Modeling Health Care Policy Alternatives

RAND researchers suggest specific improvements to data, as well as creation of a communications forum for modelers and policy-makers, to increase the utility of models that analyze health policy options, in this essay from Health Services Research in 2020.

The authors examined two types of models commonly used for health policy evaluation: microsimulation and cell-based models. Microsimulations are useful for modeling complex systems; but, in order to be accurate, they need large amounts of data about agents, their relationships and impacts on each other. Microsimulations are strong models for health policy changes, including health insurance reform. Cell-simulations are coarser and can be used when large amounts of data are not available or necessary. The simplicity and transparency of cell models makes them attractive.

The authors make specific recommendations to strengthen health policy modeling:

  • Fund several studies that approach a key question in different ways. Then, the results from these diverse studies can be synthesized to provide more reliable causal parameters and ranges of effects on critical health policy issues.
  • Improve the longitudinal study of responses to behavioral changes by standardizing across nationally-representative studies, increasing sample sizes and specifically studying the behaviors of two key actors in health policy: employers and physicians.
  • Improve access to existing data sets that link agents (e.g., patients to providers, workers to employers.)
  • Create a forum for modelers, stakeholders and policy-makers to routinely communicate.

The authors note that even models with limitations can help policy-makers, as long as those limitations are understood and clearly communicated. But, according to the authors, given the scale of current health policy reform efforts, investing in the tools necessary to provide policy-makers with useful information should be a high priority.