Philadelphia Reflections

Diagnosis Based Payment: A Warning

I was sitting in the Congressional hearing room when it happened. A proposal from the hospital association was made to Congress in 1983 that instead of paying hospitals for each step of treatment, they should be paid by the diagnosis, and Congress soon agreed to the idea for Medicare. This system was to be limited to helpless inpatients. The idea had some good features: if a patient had to be fed with a spoon, he had little interest in the cost of his treatments. Under these circumstances, market mechanisms would never restrain the cost of hospitalized patients. If they were anesthetized, it was even truer.

DRG: An Object Lesson for Control Freaks With Little Interest in What They Are Controlling.

The problem was to put a price on thousands, even millions, of diagnoses with enough difference in cost to warrant a code number. The fifty-year-old coding system of the AMA called Standard Nomenclature of Diseases and Operations, had room for a hundred million diagnosis codes, of which perhaps two million were in use. It was a useful classification in the days when hospitals and Natural Science museums were much alike, cataloging and classifying different objects. But record librarians were in a position to see what their main activity was really like. It was gathering the various pieces of previous admissions in order to be useful in managing a new episode for an individual. For this, a new code was less work for the record librarians: the International Classification of Diseases, which reduced the number of diagnoses in practical use to about a thousand. When greater detail was needed, it was simpler just to look it up, and record librarians knew very few doctors except pathologists actually did that. So hospitals were ordered by accrediting bodies to use ICD coding, and save administrative costs. After Congress made its payment decision, a committee was formed to cut the thousand down to two hundred with similar payments, and lo, the DRG (Diagnosis-Related Groups) was created. All hospital inpatients were assigned one of two hundred codes, within which the size of the payment was a dominant sorting feature. Before long, in-patients were accordingly charged one of two hundred prices. The pathologists objected and produced their own modernized coding system, SNOMed. Forget it, it was too much work, cost too much, it was too hard. (By the way, the number of DRG codes is already back up to a thousand.)

So in this way, by arranging the assignment of costs to codes, Medicare and the hospital coding clerks took over the job of pricing. No doctor understood what in the world they were doing. And by steps familiar to accountants, the DRG was enlarged back to a thousand codes and internally arranged to come out paying the hospital a 2% profit margin for inpatients. Since we were running a 3% inflation at the time, the effective push was on-- to move in-patients to the out-patient area. No matter how many tests, no matter how long the patient stayed, the DRG came out to produce a 2% profit margin. The cost the insurance had to pay was lessened, the costs the hospital actually incurred, became the hospital's problem. Meanwhile, interest rates were low, so new outpatient buildings seemed cheap. Pretty soon, hospitals were paying doctors above-market prices to fill the outpatient area. There's more to say, but the idea is clear. Once you find a rationing tool, the accountants are in charge, the doctors are out, and eventually would be really out. And the beauty part of it was, no one understood what was happening, or who did it. Except you will find a lot of empty out-patient buildings when the music stops.

Originally published: Wednesday, August 10, 2016; most-recently modified: Thursday, May 16, 2019