Research Information

The choice among many medical procedures, tests, and medicines is not clear cut. For example, one procedure compared with another may have lower cost but higher risk of side effects and lower success rate. The cost of medicine can be accomplished be improving the economic efficiency of medicine or medical rationing. Medical rationing can be accomplished by sustituting lower cost procedures for higher cost.

If society is to move towards increased medical efficiency, a careful analysis of medical alternatives needs to be performed. This requires an analysis of patient medical records to determine diagnosis, the risk factors of the patient, what was perscribed and the outcomes. It is important to point out that the identity of the patient is not required for such analysis.

Analysis of patient records falls under the area of nonexperimentally generated data. The value of such data is geatly increased if the researchers can obtain representative samples. This means that the value of such research is greatly decreased if permission to required to do the analysis. The cost is greatly increased and samples are not likely to be representative.

Many types of variations in medical practice need to be systematically analyzed to determine the cost-benefits of the alternatives. For example, one HMO in Boston has decided to use specially trained nurses as gatekeepers. Obviously such a policy greatly reduces the gatekeeper cost over having family practice specialists as gatekeepers. The impact of such a move on the health of the customers would require systematic analysis of patient records in the two alternative medical delivery systems over an extended time period.

Much research needs to be performed on the systematic analysis of risk adjusted outcomes. Given the various risk factors such as age, general health, and stage of an illness, how effective was the treatment.

How is information to be obtained. First, there is no need for labels which provides for privacy. However, the data must be carefully handled so that individual identities can not be inferred from the records. Analysis could be done at the sites of the medical records and only summary information sent forward. This would require a great deal of trust, but is the cheapest in the long run. The analysis modules would be intelligent agents with permission to search and analyse particular cases. Only summary statistics and not the patient medical files would be communicated from the computer of the medical care givers.

Such analysis is essential to provide analysis to resolve the debate between rationing and quality care at a reasonable cost.

\