Documentation

About the Project

NC Nursecast was developed by The Cecil G. Sheps Center’s Program on Health Workforce Research and Policy, with funding from the North Carolina Board of Nursing and expertise from SMAP Ltd.

It consists of two parts:

The NC Nursecast Supply & Demand Model is an interactive, web-based tool that forecasts the future supply and demand for Registered Nurses (RNs) and Licensed Practical Nurses (LPNs) in various practice settings in North Carolina. The model is based on historical nurse licensure data, state population data, and health care survey and discharge data, as well as expert input from a nursing advisory committee. These sources were used to develop baseline estimates of how many nurses NC will have and how many nurses NC will require from 2019-2033, based on historical utilization patterns by region and practice setting. The model also provides alternative scenarios to explore the potential outcomes if nurses leave practice earlier than expected, if more students graduate from nursing programs, or if fewer nurses come to NC from outside of the state.

The NC Nursecast Graduate Diffusion tool is a separate interactive, web-based tool that allows users to see where students in NC’s nurse training program practice after graduating. That tool shows the percent of each program’s graduates working in North Carolina, the percent employed in different settings, and the percent located in rural areas.

Note that the NC Nursecast tools do not currently include Advanced Practice Registered Nurses (APRNs) such as Certified Registered Nurse Anesthetists (CRNAs), Nurse Practitioners (NPs), or Certified Nurse Midwives (CNMs).

Methods: NC Nursecast Supply and Demand Model

Data Sources

Supply Data:

The supply model uses licensure data obtained from the NC Board of Nursing (NC BON) for 2015 through 2018, for all RNs and LPNs licensed in the state. These data are collected as a part of the biannual license renewal process and represent a complete census of NC nurses. These NC BON data are housed in the NC Health Professions Data System (HPDS) at the Sheps Center and include demographic, education, practice, and location variables needed in the supply model described below.

Demand Data:

The demand model uses different data sources to model aspects of demand that are relevant to each given setting as set out in this table:

Setting Data source
Hospital

NC Discharge data 2013 to 2017 for:

  • Inpatient

  • Hospital outpatient

  • Emergency

Ambulatory Care National Ambulatory Medical Care Survey (NAMCS) data from 2011 to 2016 for Southern Region States only
Nursing Home

Residency information from NC Nursing Home licensure data for 2019

Due to incomplete data on the originating location of nursing home residents, data for nursing home forecast is based on location of the nursing homes

Mental Health Mental Health cases reported in Medical Expenditure Panel Survey (MEPS) for 2017 for Southern States
Home Health Home Heath visits reported in Medical Expenditure Panel Survey (MEPS) for 2017 for Southern States

Employment Setting Details

The model includes the following eight practice settings: (1) hospital, (2) ambulatory care, (3) nursing home, extended care and assisted living facilities, (4) home health/hospice, (5) community and population health, (6) academic settings/nursing education, (7) mental health, and (8) correctional facilities. These data are based on self-reported category choices by NC nurses at the time of licensure renewal with the NC Board of Nursing.

Self-reported practice settings can be inconsistent because nurses are not provided with direction on the licensure form about to categorize themselves into different employment settings.

We did not model every possible practice setting, and this model also assigns each nurse to a single practice setting, even though some nurses split their time between multiple settings. The licensure data only captures the practice setting in which the nurse is working at the time of biannual licensure renewal, not every setting the nurse worked within over the course of the preceding two years.

Additional setting-specific methodological notes:

  1. Community and population health

    • Includes nurses who reported practice settings of “Community Health,” “Occupational Health,” “Public Health,” and “School Health Service.”
  2. Mental Health

    • Nurses working in mental health settings could not be identified from practice setting selection options, so we used nurse-reported specialty and employer name to categorize these nurses.
  3. Nursing Education

    • For nurses reporting “Academic Setting” as their practice setting, we conducted an additional review of employer names to identify the subset who worked within nurse training programs. It was this subset that we labeled as working in the setting “Nursing Education.”

Modeling Details

NC Nursecast’s supply and demand model includes four key elements:

  1. A supply model, which estimates how many nurses will be available in North Carolina in the future, based on current trends.

  2. A demand model, which estimates how many nurses will be required in North Carolina in the future, based on current trends.

