兰德-使用系统动力学对医生劳动力供应_需求和需求进行建模-模型结构和假设（英）-2025_39页_1mb

Modeling Physician Workforce Dynamics Using System Dynamics:

1. Purpose & Innovation

   • The RAND-AAMC model employs system dynamics (SD) to simulate physician and advanced practice provider (APP) workforce dynamics, incorporating causal feedback loops and delays.
   • Unlike traditional ratio-based models, this approach captures complex interactions (e.g., supply-demand-need feedbacks) and unintended policy consequences.

2. Key Model Components

   • Modules: Tracks supply (physicians, APPs), perceived clinical need, demand, and utilization.
   • Feedback Loops: Includes reinforcing/vicious cycles (e.g., supply-driven need increase) and balancing loops (e.g., utilization reducing perceived need).
   • Policy Levers: Supports interventions like task shifting, training slots, and APP scope adjustments.

3. Methodology & Data

   • Built using SD software (Stella Architect), with 720 equations, 30 stocks, and 64 flows.
   • Grounded in expert stakeholder discussions and empirical data from sources like AAMC, AANP, and MEPS.
   • Verification: Model calibrated to historical data (2002–2052 projections).

4. Strengths

   • Flexibility: Tests diverse policy scenarios (e.g., residency slots, burnout, payment dynamics).
   • Holistic View: Simultaneously analyzes supply, demand, utilization, and need across primary and specialty care.
   • User Interface: Publicly accessible for policymakers to explore scenarios dynamically.

5. Limitations

   • Excludes geographic details and clinician demographics (race, ethnicity).
   • Aggregates national-level data; state-specific models would require disaggregation.
   • Implicit rather than explicit: Factors like payment or salary influences.

6. Outcomes of Interest

   • Tracks 12 imbalances (e.g., clinical need vs. supply) and ratios (e.g., adequacy of supply).
   • Reports outcomes at granular levels (e.g., hourly supply-demand gaps).

7. Future Directions

   • Geographic disaggregation, integration of payment systems, burnout modeling, and state-level adaptations.
   • Exploration of policy impacts (e.g., ACA, COVID recovery).