Financing graduate medical education.

The direct costs of residency training in the United States are over $1 billion per year. These educational programs have been organized predominantly around hospital services and supported by hospital revenues. Pressure has been increasing to reduce the rate of increase in hospital expenditures or costs or both. This article describes alternative methods for financing graduate medical education. Debate over the current sources of financing reveals several troublesome issues: the presence of residents allegedly decreases the productivity of professionals and leads to overusage of ancillary services, proposed methods to pay for faculty salaries and services have created confusion and concern, and the financing of ambulatory-care training has been insufficient and poorly coordinated. The medical-education community must resolve these professional and educational problems so that financing issues can be debated and properly defended.     

Adapting industry-style business model to academia in a system of Performance-based Incentive Compensation.

Performance-Based Incentive Compensation (PBIC) plans currently prevail throughout industry and have repeatedly demonstrated effectiveness as powerful motivational tools for attracting and retaining top talent, enhancing key indicators, increasing employee productivity, and, ultimately, enhancing mission-based parameters. The University of Arkansas for Medical Sciences (UAMS) College of Medicine introduced its PBIC plan to further the transition of the college to a high-performing academic and clinical enterprise. A forward-thinking compensation plan was progressively implemented during a three-year period.After the introduction of an aggressive five-year vision plan in 2002, the college introduced a PBIC plan designed to ensure the retention and recruitment of high-quality faculty through the use of uncapped salaries that reflect each faculty member's clinical, research, and education duties. The PBIC plan was introduced with broad, schoolwide principles adaptable to each department and purposely flexible to allow for tailor-made algorithms to fit the specific approaches required by individual departments.As of July 2006, the college had begun to reap a variety of short-term benefits from Phase I of its PBIC program, including increases in revenue and faculty salaries, and increased faculty morale and satisfaction.Successful implementation of a PBIC plan depends on a host of factors, including the development of a process for evaluating performance that is considered fair and reliable to the entire faculty. The college has become more efficient and effective by adopting such a program, which has helped it to increase overall productivity. The PBIC program continues to challenge our faculty members to attain their highest potential while rewarding them accordingly.     

Timing of revenue streams from newly recruited faculty: implications for faculty retention.

PURPOSE: To determine the timing and magnitude of revenues generated by newly recruited faculty, to facilitate configuration of recruitment packages appropriately matched to expected financial returns. METHOD: The aggregate of all positive cash flows to central college of medicine administration -- from research, clinical care, tuition, philanthropy, and royalties and patents, from all faculty newly recruited to the University of Arizona College of Medicine between 1998 and 2004 -- was quantified using the net present value (npv) methodology, which incorporates the time value of money. RESULTS: Tenure-track faculty and, in particular, those with laboratory research programs, generated the highest positive central cash flows. The npv for positive cash flows (npv[+]) during 6 and 10 years for newly recruited assistant professors with laboratory research programs were $118,600 and $255,400, respectively, and, for professors with laboratory research programs, $172,600 and $298,000, respectively (associate professors were not analyzed because of limited numbers). Faculty whose appointments at the University of Arizona College of Medicine exceeded 15 years in duration were the most productive in central revenue generation, far in excess of their numbers proportionate to the total. CONCLUSIONS: The results emphasize the critical importance of faculty retention, because even those newly recruited faculty who are most successful in central revenue generation (tenure track with laboratory research programs) must be retained for periods well in excess of 10 years to recoup the initial central investment required for their recruitment.     

Developing a new faculty practice plan with a model for funds flow between the hospital and the plan.

The majority of academic health centers are experiencing significant difficulties balancing their research, teaching, and clinical missions while maintaining adequate financial performance. One of the major areas under intense scrutiny is the specific financial relationship between the hospital and the full-time faculty. A realignment of the funds flow between these two entities is becoming essential to the future viability and ultimate survival of many health systems. The authors describe a model that evolved as part of the integration of the faculty practice plans of their institutions when they merged into a single health system, and that provides a framework that specifically addresses these issues of funds flow. The model includes (1) a strong partnership between the hospital and the full-time faculty; (2) a governance model of chairmen, faculty, and administration; (3) flexibility for the department chairs to set salaries and make significant financial decisions relative to their departmental performances; (4) a specific formula for funds flow for graduate medical education dollars from the hospital to the clinical departments; (5) local front-end charge capture and back-end central collections; and (6) clear and consistent definitions of revenue and expense items for both partners.     

