Exploring the Cost of Undergraduate Medical Education

Exploring the Cost of Undergraduate Medical Education Background Medical students, public policy makers, and others look to the AAMC to explain the cost of medical education, especially during times when resources are constrained. It has been nearly fifteen years since the AAMC closely examined the cost of undergraduate medical education. Rapid enrollment growth, new medical schools and regional campuses, new technologies, and curricular changes raise new questions about the cost and efficiency of teaching medical students as do pressures on resources that fund medical education, rising medical student debt, and the impact of the national recession, national debt, and healthcare reform. The objective of this paper is to identify what is known about the cost of undergraduate medical education (UME) and to identify opportunities for further study. Defining the Cost of UME While everyone agrees that educating medical students is expensive, much of the recent discussion about costs relates to how much a student should pay for an M.D. degree. The debate is fueled by the impact of tuition increases and rising student debt on accessibility and student specialty choice. It is important to distinguish between the price students pay for their education and the cost of educating medical students. The term cost is frequently used in reference to tuition rates and cost of attendance (the amount a student pays for all education-related expenses). The cost of educating medical students in the U.S. is shared by students, government, private donors, and patients. According to one recent estimate, tuition covers about 20% of the annual cost to educate a medical student (Medical School Economics: The Dollars and Sense of Educating Physicians, University of Texas San Antonio, Winter 2010). There is no direct relationship between the cost of UME and tuition rates. Tuition is but one of the revenue sources that support the cost of UME, and tuition rates often fluctuate based on the capacity of a medical school to raise revenue from other sources. So, for example, when state appropriations or endowment earnings fall, university boards sometimes authorize tuition rate increases across disciplines to recover lost revenues. Because medical schools have multiple missions, there is no direct relationship between the size of a medical school and the cost of UME, nor is there a direct relationship between the cost of UME and the magnitude of financial resources at a given medical school. Attempts to relate medical school revenue sources to the cost of UME, such as dividing revenues by number of students, fail to recognize the breadth of medical school activities supported by medical school revenues. 1

In the context of this report, cost refers to the production costs associated with the resources required to educate students enrolled in academic programs that lead to an M.D. degree, regardless of where UME activities occur and who pays for those resources. Factors that Influence the Cost of UME Essential resources required of all M.D. degree programs include faculty, support staff, administrators, supplies, equipment, and space. Factors that affect the cost of these resources include curriculum design, educator mix, class size, faculty teaching loads, location, institutional culture, institutional development stage, organizational structure, government-imposed regulations, and accreditation standards. There is general agreement that educating medical students is more costly than educating students in other disciplines because UME is more labor intensive. The low student faculty ratio in clinical settings accounts for much of the cost of educating medical students, and UME has become more labor intensive in recent years as students moved out of lecture halls and into small group settings and clinics beginning in their first year of medical school. Most observers believe that curriculum reform has increased the cost of educating medical students; however, the impact of curriculum reform on UME costs depends not only on the number of hours faculty spend with medical students, but also on the mix of faculty and other educators involved in the curriculum. Differences in compensation rates for basic and clinical faculty, full-time and part-time faculty, and other educators involved in the curriculum affect UME costs. Although the extent to which a medical school is successful in negotiating pro bono services from affiliated hospitals and community physicians affects the amount a medical school pays for UME in a given year, services provided by volunteer faculty are a cost of UME, even if not covered by medical school funds. Isolating the Cost of UME Isolating the cost of UME can be a formidable task because much of UME teaching takes place in a clinical setting, where faculty members are simultaneously training medical students and graduate medical residents while caring for patients. Differences in methods and assumptions used to attribute labor costs across activities can result in significant differences in the imputed cost of UME. To a lesser extent, methods used to allocate administrative, staff, library, student services, and infrastructure expenses across multiple activities also affect costs. Although the challenge of allocating costs across multiple activities is not unique to UME and permeates cost studies throughout higher education, there is no agreement on how to best apportion costs across joint activities. In its 1996 report, the AAMC Task Report on Medical School Financing recognized that medical education is best conducted in an environment that provides for and demands ongoing faculty scholarship, and that promotes excellence and innovation in patient care. The Task Force took the view that limiting costs only to resources that can be directly linked to teaching represents a narrow and incomplete representation of the cost of medical student education. Methods to allocate a portion of environmental or milieu costs to UME generally involve examining faculty activity profiles to determine the amount of time a typical faculty member should spend on research, clinical care, and scholarly activities in order to be an effective teacher. Because 2

