Thursday, January 7, 2010
Analytics and decision sciences generally rely on logic, though humans are not necessarily logical. That’s why computers are helpful, they are fundamentally logical, though incapable of creativity. In healthcare (and national security) we are drowning in information, thanks to the efficiencies and sensitivities of computers. And we are relying too much on them to provide answers they aren’t capable of generating. Here’s why.
We have so much data, we are forced to store it using logical systems for later analysis. Information about “Malbec”, an Argentine red wine, for example, could be stored under any of the following categories: Argentina, wine, red. And you could look in those places, or cross reference them, to find malbec mentions. “Claw hammers” is filed under “tools,” subsection “hand tools,” and perhaps a further subsection “hand tools-carpentry.”
Logical analytics are useful for identifying problems or weakness: 100% of carry-on luggage is screened, but only 10% of checked baggage is screened (I’m making those figures up, but someone knows them.) This can be found by simple measurement. Logically this tells you that checked baggage is a better choice for smuggling contraband. That’s the logical decision reached by analysis of known data. A computer could reach that same decision if provided the same variables. But a computer cannot tell you how to smuggle your contraband in checked baggage. That takes some level of creativity. Sick and illegal creativity, but creativity none-the-less. So creativity is used to exploit weaknesses identified with logical analysis. Operationalization is an inherently creative process. Foiling such threats thus requires creative, not logical, analysis.
Risk calculators can forecast your 10-year risk of heart attack. For example, a 55 year-old non- smoking male, with total cholesterol of 280 mg/dL, HDL of 60 mg/dL, systolic blood pressure of 100mm/Hg who is not taking blood pressure meds, faces a 7% risk of heart attack in the next 10 years. A computer calculates that risk using an algorithm based on clinical trial data. That’s a logical analysis using known data. Changing those factors, i.e. lowering blood pressure or increasing HDL by a certain amount, will reduce that risk by a predictable percentage. Again, its logic, based on analysis of known variables. But getting the person to change behavior in a way that results in those desired changes (operationalization) requires creativity. No computer can tell you how to do that. Once again, the magnitude of the problem (heart attack risk) can be established using logical analysis of data gathered through measurement, but the solution requires the human attributes of creativity.
In a world awash in data and health threats, we may be leaning too heavily on logic and not enough on creativity. Think of some ways that you could make yourself more happy, more likely to exercise, or to eat better. Sure, a computer can tell you how much you might reduce your diabetes risk by exercising one hour a day or eating 100 fewer grams of fat. But can a computer tell you how you will achieve that?
Share and Enjoy:
Posted by Ano Lobb at 11:29 PM