  3. Estimates of future nursing workforce shortages or surpluses in North Carolina, calculated by combining the supply and demand models.

  4. Alternative supply scenarios which estimate future nursing workforce shortages or surpluses if current trends change: if nurses accelerate or delay retirement, for example, or if nursing schools increase class sizes.

See details on each, below.

1. The Supply Model

The supply model projects the future headcount and FTE of RNs and LPNs from 2019-2033. The basic approach in this model, as in any workforce supply model, is to take the current workforce, subtract “leavers” (nurses who retire, leave the workforce, or move out of state) and add “joiners” (newly-trained nurses or nurses who moved into the state) for each year to generate a forecast for the next year. The approach is summarized visually below:

Flow chart showing the different components that contribute to the supply forecast.

Figure 1: Supply Model for Estimating North Carolina RN and LPN workforce, 2019-2033

The supply model is based on an agent-based microsimulation. This involves creating a “synthetic” version of the North Carolina nurse workforce (based on the characteristics of the actual workforce), applying assumptions based on current leaving and joining trends, and performing many repeated computer simulations of possible trajectories of the individual synthetic nurses based on their age, nurse type (RN or LPN), degree (ADN, BSN, or LPN), gender, and employment setting (hospital, ambulatory care, nursing home, extended care and assisted living facilities, home health/hospice, community and population health, academic settings, mental health, and correctional facilities).

For more details on these employment settings, see the “Data Sources” and “Employment Setting Details” sections, above.

Limiting Analysis to RNs and LPNs

NC Nursecast includes RNs and LPNs. Advanced Practice Nurses— nurse practitioners (NPs), certified nurse midwives (CNMs), and certified registered nurse anesthetist (CRNAs)—are excluded from the model. However, because APRNs are licensed as RNs in NC, we needed to develop a method to exclude them from the data used by the model.

NPs and CNMs with an active practice agreement in NC were identified as APRNs. CRNAs and CNSs self-report their roles in the licensure data. NPs, CNMs, and CRNAs in active practice in North Carolina were removed from the RN workforce data. However, CNSs remain in the model because of the significant overlap between RN and CNS roles. NPs reporting an active RN practice location but not an active NP practice location remain in the RN workforce data. APRNs who are not currently practicing as an APRN also remained in the RN workforce modeled.

Modeling Geographic Mobility

Based on licensure data between 2015 and 2018, the probability of a nurse changing settings and geographic location was calculated. The model uses these probabilities to estimate whether an individual nurse changes his or her geography and setting in each year of the forecast.

2. The Demand Model

Demand was modeled at the county level for each employment setting. There is significant variation in the data available for measuring existing demand in different employment settings, so we used several different approaches for calculating future demand.

For community and population health, nursing education, and correctional facilities, the demand forecast is dependent on population change (which means, for example, a 10% growth in a county’s population would generate a 10% growth in the demand forecast for these settings in that county).

For the remaining settings, we developed separate regression models to forecast demand based on predictors such as gender, age, and race/ethnicity.

For ambulatory care settings, there has been an increase in the number of nurses working in this setting, much of which as a result of nurses undertaking new roles in ambulatory care rather than an increase in total demand for ambulatory care. Therefore, to forecast ambulatory demand based on changes in population alone will not reflect the expected continued increase in ambulatory care nurses. Following input from chief nursing officers and nursing leaders in ambulatory care, it was agreed that the model would continue the historical increase in ambulatory care nurses until 2023, between 2023 and 2025 there is a smoothed transition to demand driven by population growth which is used beyond 2025.

Forecasts of future demand were then generated using the population forecasts from the NC Office of Budget and Management to model the effect that population growth and the changing composition of the population by age, gender, and race would have on demand. Non-population forecast data such as percent insured was assumed to be constant and remains unchanged for all years.

Demand for future visits depending on setting were converted to demand for FTE by using the ratio of visits to FTE in 2018 for all future years. For example, a 10% increase in demand for ambulatory visits is modeled as a 10% increase in demand for ambulatory nurse FTEs from the baseline of 2018.

Demand forecasts by headcount were estimated by using the average headcount to FTE ratio for a given setting.