Implementing a simpler approach to mission-based planning in a medical school.

Changes in the education, research, and health care environments have had a major impact on the way in which medical schools fulfill their missions, and mission-based management approaches have been suggested to link the financial information of mission costs and revenues with measures of mission activity and productivity. The authors describe a simpler system, termed Mission-Aligned Planning (MAP), and its development and implementation, during fiscal years 2002 and 2003, at the School of Medicine at the University of Texas Health Science Center at San Antonio, Texas. The MAP system merges financial measures and activity measures to allow a broad understanding of the mission activities, to facilitate strategic planning at the school and departmental levels. During the two fiscal years mentioned above, faculty of the school of medicine reported their annual hours spent in the four missions of teaching, research, clinical care, and administration and service in a survey designed by the faculty. A financial profit or loss in each mission was determined for each department by allocation of all departmental expenses and revenues to each mission. Faculty expenses (and related expenses) were allocated to the missions based on the percentage of faculty effort in each mission. This information was correlated with objective measures of mission activities. The assessment of activity allowed a better understanding of the real costs of mission activities by linking salary costs, assumed to be related to faculty time, to the missions. This was a basis for strategic planning and for allocation of institutional resources.     

Improving clinical productivity in the academic setting: a novel incentive plan based on utility theory.

PURPOSE: Academic internal medicine practices face growing challenges to financial viability due to high overhead, competing institutional missions, and suboptimal physician productivity. The authors describe the development of a clinical incentive plan for a group of academic subspecialty physicians at the Dana Clinic, an outpatient setting at Yale School of Medicine, and report on results of the first year's experience under the plan. METHOD: Utility theory was used to assess the risk profile of clinic faculty and identify incentive payments that would optimize faculty benefit or "utility" while minimizing departmental costs. Under the plan, physicians who reached a productivity target based on work Relative Value Units (wRVUs) between October 2003 and November 2004 had overhead costs covered and received a fixed payment to support salary; additional incentive payments were available for those exceeding the target. Physicians failing to reach the target were responsible for their own overhead costs and received no fixed payment. Physician productivity as measured by wRVU per full-time equivalent (FTE) was compared for the year prior to, and the year following, incentive plan introduction. RESULTS: Forty-seven members of eight academic sections were included in the analysis. Median productivity improved by 34%, with 42 of 47 physicians showing improvement. Significant improvements were also noted in collections (62%) and visit volume (23%), and shifts were observed in coding patterns. CONCLUSIONS: The unique threshold-based structure of the incentive plan, as determined through utility theory modeling, as well as permitting physicians to choose how to achieve the wRVU target were key features of its success, resulting in improved productivity without increasing practice resources or faculty salaries.     

The cost of medical education in an ambulatory neurology clinic

Decreased revenue from clinical services has required academic hospitals and physicians to improve productivity. Medical student education may be a significant hindrance to increased productivity and income. This study quantifies the amount of time spent by faculty members teaching medical students in an ambulatory neurology clinic as well as the amount of time students occupied rooms when seeing patients on their own. Over a three-week period in an ambulatory neurology clinic, an observer noted these quantities of time, and the opportunity costs of both amounts of time were determined. Attending physicians spent an average of 19.6 minutes per medical student per half-day teaching, which translates to an average cost of $20.78 per half-day clinic. Students spent an average of 49.9 minutes per half-day seeing patients in the absence of an attending physician, an opportunity cost to the clinic of $142.50 per student per half-day.     

Analysis of downstream revenue to an academic medical center from a primary care network.

