Appropriateness: the next frontier

The health care systems of developed countries share common problems. Firstly, the explosion of costly medical technologies increasingly jeopardises our ability to give everybody all the care that would benefit them. And, secondly, the explosion in medical services has made it virtually impossible to remember the indications, complications, and costs of procedures and drugs - that is, to practise good medicine without additional help.

Studies of appropriateness underline the seriousness of these problems. By appropriate care I mean that for which the benefits exceed the risks by a wide enough margin to make it worth providing. If we could increase appropriate and decrease inappropriate care, the benefits to patients and society in terms of health and wealth would be enormous. Indeed, without methods to detect inappropriate care, society's ability to maintain universal insurance coverage may disappear.1

But how do you measure the appropriateness of care?2 Although the clinical literature is the place to start, it mostly concerns the efficacy of a procedure performed under ideal conditions and tells us little about what happens when the procedure is done under less than ideal conditions.3 Furthermore, research rarely includes outcome measures that are relevant to patients and practitioners - for example, effects on health status or function.4

To measure appropriateness, colleagues and I at the RAND Corporation and the University of California, Los Angeles, have developed an explicit method, beginning with a literature analysis that summarises what is known about a procedure's efficacy, effectiveness, indications, cost, and use. The next step is to develop a list of specific clinical indications based on that review. Using the list of indications and the literature review, a panel then rates appropriateness on a scale of 1 to 9. On the basis of these ratings and clinical data collected from medical records we can measure appropriateness in actual practice.

Some of the findings have been worrying. For example, among Americans aged over 65 being treated in the fee for service system, carotid endarterectomy was performed for reasons that were equivocal, at best, in two thirds of cases. The same was true for upper gastrointestinal endoscopy and coronary angiography in about one in four cases.5 In a randomly selected group of hospitals in the western United States the proportion of coronary artery bypass surgery that was inappropriate or equivocal varied among hospitals from 23% to 63%.6 In a study in the Trent region of Britain coronary angiography and coronary artery bypass operations were performed for inappropriate or equivocal reasons in about half of cases.7 And the same was true in the North West

Thames region for 60% of cholecystectomies - regardless of whether they were performed in the public or private sector.8 Studies of the underuse of necessary care are now beginning; their previous absence is disturbing.

Two important conclusions may be drawn from the research. Firstly, inappropriate care is there if you look for it, and, more importantly, restricting the volume of care by global budgeting or planning does not eliminate it.2 Secondly, the rates of less than appropriate care are too large to be ignored.

How could appropriateness ratings revolutionise health care? Purchasers could decide to buy services only from doctors who agreed to operate within guidelines based on appropriateness, or services that satisfied generally accepted criteria of appropriateness. Ratings could also be used to prevent the underuse of necessary care. Decisions about licensing and recertifying doctors could be partly based on audit of the appropriateness of the care provided.

As a doctor, I hope that doctors will use appropriateness guidelines to improve the care that they deliver. Suppose that, when a patient was being considered for one of the most common 100 or so procedures, this happened: the doctor and the patient would take a few minutes to enter into the office computer all the clinical data that were critical to determine whether the procedure should be done. In a few seconds the computer could produce an appropriateness rating, an analysis explaining the rating, and an indication of the basis for the rating (that is, mostly scientific literature or expert opinion). The patient and doctor could do their own sensitivity analysis (that is, examine with the aid of the computer how changes in the patient's symptoms, signs, or responses to treatment would alter the appropriateness of the procedure) and explore extenuating clinical circumstances.

Of course, the above process represents only a starting point. None the less, it is time to assure our patients that, before they are subjected to a procedure or denied its use, its appropriateness has been explicitly verified. Methods to assess appropriateness are available. It is time they were used by doctors to eliminate both underuse and overuse of clinical interventions.