there is no agreed-upon standard or guideline for allocating milieu costs, many cost studies ignore them altogether. UME Cost Studies A number of studies have attempted to isolate the recurring annual costs of delivering UME programs in the U.S. Almost all of these studies use some variation of activity-based costing to impute the cost of resources associated with UME, and all make assumptions about how to allocate costs across joint activities. The most extensive literature review of UME cost studies was conducted by Jones and Korn. Their findings, together with the results of several cost studies done in the mid-90s, were published in the March 1997 issue of Academic Medicine (Jones and Korn, Academic Medicine, Vol. 72, No.3/March 1997, pp. 200-210). In order to compare cost per medical student, they divided studies into two categories. Cost estimates that included milieu costs were labeled total educational resource costs. Cost estimates that limited costs to those directly related to UME, including a portion of infrastructure and administrative costs that could be attributed to UME, were labeled instructional costs. Based on their review, Jones and Korn determined that estimates of the annual instructional cost over a period of two decades fell within a remarkably narrow range of $40,000 to $50,000 per medical student per year (1996 dollars). Total educational resource costs showed greater variability ($72,000 to $93,000 per student per year), due largely to assumptions about faculty workload, methodologies used to allocate joint costs across activities, and judgments about milieu costs. Two of the most important UME cost studies reviewed by Jones and Korn were published in 1974 and based on a collaborative effort of the AAMC and the Institute of Medicine of the National Academy of Sciences (IOM). The AAMC and IOM recognized that instructional costs alone do not represent the full cost of UME, and they convened a panel of medical educators to analyze faculty effort reports and develop a profile for a faculty member fully involved in education. This faculty profile was used to estimate the cost of research and clinical activity components of the education program leading to the M.D. degree (AAMC Proceedings and Annual Report for 1974, p. 541). Faculty activity profiles are a key component of program costs analysis construction (PCAC), a method used in the 1970s to develop medical education funding guidelines in New Jersey. One of the studies reviewed by Jones and Korn used PCAC methods refined by investigators in Canada to predict UME resource requirements using faculty profiles (Valberg LS, et al, ISBN 0-920169-51-1). Workload expectations used when developing faculty profiles can have a significant effect on cost per medical student. Variations of PCAC were used to estimate UME costs in several of the studies reviewed by Jones and Korn and have been widely used to analyze costs in other health professions programs. Although there have been a number of published cost studies since 1996 that focus on aspects of UME (e.g., ambulatory care, problem-based learning), a review of the literature revealed only one new study conducted in the U.S. that focuses on annual costs of the UME program. This 3

study uses the PCAC method to compare UME costs over time. Using assumptions applied in their original study, investigators showed that UME costs at UT Houston in 2007 fell 9% compared to 1995 (E. Gammon and L. Franzini, Revisiting the Cost of Medical Student Education: A Measure of the Experience of UT Medical School-Houston, J Health Care Finance 2011; 37(3):72 86). The drop in costs was attributed to a reduction in educator contact hours (31% in the first two years and 15% in the last two years of the curriculum) and a shift in the curriculum from clinical sciences to less expensive basic sciences faculty. According to the report, faculty-intensive classroom instruction that required clinical faculty was reduced over the 12-year period in order to maximize billable clinic hours. Of particular note was a 28% reduction in hours taught by graduate medical residents attributed to the mandated cap on resident work hours. The drop in costs coupled with a 9% medical student enrollment increase resulted in a 16% drop in the cost per student. Other Approaches and Methods Although the UME cost studies reviewed used activity-based costing to assess recurring annual costs, there are other ways to approach a cost analysis that are useful in certain circumstances. For example, a marginal cost approach can help to evaluate the need for additional resources when class size is increased whereas an approach that assesses both capital investment and recurring costs is required when planning a new campus. Regardless of the approach used, an analysis of costs should begin with a clear statement of purpose. As indicated in the 1996 Task Force Report on Medical School Financing, methods to estimate instructional costs require first and foremost an accurate analysis of faculty time devoted to teaching medical students and associated activities. Because an accurate curriculum inventory forms the backbone of a UME cost analysis, tools developed to support mission-based management systems may be useful in determining the cost of UME. One study, published in 2002, identified 41 medical schools or medical school departments that used systems to track the educational activities of faculty members (Mallon, W.T. and Jones, R.F., Academic Medicine, February 2002). These systems used either contact hours as the unit of analysis or a relativevalue system that weighted teaching activities similar to the resource-based relative-value scale used for clinical activities. These systems, developed since the mid-90s, could form the basis of more accurate estimates of UME costs. Discussion There has been no seminal work on the annual cost of UME since the mid-90s. In recent years, requests for UME cost data have come primarily from state policymakers involved in developing and revising UME funding formula, but there is increasing demand from students struggling to pay their share of costs and teaching hospitals wrestling with impending changes of national health care reform. Since the recession of 2008, medical schools have felt increasing pressure on resources that support UME. New data on the cost of UME will help medical school leaders support the need for additional resources. 4