PURPOSE: Many academic medical centers (i.e., teaching hospitals) have established primary care networks for not only assuring a referral base but also for educating students in the primary care setting. Such networks generally are not profitable when analyzed on an individual facility basis. However, revenues generated at the medical center in terms of inpatient admissions, laboratory testing etc., usually are much larger than generated on site. In this study, the downstream revenue from 18 practice sites was evaluated at The Ohio State University Medical Center. METHOD: Revenues in fiscal year July 1, 2003, to June 30, 2004, were broken down into four streams, including inpatient and outpatient charges and collections for both network and specialist physicians. A fifth stream evaluated specialist professional fees. The authors developed a novel conservative weighting system to capture the concept that not all revenues generated from network patients were actually dependent on the use of the network. RESULTS: Findings included that the downstream direct contribution margin of US dollars 14 million just from the admissions and outpatient tests and procedures directly generated by network physicians alone was nearly twice the US dollars 8.3 million network operating loss. The total downstream net revenue of nearly US dollars 115 million was more than 6 times the US dollars 18.9 million net revenue to the network. The downstream direct contribution margin of US dollars 52 million was 6.3 times the network loss. Total downstream gross revenue (charges) to the medical center was over US dollars 250 million and over US dollars 300 million when the specialist gross revenues were included. CONCLUSIONS: This study demonstrates that a primary care network can generate significant financial support for an academic medical center.     

Mission-based budgeting: removing a graveyard.

Many activities in today's medical schools no longer have medical students' education as their central reason for existence. Faculty are hired primarily to provide clinical service or to make discoveries, with the role of educator of secondary importance. Budgeting in medical schools has not evolved in concert with these changing roles of faculty. The cost of medical students' education is still calculated as if all faculty were hired primarily to teach medical students and their other activities were to support this "central" mission. Most medical schools still mix revenues without regard to intent and cannot accurately determine costs because they confuse expenses with costs. At the University of Florida College of Medicine, a group of administrators, chairpersons, and faculty developed a budgeting process now called mission-based budgeting. This is a three-step process: (1) revenues are prospectively identified for each mission and then aligned with intended purposes; (2) faculty productivity, i.e., faculty effort and its quality, is measured for each of the missions; and (3) productivity is linked to the prospective budget for each mission. This process allows the institution to understand the intent of its revenues, to measure how productive its faculty are, to learn the true costs of its missions, to make wise investment decisions (subsidies), and to justify to various constituents its use of revenues. The authors describe this process, focusing particularly on methods used to develop a comprehensive database for assessment of faculty productivity in education.     

Financial impact of physician assistants on medical practice.

A study of revenues generated and expenses incurred by 12 physician assistants (MEDEX) who had held salaried positions for at least one year was conducted to determine their financial impact on primary-care practices. Daily charge logs were used to make annual estimates of MEDEX-generated revenues. One method of estimating annual revenues produced a mean of $28,190 per year, and a second method yielded a figure of $30,210 per year. Financial statements were used in two different ways to estimate annual expenses related to the employment of the MEDEX. One procedure indicated average costs of employment were $15,900, and the other $20,100. Ten of the 12 practices in the study experienced substantial gains of estimated revenue over expenses ascribed to the activities of the MEDEX.     

Contribution of clinical trials to gross domestic product in Hungary

Aim

To determine the contribution of clinical trials to the gross domestic product (GDP) in Hungary.

Methods

An anonymous survey of pharmaceutical companies and clinical research organizations (CROs) was conducted to estimate their clinical trial-related employment and revenues. Clinical trial documents at the National Institute of Pharmacy (NIP) were analyzed to estimate trial-related revenues at health care institutions and the value of investigational medical products (IMPs) based on avoided drug costs. Financial benefits were calculated as 2010 US $ purchasing power parity (PPP) values.

Results

Clinical trials increased the revenue of Hungarian health care providers by US $165.6 million. The value of IMPs was US $67.0 million. Clinical trial operation and management activities generated 900 jobs and US $166.9 million in revenue among CROs and pharmaceutical companies.

