Dealing with medical practice variations: a proposal for action

• Published in: Health Affairs 1984, Vol.3 (2), p.6-32
• Main Author: Wennberg, J E
• Format: Article
• Language: English
• Subjects: Administrators; Aged; Angina pectoris; Attitude of Health Personnel; Catchment Area (Health); Classified information; Coalitions; Community; Containment; Cost Control; Decisions; Diagnosis-Related Groups; Female; Financial Management, Hospital; Health care; Health care delivery; Health care expenditures; Health care policy; Health Services - statistics & numerical data; Health Services Misuse; Hospitals; Humans; Hyperplasia; Hysterectomy; Local population; Male; Markets; Medical personnel; Medical treatment; Medicare; Medicine; Middle Aged; Patient Admission; Patients; Per capita; Physicians; Planners; Prices; Prostate; Public health; Regulators; Resource allocation; United States; Variation
• ISSN: 0278-2715; 1544-5208
• DOI: 10.1377/hlthaff.3.2.6

[...]I recommend that the medical community make greater efforts to deal with the cost-containment problem by reducing the use of hospitals for marginally indicated conditions, as may be determined from the monitoring of medical practice called for above. Yet the data show that the most important determinant of variations in per capita costs, or the "bottom line" for payers, are physicians' decisions to admit patients to the hospital or to employ a specific treatment, not the decisions they or other health care providers may make that affect the efficiency of medical care as reflected in the unit price of service or the length of a hospital stay. Since more than 85 percent of hospitalizations classified under the DRG system appear to have greater variation in per capita use rates among hospital market areas than hysterectomy, the above example is a conservative demonstration of the problems that attend unit pricing approaches to cost containment. The way the markets are organized assures a close association between the medical care experience of the local population and decisions made by health planners, regulators, local administrators, hospital trustees, clinicians, and, potentially, business coalitions. Since information on resource allocation and service use rates is available from all relevant places where care is given (whether in- or out-of-area), the per capita rates are truly population-based and thus may be validly compared. Using claims pooled from all Blue Cross accounts, John Putnam of Maine Blue Cross has shown how that organization can provide very important information on variations. Because of its national coverage and the richness of its data base, the Medicare program offers the best immediate opportunity to implement feedback in all parts of the country.