Friday, October 31, 2014

The Dirty Little Secret about Analytics

John Cage used to talk frequently about how he would consult the I Ching to make choices for him when composing music. What I remember most is when he said that the I Ching never made a bad decision. If a decision turned out to be problematic (leading, for example, to music that was impossible for a single performer to play), then Cage took this as an indication that he had formulated his question poorly.

This is not just an anecdote about avant-garde practices. It may also tell us something about what happens when Microsoft, Google, and Apple all decide that they need "market share" in the health care industry. This is just the latest data point in the litany of evidence as to why health care is in such a sorry state because industry and business groups feel it is necessary to see to the "health" of their revenue streams by telling expertly-trained doctors and nurses how to treat patients. These days it is all about getting everyone to invest in wearables that will automatically transmit all necessary "health care data" to some cloud site where the resulting data base can enjoy all the benefits of being massaged by the best analytic software that can be dreamed up by analysts (most of whom, of course, are business analysts, rather than health care practitioners).

The thing is that such a data base is ultimately no different from the I Ching. The quality of the answer you get depends on the quality of the question you ask. This is particularly true in non-routine situations. Good doctors know when they are stumped, but they also know that they are stumped because they have not yet figured out the right question to ask. The best ones persist until they finally land on that right question.

The problem is that we live (for over half a century now) in a world in which the judgment of an algorithm is always taken as superior to human judgment. Indeed, we are so addicted to that believe that we cannot conceive of an alternative. This latest phase in the industrialization of health care is yet another step in pushing the most valuable expertise of practitioners out of the system. In other words the price of a higher revenue stream will be a lot more sick people, many of whom will probably turn out to be incurable.

1 comment:

DigitalDan said...

I will admit I enjoy the heart rate sensor built into my Moto 360 (as well as the barometer in my phone), but not because I want to share the information with anyone else (although I probably am anyway.) Guess they're welcome to it. The Moto's step counter, however, seems pretty bogus.