Modus operandi

Recent research at the Dartmouth Medical School has produced a startling finding: The type and frequency of medical treatments vary widely from one area to another, not because of illness factors in the population, or because of access to medical care, but because of major differences in the way physicians practice medicine. In other words, the range of variation reflects the amount of "uncertainty" within the profession, the fact that there is widespread disagree,ment among physicians regarding the best way to diagnose and treat common conditions.

These findings are the result of recent research by Dr. John E. Wennberg, an epidemiologist at the Medical School. Wennberg arrived at these conclusions using a research technique known as small area analysis. He has used this data-based technique to study variations in the way medical care is used in hospital service areas in New England. His earliest studies were in Vermont, where among different hospital service areas extensive variation was found in per capita expenditures, per capita numbers of beds, and per capita surgery rates. The medical causes of admission showed even greater variation than surgical ones.

What accounts for such differences? Wennberg determined that the varia variations could not be explained by differences in consumer characteristics. It became clear, however, that there was a strong relationship between the specialists practicing in a given area and the patterns of hospital use: the greater the number of specialists, the greater the use of hospitals; the more surgeons, the more surgery; the more internists, the more diagnostic tests.

What was most dramatic, however, were results which suggested that individual surgeons' differences in judgement play a key role in deciding which types of surgery particularly the highly variable "elective" ones are used. While some areas had similar overall ratesof Surgery, these were often due to different types of operations being performed. (This characteristic, observed over time, is known as an area's "medical care signature.") It appears, then, that doctors in different areas treating similar populations practice medicine very differently.

At a recent meeting of the National Academy of Sciences in Washington, Wennberg indicated that 85 percent of medical procedures vary more than threefold from one region to another. His observations have led him to propose the professional uncertainty theory: The variations are due to the different judgments made by different physicians when confronted by a similar problem. Why this professional uncertainty? Because often there seems to be no underlying consensus among physicians on the use of a given procedure.

These findings are disturbing to the medical profession. Dr. Frederick C. Robbins, president of the National Academy of Science, stated: "You can cover it up all you want, but it looks bad, and it looks bad because it is bad. It is not an appropriate way for a profession to behave." So what's to be done? Wennberg sets forth two main problems to be confronted: medical outcomes and cost containment. Each calls for different solutions.

Regarding medical outcomes, physicians must first be made aware of the variations, the reasons behind them, and what can be done to reduce them. Wennberg is working on a pilot project to address this issue with physicians in Maine. For example, high hysterectomy rates in one area of Maine dropped after information on variations was provided to physicians.

In terms of cost containment, Wennberg does not believe that a reduction in rates for a few procedures would make much of a dent in overall costs. Awareness of the highly variable causes for hospital admissions will lead to cost redactions only when efforts to reduce hospital use are broad-based and when consensus is reached on which procedures can safely be done outside hospitals. The numbers of hospital employees and beds can then be reduced to fit the lower demand.

Wennberg's findings are currently receiving widespread attention. Corporations, public officials, and consumer advocates view small area analysis as a potentially powerful tool that not only can help control costs but can improve the quality of health care as well.