1. Innovations in Health Workforce Modeling Academy Health Presentation June 5, 2004 Presented by Tim Dall The Lewin Group

2. 2 Acknowledgements Sponsored by the National Center for Health Workforce Analysis special thanks to: Steve Tise: project officer for physician workforce study Marshall Fritz: project officer for nursing demand study Bill Spencer: project officer for nursing supply study Marilyn Biviano: former director of NCHWA The Lewin Group Paul Hogan, Atul Grover, Elisa Golfinopolous Altarum Charlie Roehrig, Mary Bannister, Sara Eisenstein, Carolina Fulper Disclaimer: No official endorsement of this work by the U.S. DHHS or any of its components is intended or should be inferred.

3. 3 Agenda Data, assumptions and methods used to create four workforce models –Nursing Supply Model (NSM) –Nursing Demand Model (NDM) –Physician Supply Model (PSM) –Physician Demand Model (PDM) Additional research needed to make these models more valuable as policy tools

4. Nursing Supply Model Overview

5. 5 50 States plus the District of Columbia RNs at three education levels –Diploma or associates degree –Baccalaureate degree –Masters or higher degree RN age Years 2000 to 2020 NSM Projection Dimensions

6. 6 New graduates from U.S. nursing schools Immigration of RNs from outside the U.S. Change in educational attainment Cross-state migration Attrition from the licensed RN population (e.g., due to career changes, retirement, death and disability) RN labor force participation rates NSM Projection Determinants

7. 7 Nursing Supply Model Overview Beginning of the Year End of the Year US Graduates Foreign Graduates Retirement Death/disability Career change Education upgradesCross-state migration Nurse aging Nurse Population

8. 8 RN population: the number of licensed RNs Active RNs: the number of RNs employed in nursing or seeking employment in nursing FTE supply: the number of full-time- equivalent (FTE) RNs employed in nursing Three Measures of RN Supply Projected

9. 9 Sample Survey of RNs (1992, 1996, 2000) –To estimate base year (2000) RN population by age, education level and State –To estimate labor force activity rates by RN age and education level –To estimate cross-state migration patterns Current Population Survey (1999-2001) –To estimate RN retirement and disability rates Census Bureau population projections 2000 to 2020 –To estimate changes in the size of the female population age 20-44, the traditional candidate pool for RNs NSM Primary Data Sources

10. 10 NSM Limitations Parts of the NSM are static –Retirement, education upgrades, and labor force participation patterns constant across States and over time, but differ by RN age and education level –Cross-state migration patterns constant over time, but vary by State and RN age and education level –NSM user can change assumptions regarding key variables RN supply projections are independent of –RN demand projections –Projected supply of other health workers (e.g., LPNs) Small sample size in some surveys (e.g., RN Survey) reduces reliability of projections for smaller States

11. Nursing Demand Model Overview Nurse Demand is defined as the number of full time equivalent (FTE) nurses that society is willing to pay for given population needs, economic considerations, the healthcare operating environment, and other factors.

12. 12 12 settings (e.g., hospitals, nursing homes, home health, doctors offices, public health) 50 States plus the District of Columbia Years 2000 to 2020 RNs, LPNs, nurse aides, home health aides NDM Projection Dimensions

13. 13 Demographics and geographic location of population Characteristics of the health care system—e.g., –Shift from hospital inpatient to outpatient services –Percent uninsured –HMO saturation rate Economic factors—e.g., –Nurse wages (e.g., ratio of RN to LPN hourly wages) –Reimbursement rates –Per capita personal income NDM Projection Determinants

14. 14 Nursing Demand Model Overview Nurse Demand Projections Health care use projections Nurse staffing intensity equations Health care use equations Per capita health care utilization Trends in health care market conditions, economic conditions, patient acuity, etc. Population projections

15. 15 Nurse demand (data for 1996 to 2000) –Sample Survey of RNs: estimates of RNs by employment setting –American Health Care Association: estimates of FTE RNs employed in nursing facilities Health care demand (data for 1996 to 2000) –Area Resource File –National Inpatient Sample –National Ambulatory Care Survey –National Hospital Ambulatory Care Survey –National Nursing Home Survey –National Home and Hospice Care Survey Population projections: Census Bureau (data for 1996 to 2020) NDM Primary Data Sources for RNs

16. 16 NDM Limitations Simplified model of a complex health care system –Contains limited number of demand determinants –Has limited ability to model effect of economic determinants on substitution of nurse types –Technology is static State-level data used to estimate relationships that occur at the local level Small sample size in some surveys (e.g., RN Survey) reduces reliability of projections for smaller States and settings that employ fewer nurses