The model aggregates the county-level results to produce forecasts of demand at the Area Health Education Center (AHEC), Medicaid , and Metro/Non-Metro regions. A crosswalk of counties to these larger units of geography is available.

3. Estimating Shortage and Surplus

Like most other workforce models, NC Nursecast assumes that in 2018 demand is matched by supply - neither a shortage nor a surplus. This assumption was made due to the lack of data to reject this assumption, as vacancy data are not collected in a consistent way for North Carolina. The one exception is for RNs in nursing education for which vacancy data were available. The data indicated a 7% shortage in FTE in 2018, which was incorporated into the model with demand exceeding supply by 7%.

Confidence Intervals

Confidence intervals are used to convey the uncertainty in the forecast and to provide an estimated range around a predicted value. As a comparison, it is generally accepted that the forecasting of a hurricane trajectory has a degree of uncertainty in that the farther the forecast goes into the future, the less certain the path of the hurricane. Workforce modeling has a similar degree of uncertainty generated by minor changes within a year that over time have a cumulative effect. To model this, the multiple iterations of the scenario are completed, within each iteration there will be variation in, for example, the year a nurse leaves the workforce, which means each iteration will generate slightly different forecasts. The calculation of 95% confidence intervals takes into consideration the cumulative effect of individual variation of iterations of a scenario.

4. Modeling Alternative Supply Scenarios

The baseline model assumes that the factors affecting the supply and demand for nursing services in 2019 will continue as they have in prior years. Yet, we know that nursing workforce participation patterns, models of care, and other factors are likely to change. To account for these deviations from the current status quo, we modeled five alternative future scenarios known to affect supply and demand:

  1. Early exit from the workforce by 2 years: nurses leave the workforce two years earlier than they would have based on historical data, with all nurses retiring by age 68

  2. Early exit from the workforce by 5 years: nurses leave the workforce five years earlier than they would have based on historical data, with all nurses retiring by age 65

  3. Delayed exit from the workforce by 2 years: nurses delay leaving the workforce by two years, although all nurses still retire by age 70

  4. Reduction in out-of-state supply by 2.5%

  5. Increase in NC new graduate nurse supply by 10%

We also modeled a sixth scenario that combines three of the five scenarios and reflects a situation that could occur together given some early indicators of the current NC nursing environment and the COVID-19 pandemic:

  1. Combined scenario: nurses exit the workforce five years earlier than they would have, with all nurses retiring by age 65 + a reduction in out-of-state supply by 2.5% + an increase in NC graduate supply by 10%

Do these Scenarios Estimate the Effects of COVID-19?

The alternative supply scenarios were developed prior to the COVID-19 pandemic, but they can be used to illustrate some possible types of pandemic-related workforce effects. For example, the “early exit” scenarios might inform discussions about burnout, and the “increase in graduate supply” scenario could reflect a pandemic-related “hero effect” that motivates students to pursue careers in nursing. A combination scenario of early exit + decreased out-of-state supply + increased graduate supply is available to visualize the combined effects of these scenarios that might occur together due to the pandemic. Note, again, that the magnitude of any COVID-19-related changes is still unknown, but these scenarios can provide a starting point for discussion.

Supply & Demand Model Terminology and Calculations

  • Supply: The supply model projects the future headcount and FTE of RNs and LPNs from 2018-2030.
  • Demand: The demand model projects the future headcount and FTE of RNs and LPNs from 2018-2030.
  • Supply-Demand by FTE: The absolute difference between the supply and demand forecasts, calculated as supply minus demand for a given year. This calculation provides an estimate of the absolute number of nurses in surplus or shortage.
  • % Surplus or Shortage: The percentage shortage or surplus of supply relative to the demand forecasts. When this percentage is negative, it means demand is greater than supply. The calculation is useful for understanding the relative surplus or shortage for a projection.

Methods: Graduate Diffusion Tool

North Carolina has over one hundred nursing programs. Some programs educate licensed practical nurses (LPNs) while others educate registered nurses (RNs) and award an associate’s or bachelor’s degree. Where do these nurses go after they have graduated and how do the patterns of diffusion differ between different types of programs and individual programs?

