Cost-benefit analyses of California family practice residencies

Journal of Family Practice, May, 1989 by Paul G. Barnett, John E. Midtling, William H. Burnett, Franklyn D. Dornfest, J. Edward Hughell, Nornam B. Kahn, Fran S. Larsen

In its first report to Congress, the Council on Graduate Medical Education (COGME) found that despite an overall surplus of physicians, the United States has too few primary care physicians, and that family physicians are in especially short supply. [1] The council recommended that federal, state, and private support for family physician training be expanded.

Six previous national commissions found that primary care training programs cannot earn enough patient care income to meet expenses and recommended that these residencies be subsidized. [2-7] Neither these commissions nor COGME explicitly stated how significant this support must be. Such imprecision can be attributed to the dearth of studies documenting the cost of graduate medical education in primary care.

Conceptually, it is not difficult to understand why support is needed. Ambulatory-based training generates less revenue than programs in the procedurally oriented specialties, where hour for hour, physicians earn five to ten times as much as can be garnered from giving outpatient care. [8] Moreover, ambulatory programs collect a smaller percentage of charges, as third-party payers reimburse fully for hospitalization but require out-of-pocket contributions for outpatient visits.

The low reimbursement is compounded by the high cost of teaching in an outpatient setting. Much of hospital-based teaching occurs with the faculty physician teaching a group of residents in regularly scheduled rounds. In contrast, ambulatory patients are available for a short time each visit, and the faculty physician not only teaches a single resident at a time, but also must continue to be available throughout clinic hours. In short, ambulatory-based training is a more labor-intensive process. Furthermore, the resources needed to- teach behavioral sciences constitute an additional expense unique to family practice residency training. This paper describes the issues involved in quantifying the finances of family practice training and the methods used, and reviews published studies that have employed cost-allocation methods to evaluate family practice training programs. Also presented are financial data from eight residency programs as well as an estimate of the pure cost of education in one of these programs using a joint-products cost analysis.

GRADUATE MEDICAL EDUCATION COST-BENEFIT METHODS

Two methods have been used to study the revenues and expenses of graduate medical education. Each method has its appropriate use and its drawbacks.

The cost-allocation method is used to prepare hospital cost reports. It generates a number that purports to represent total educational cost. The costs of the hospital (or medical school) are assigned to mutually exclusive "products" of education, research, and patient care, using a time analysis of physician activities. Cost-allocation studies rarely have guidelines or criteria for the activity analysis, and physician time reports are often completed by an accountant or administrator, not the individual physician. Even when objective standards are established, cost allocation is inherently arbitrary in an enterprise such as a teaching hospital, where products are produced simultaneously. For example, there is no objective way to divide the cost of conducting hospital rounds between patient care and teaching. For these reasons, data generated in cost-allocation studies are quite variable and not reproducible.

The joint-products cost-allocation method recognizes that patient care and education are produced simultaneously, and that most costs of a teaching institution are attributable to both of these products. [9] Expenditures are divided into joint costs, the costs of activities in which products are simultaneously produced, and pure costs of each product, that is, costs that are strictly assignable to that product. The pure cost of education in a teaching hospital is the hospital's current costs less the estimated costs of delivering the same amount of patient care without a teaching program.

Under the joint-products cost method, the question of how to apportion the cost of rounds between patient care and resident teaching is abandoned as unanswerable. Instead, the analysis focuses on how the cost of rounds will be affected by changes in the number of patients being given care or by changes in the number of residents being taught.

As the joint-products cost method deals with hypothetical situations, it is not empirical. Results are highly dependent on the assumptions made by the analyst.

Whichever method is used, it must be recognized that residency programs are rarely independent businesses with their own accounting systems. They may be constituted as a department in a medical school, a cost center within a hospital, or as a freestanding nonprofit corporate practice. The activities of residency programs often occur at the intersection of several institutions. Residency finances may be accounted for not only by a hospital, but also by the medical school, its clinic, and, increasingly, a faculty practice plan as well. Two surveys of US family practice residencies show that residencies take many different forms. One survey found that only 21% of the programs were operated as a hospital cost center. [10] The second survey found that most (59.5%) residencies were a hospital cost center. [11] This difference may reflect sampling bias or differences in study methods, or it may genuinely reflect that more family practice programs have become hospital based to realize the Medicare payment for graduate medical education.