Dynamic programming in healthcare

The goal of healthcare is often ambiguous. Of course, the goal is dynamic depending on the state of the individual. The optimization variable, presumably, is the current value of the individual. That value depends on the likely decisions the individual would make in the future at different points in time. The information available to the individual at those decision points are uncertain and the individual has to reach optimal decisions by dynamic programming from her anticipated end.

It is a complex problem. But the value of the individual is an important consideration in healthcare. Arbitrary treatment and end of life decisions will be unambiguously suboptimal for the individual and society. From a policy perspective, one has to ask if an objective and individual customized analysis and decision are possible. If in the status-quo, such decisions are arbitrary, then a systematic process can only improve the situation.

Individual’s perception of her own utility differs from others and it depends on a plethora of factors, including but not limited to pain, expectation of remaining horizon, arrival of major events in that horizon, and her environment, family and significant others. The individual, thus, has lot more information to value herself than any bystander. Anybody other than the individual, including all healthcare participants, manufacturers, payers, providers and regulators, appear ill-quipped to improve on the individual’s decision. Unfortunately, any ex-ante attempt by the individual to influence decisions without information on her future state cannot be shown to be optimal. It is a catch 22, prescriptions by the individual without knowing future state and decisions by others who has the future state, are both suboptimal.

A healthcare decision framework that optimizes decisions for the individual without her ability to make such decisions may be needed.