To help answer those questions, we use licensure data to determine where a nurse completed their first nursing degree for licensure and where they were working two years later. Then we put those origin and destination points on a map for each of North Carolina’s nursing programs. We drew an ellipse around those points to show the spatial distribution of approximately 23 of the graduates from each program, giving a general idea of the diffusion pattern of a school’s graduates.

Data Sources:

The in-state retention data available in the tables comes from NC Tower, which provides aggregate information on students who attended public universities and community colleges in North Carolina. Therefore, only public institutions have a retention value listed.

Web Model Interface Details

The website and data visualizations are built using a variety of open-source software, most notably the D3 data visualization library, the JavaScript framework Svelte and the JavaScript bundler Rollup. The nurse diffusion map also makes heavy use of the Mapbox GL JS JavaScript library. Most of the website style (i.e., CSS) is defined using Bulma, a CSS framework. The website is composed and generated using Hugo.

The data are housed on a server provided by Carolina CloudApps within a container running Node.js. The main job of this server is to pass data from a mySQL database to the user interface. The database is loaded with precalculated projection data for all the parameter combinations in the visualizations.

Definitions

Advanced Practice Registered Nurse, APRN
A registered nurse (RN) who has completed additional master’s or post-master’s level education in a specialty area. APRNs include nurse practitioners, certified nurse midwives, certified registered nurse anesthetists, and clinical nurse specialists. APRNs are excluded from the NC Nursecast RN supply model.

Area Health Education Center (AHEC) Regions
North Carolina’s Area Health Education Center Program is a statewide program to develop the state’s healthcare workforce, which is administered through nine regions made up of adjacent counties. Each region has its own AHEC center focused on the unique workforce needs of the region.

Full Time Equivalent, FTE
One of two ways to count a workforce (see headcount). Each nurse is counted as a proportion of full-time based on the number of hours they work. Our model defines full time as 40 hours per week. For example, a nurse may only work as a nurse for 4 hours per day on Monday through Thursday, a total of 16 hours and an FTE of 0.4 (1640). The other way to count is by headcount. We modeled FTE in the probabilistic model based on the mean and standard deviation of hours worked as a function of age, gender, and employment setting using NCBON licensure files which contain self-reported average hours worked per week in nursing.

Headcount
One of two ways to count a workforce (see full time equivalent). Each nurse is counted as 1.0 regardless of whether they are full or part time.

Licensed Practical Nurse, LPN
A nurse who has completed a practical or vocational nurse education and clinical training program (approximately 12-18 months in duration), passed a licensure exam, and completed licensure with a state board of nursing.

Medicaid Region
North Carolina launched the NC Medicaid Managed Care model on July 1, 2021. To administer the program, the NC Department of Health and Human Services has divided the state into five regions. Forecasting demand at this level is particularly useful when looking at the demand of the Medicaid beneficiary population.

Metropolitan
A county that contains a core urban area of 50,000 or more people, as defined by the United States Census Bureau and the Office of Management and Budget. Typically, this designation is used to define counties that are urban or not rural. This model uses the 2017 vintage of the county delineation file in which 46 counties in North Carolina are metropolitan. See non-metropolitan.

Non-metropolitan
A county that does not contain a core urban area of 50,000 or more people, as defined by the United States Census Bureau and the Office of Management and Budget. Typically, this designation is used to define counties that are rural. This model uses the 2017 vintage of the county delineation file in which 54 counties in North Carolina are non-metropolitan. See metropolitan.

Registered Nurse, RN
A nurse who has typically completed either a two-year postsecondary degree (an Associate’s Degree in Nursing, or ADN), a diploma in nursing (typically 18-32 months of training offered by a hospital), or a four-year postsecondary degree (Bachelor of Science in Nursing, or BSN), passed a licensure exam, and completed licensure with a state board of nursing.

Explore

Supply & Demand

Want to learn about the future supply and demand of our state's licensed practical nurses and registered nurses across settings and geographic regions?

See projections

Graduate Diffusion

Want to see how different North Carolina nursing programs impact the distribution of health professionals in their area and across the state?

Examine graduate diffusion

Key Findings

What are the main takeaways from the nursecast projections? In what regions and settings will nurses be in shortage?

Read key findings

Who are we?

This project is brought to you by The Cecil G. Sheps Center For Health Services Research at the University of North Carolina in partnership with the North Carolina Board of Nursing.