Optimization of Drug Administration by a Tauberan Approach

The problem of achieving a desired time profile of the used drug's blood level is studied. It is shown how the profile can be approximated as a weighted sum of responses to a single dose according to Tauber. The problem of attaining the desired profile is shown to be reducible to a classical parameter optimization problem. A solution to this problem is proposed which makes use of Prony's algorithm and dynamic programming for the optimization of the drug administration. The doses and the timings of the drug administration are optimized to result in a sensible compromise between the therapeutic effect and keeping the total amount of the administered drug small, bearing its side effects in mind. As an example of the application of this technique, levodopa therapy in Parkinson's disease is discussed.