17. Physician Supply Model Overview

18. 18 National projections through 2020 For 36 medical specialties Activity Distribution (e.g., patient care, administrative, research, teaching, other) Physician age and sex IMGs, USMGs, CMGs MDs and DOs PSM Projection Dimensions

19. 19 Current age, sex, USMG/IMG/CMG, and specialty distribution of physician workforce (starting with 2000) Annual number of US and international medical graduates Recent patterns of specialty choice and activity distribution Physician retirement patterns Physician productivity (especially average hours worked) PSM Projection Determinants

20. 20 Physician Supply Model Overview STOCK GROWTH TOTAL SUPPLY: Specialty, Activity, Age, Sex, Edu.t+1 BASELINE PHYSICIAN SUPPLY t+1 ACTIVITY DISTRIBUTION Losses NEW PHYSICIAN SUPPLY t+1 BASELINE PHYSICIAN SUPPLY: Specialty Distribution Activity Distribution Age Sex IMGs, USMGs, CMGs MD’s and DO’s Losses MEDICAL SCHOOL GRADUATES: Age Sex PGY IMGs, USMGs, CMGs MDs and DOs SPECIALTY DISTRIBUTION

21. 21 American Medical Association Masterfile –To estimate demographics and specialty distribution of current MD workforce –To estimate physician activity distribution –To estimate physician retirement rates by age, sex, USMG/IMG –To estimate specialty distribution of new physicians by sex, USMG/IMG American Osteopathic Association Masterfile –To estimate demographics and specialty distribution of current DO workforce AMA Socioeconomic Monitoring Survey –To estimate variation in hours worked by physician age, sex, specialty PSM Primary Data Sources

22. 22 PSM Limitations Parts of the PSM are static –Retirement patterns vary by physician age, sex and USMG/IMG but are constant over time Rates reflect long-term historical patterns Rates do not change with economic conditions –Specialty choice reflects recent historical patterns, but does not reflect projected demand

23. Physician Demand Model (PDM) Overview

24. 24 National projections through 2020 For 18 medical specialties For specific population groups defined by age, sex, metropolitan/non-metro location PDM Projection Dimensions

25. 25 Population projections by age, sex, metro/non- metro location Projected insurance distribution Detailed physician-to-population ratios Physician/NPC substitution ratios (not used in baseline projections) Adjustments to demand projections made outside the model –Economic growth –Changes in patterns of care PDM Projection Determinants

26. 26 Physician Demand Model Overview PRACTITIONER REQUIREMENTS ESTIMATES Assigns 18 specialties and 12 NPC disciplines to each sex, age, metro/nonmetro and insurance category POPULATION PROJECTIONS --Sex, age --Metro/nonmetro --Insurance PRACTIONER STAFFING RATIOS --Physicians (18 specialties) --NPCs (12 disciplines) TRENDS --Sex, age --Metro/nonmetro --Insurance

27. 27 Census Bureau population projections Estimate utilization of physician services by patient characteristics, insurance status and delivery setting using: –National Ambulatory Medical Care Survey –National Hospital Ambulatory Medical Care Survey –National Inpatient Sample –National Nursing Home Survey –National Home and Health Survey PDM Primary Data Sources

28. 28 PDM Limitations Demand in base year (2000) is defined by current utilization and service delivery patterns –Assumes physician supply and demand in equilibrium in base year –Inefficiencies in current utilization and service delivery patterns extrapolated into the future Physician requirements do not reflect overlapping scope of practice (i.e., substitution) between medical specialties

29. Additional Research Needed

30. 30 Additional Research Needed: NSM Estimate relationship between number of new RN graduates and –Wages –Working conditions –Teaching capacity –Training costs Estimate relationship between RN activity status and –Wages –Working conditions RN cross-state migration patterns Add LPN component

31. 31 Additional Research Needed: NDM Model demand by skill/training Estimate employers’ propensity to substitute between RNs, LPNs and nurse aides based on changes in economic factors (e.g., wages) Impact of changing technology on nurse staffing patterns Identify additional determinants of demand for nurses and quantify their impact

32. 32 Additional Research Needed: PSM Physician retirement patterns –AMA data may underestimate retirement –Physician Worklife Survey (conducted by Sheps Center, UNC) finds higher retirement rates than suggested by AMA data Physician specialty choice –How do projected shortages/surpluses for particular specialties affect specialty choice for new physicians

33. 33 Additional Research Needed: PDM Alternative estimates of base year demand that do not necessarily assume equilibrium in the base year Impact of a growing supply of non- physician clinicians on demand for physician services Effect of economic growth on demand for physician services—and the implications for government policies/programs

34. Tim Dall Vice President Email: tim.dall@lewin.com Ph: (703)269-5743 Fax: (703)269-5503 www.lewin.comtim.dall@lewin.com