Dynamic Microsimulation of Health Care Demand, Health Care Finance and the Economic Impact of Health Behavior. Part I: Background and a Comparison with Cell-Based Models
Cell-based health care models, as well as macro-level projections of future population and economic trends used as input to health care models, are limited to a few variables, which makes microsimulation an interesting modeling option, especially as it allows for modeling of the interaction of demographic with social, environmental and economic variables. Micro-approaches can incorporate the wealth of substantive analysis gained from a large number of micro- and macro-level studies with regard to demographic, economic and health behavior. Compared to cell-based macro models, microsimulation can produce useful projections for the analysis of different health-related phenomena considering additional dimensions, i.e., detailed issues regarding health care finance (insurance schemes, individual accounts etc.) and individual risk exposure. This paper constitutes the first part of an investigation of the potential of dynamic microsimulation for the modeling and projection of health care demand, health care finance and the economic impact of health behavior. The main purpose of this part is to provide a brief theoretical background with regard to the dynamic microsimulation approach and a comparison of the microsimulation approach with the cell-based macro approach. Starting with a definition of dynamic microsimulation and a classification of the types and approaches, microsimulation modeling is brought into the context of the life-course paradigm. This paradigm, meanwhile being the dominant paradigm in demography, can also be a useful organizational principle for the study and projection of health-related phenomena using microsimulation. Microsimulation is then compared with cell-based approaches, and the potential strengths as well as drawbacks of the microsimulation approach with regard to health care modeling are investigated. Dynamic microsimulation might turn out to be increasingly appropriate as a modeling approach in this field, which is currently dominated by cell-based macro-models.