I agree completely.The editorial should not have been published.Andy Grove is obviously a brillant,accomplished former CEO of a big time high tech company but his comments suggest he has no concept of the medical profession or what should be done to improve health care.I had planned to pen a post critical of the editorial but Dr.Poses was quick on the draw.

I haven’t read the editorial and I’m not a doctor, but I would think carefully about anything Andy Grove has to say.

“In the chip industry, one can control almost all the variables.”

Wouldn’t that be nice. When a new chip factory is put together, nothing works, and nobody knows why. Every step of the process is intentionally up against the limits set by the laws of physics. It’s black magic and voodoo all the way. What they do is measure every aspect of the outcomes and procedures. That data is analyzed and the manufacturing procedures are ruthlessly refined. Ego and politics are strongly frowned upon, because the company is losing a couple of million dollars EVERY DAY the plant isn’t selling chips. Gradually the factory starts to make more and more working chips. If they do a good job, the plant becomes profitable. Refinement continues forever, to cut costs and improve chip performance.

The essential thing is the closed loop feedback cycle: information -> analysis -> improvement -> information -> and so forth. If the plant were, for instance, dispensing an incorrect chemical 0.01% of the time, this would be discovered and fixed in the normal course of business; if the boss discovered on his own that this problem existed, there would be hell to pay. If a process engineer discovered that a $50,000 machine would be cost effective, the purchase order would be signed ASAP.

Compare this to hospitals. Very little is measured. What is measured is locked away in unanalyzable handwriting. Budgeting concentrates on costs and tends to ignore returns. PR tends to ignore returns too, so there is little competitive advantage. Epidemiology is considered exotic. Oddball drug interactions and rare effects are invisible for lack of computer-readable data.

“No one yet knows how to store, for example, the contents of the medical history in anything other than a text file.”

How silly. Diagnoses have long been encoded with computer-readable numbers. Drug regimens and analytical test results are equally trivial to store in a computer. Likewise for number of visits, treatment cost, time of day of procedures, number of days of hospitalization, specialist types consulted, analog scales for symptoms, number of days off work, a simple disability metric, and so forth. Even limited data like this is an statistical gold mine that can drive process improvements. But if you don’t measure it, you’re flying blind.