA Quarterly Summary. U.S. Department of Health and Human Services; 2019.
- 7Altschul DB, Bonham CA, Faulkner MJ, et al. State Legislative Approach to Enumerating Behavioral Health Workforce Shortages: Lessons Learned in New Mexico. American Journal of Preventive Medicine. 2018;54(6):S220-S229. doi:10.1016/j.amepre.2018.02.005.
- 8Ellison EC, Pawlik TM, Way DP, Satiani B, Williams TE. Ten-year reassessment of the shortage of general surgeons: Increases in graduation numbers of general surgery residents are insufficient to meet the future demand for general surgeons. Surgery. 2018;164(4):726-732. doi:10.1016/j.surg.2018.04.042.
- 9National Council Medical Director Institute. Psychiatric-Shortage_National-Council-.pdf (PDF - 5 MB). Published March 28, 2017. Accessed January 18, 2018.
HWSM consists of self-contained modules that describe different components of the health care system, which is consistent with best practices for modeling complex systems.10 HWSM runs using R software. HWSM continues to be maintained and refined with new occupations added periodically and scenario modeling capabilities enhanced. Each year, the model is used to project supply and demand for selected occupations and updated with the most recent data from key sources. Recently modeled occupations use more current data than occupations modeled in previous years. Substantial efforts continue to make HWSM transparent and peer-reviewed, with feedback used to refine HWSM inputs and assumptions.
The remainder of this report documents the logic, methods, data, assumptions, and validation processes for HWSM in general, and how the model is adapted to individual health occupations. Modeling Supply of Health Professionals and Modeling Demand for Health Care Services and Providers, respectively, describe the supply and demand components of HWSM. Other modules provide information on specific health occupation categories. HWSM Improvement, Validation, Strengths, and Limitations describes model validation activities, strengths and limitations. Other reports and fact sheets published by HRSA summarize model supply and demand projections by occupation.11
- 10Citro C, Hanushek E. Improving Information for Social Policy Decisions: The Uses of Microsimulation Modeling – Volume I: Review and Recommendations. National Academy Press; 1991:360.
- 11Health Resources and Services Administration. Projecting Health Workforce Supply and Demand. HRSA Health Workforce. Published 2021. Accessed January 30, 2021.