Just before the national recession of 2008, medical education entered a period of rapid expansion in response to predicted shortfalls in the physician workforce and the AAMC s goal to increase medical student enrollment 30% by 2015. Although the LCME carefully evaluates the resources of each new medical school, there has been no assessment of expansion costs across the academy. Data on the cost of UME could prove valuable for predicting resource requirements at new and developing schools. There is a perception that students are paying more than their fair share of costs, that students are subsidizing the cost of research, that medical schools have become bloated by administrators, and that the cost of educating medical students at allopathic schools is significantly higher than at osteopathic schools. A recent report argued that the cost of educating students at Michigan State University in 2009 was $19,730 per DO student per year compared to $60,000 per MD student and that the excessive difference in cost is due to the LCME infrastructure-related requirement. (T. DeLengocky, DO and B Bartoz, A Cost Analysis of the Proposed MD Program at UNTHSC: Spending More and Getting Less, Report submitted to Texas Osteopathic Medical Association for the 82nd Texas Legislature to review); however, the cost argument was made based on comparing operating budgets, not on an analysis of costs. There is no recent empirical cost data on which to base the facts. A number of new medical schools have developed innovative organizational structures and unique arrangements with hospitals and community physicians while other medical schools have recently reorganized long-established department and administrative structures and consolidated clinical services. The impact of efforts to trim budgets on the cost of UME is not known. Because there is no standard methodology for isolating the cost of UME, cost studies have heavily relied on assumptions about student faculty ratios, expected faculty workload, and faculty profiles that are both debatable and dated. The lack of standards and guidelines for apportioning costs across joint activities and for allocating milieu costs makes it difficult to compare cost study results across medical schools. Directions for Further Study While this paper offers some insights into the challenges associated with estimating the cost of educating a medical student, there are many unanswered questions that may be worthy of exploration such as What is the best methodology for estimating the cost of UME? How have UME costs changed over the past 15 years as a result of o curriculum reform o mandated cap on resident work hours o university expansions and debt obligations o medical school restructuring o payments to affiliated hospitals for teaching medical students What is the share of UME costs paid by medical students? 5

How does the cost of training medical students impact the cost of healthcare? How do current pressures on revenue streams impact expansion goals? Are there more cost-efficient ways to deliver UME? What are the appropriate outcome measures that can be used to evaluate costs? How do costs of UME at allopathic medical schools differ from those of other schools (e.g., osteopathic schools, offshore schools, Canadian schools), and what accounts for cost differences? Conclusion A theory introduced by Howard Bowen in the late 1970s, known as the revenue theory of costs, concluded that the only way to control higher education costs is to limit revenues. Some would argue that this is true for medical schools. However, there is clearly a need to understand resource requirements and costs if medical schools are to respond to pressures on the revenue sources that have sustained UME programs. The AAMC can assist in bringing the academic community together to identify best practices in isolating the cost of UME. This might include developing standards or guidelines for apportioning costs across joint activities, allocating milieu costs, and faculty activity profiles. The AAMC can also assist by articulating the difference between the cost of UME and tuition rates. In 1972, when the Institute of Medicine (IOM) decided to undertake a study on the costs of medical education, an article in Science stated that it could hardly have found a more difficult, conspicuous, and potentially controversial assignment (Science 2 June 1972:997-999). The same would be said today. 6