Thursday, July 23, 2009

The President's Press Conference on Health

Professor Obama sees an inefficient health care system. Health is a prominent example of "household production theory". You do not have an innate desire to take blood pressure medication. Such a drug is an "input" into feeling good and not having a heart attack. The President has claimed that if doctors were compensated for keeping you healthy rather than being paid for by "fee for service", we would have a win-win of healthier people at a lower social cost.

How would this work? First, patients would need to be randomly assigned to doctors. If doctors could choose their patients, then doctors would cherry pick the healthy ones and would be paid a lot of money for "keeping them healthy" but we all know the counter-factual here. The healthy would be healthy even if Dr. Nick from the Simpsons' was their doctor.

Now don't forget transport economics. Doctors are located at a physical place and different neighborhoods of different racial and income levels are located at different distances from the Doctor. You don't have to be Hotelling to anticipate that Doctors in an inner city tough community will demand an "adjustment" because their patients will be sicker than the average. So, how will the Obama Team's pay to these doctors work? Will it be based on a "value added" criteria relative to a baseline? But who will determine what the baseline health of these residents is?

This is a similar issue when we reward teachers for student test scores. What is the right control group? The national average? The child's score at the start of the year?

Perhaps a solution is "Yardstick competition". Suppose that for each person, we knew five of his family members and five of his friends at work and in the residential community. If there were a set of observable indicators such as a cholesterol reading, one could calculate group averages and compare the individual's time series of cholesterol readings to these two peer groups. If the individual improves relative to the peers, the doctor team gets a bonus. This approach would control for a number of trend issues. Basic statistical techniques could be used to correct for age differences but the key issue is identifying a valid control group for identifying quality doctor "inputs" when they are delivered.

Swithing Subjects: Here is the future of UCLA. Like Elton John, He is a Rocket Man.