Conclusions

The contribution of clinical trials to the Hungarian GDP in 2010 amounted to 0.2%. Participation in international clinical trials may result in health, financial, and intangible benefits that contribute to the sustainability of health care systems, especially in countries with severe resource constraints. Although a conservative approach was employed to estimate the economic benefits of clinical trials, further research is necessary to improve the generalizability of our findings.Active participation in international clinical trials may provide health benefits to patients and financial and professional benefits to health care providers. In lower income economies, such as those in Central-Eastern Europe (CEE), the relative benefits of clinical trials are even greater than in the high income countries of Western Europe and North America. Consequently, the contribution of emerging markets to international clinical trials is growing substantially (1). This phenomenon is especially visible in CEE, where the number of clinical trials has increased significantly over the past 15 years and is expected to increase even further in the near future (2). In CEE, international clinical trials offer opportunities for site personnel to improve their professional networking and be remunerated on higher-than average income level. For health care institutions with substantial budget constraints, trial-related payments can represent an important source of liquid cash. A supportive attitude of hospital management toward clinical trial activities, in terms of providing better working environment or increased remuneration, may help to prevent the migration of qualified professional staff to higher income countries. In CEE countries, the health status of the population is worse than in higher income Western European countries (3) and the accessibility of new medicines is relatively limited (4). Therefore, through clinical trials, CEE patients can obtain access to standardized modern health care services, technologies, and investigational drugs without waiting lists or co-payments. However, investigational medical products (IMPs) may represent considerable health risks for patients.The societal gain associated with clinical trials is multifactorial. Clinical trials contribute to the evolution of evidence-based medicine. They systematically investigate side effects and health outcomes not only for IMPs but also for the control treatment arms. Therefore, safety information, even about marketed therapies, is captured and no public investment is necessary.The most tangible benefit may be the financial impact, including the contribution of trials to the revenues of health care providers and clinical research organizations (CROs). However, there are also indirect benefits, such as avoided health care expenses due to the free delivery of IMPs and services.Few scientific publications have addressed the financial benefits of clinical trials. These publications examined avoided drug costs and additional revenues primarily from the viewpoint of health care institutions (5-9). There is also one Polish study on the national economic impact of clinical trials, but the approach was not comprehensive enough to capture all direct and indirect financial benefits (10).Hungary currently has a favorable position for implementation of clinical trials (11). It has high-level professionalism at investigational centers, rapid regulatory and ethical endorsements of applications, complex but manageable contracting processes at clinical sites, sufficient contributions to patient recruitment, and high Good Clinical Practice (GCP) quality according to Food and Drug Administration (FDA) inspections (12). However, similarly to other CEE countries, the capacity for clinical trial participation in Hungary has not been maximized. The aim of this study was to determine the contribution of clinical trials to the national economy in Hungary. We estimated the clinical trial-related revenues of CROs, investigators, and health care institutions and the financial benefits of avoided drug costs due to IMPs as the percentage of the gross domestic product (GDP).     

The economic cost of sleep disorders

STUDY OBJECTIVES: To determine the economic cost of sleep disorders in Australia and relate these to likely costs in similar economies. DESIGN AND SETTING: Analysis of direct and indirect costs for 2004 of sleep disorders and the fractions of other health impacts attributable to sleep disorders, using data derived from national databases (including the Australian Institute of Health and Welfare and the Australian Bureau of Statistics). MEASUREMENTS: Direct health costs of sleep disorders (principally, obstructive sleep apnea, insomnia, and periodic limb movement disorder) and of associated conditions; indirect financial costs of associated work-related accidents, motor vehicle accidents, and other productivity losses; and nonfinancial costs of burden of disease. These were expressed in US dollars (dollar). RESULTS: The overall cost of sleep disorders in Australia in 2004 (population: 20.1 million) was dollar 7494 million. This comprised direct health costs of dollar 146 million for sleep disorders and dollar 313 million for associated conditions, dollar 1956 million for work-related injuries associated with sleep disorders (net of health costs), dollar 808 million for private motor vehicle accidents (net of health costs), dollar 1201 million for other productivity losses, dollar 100 million for the real costs associated with raising alternative taxation revenue, and dollar 2970 million for the net cost of suffering. CONCLUSIONS: The direct and indirect costs of sleep disorders are high. The total financial costs (independent of the cost of suffering) of dollar 4524 million represents 0.8% of Australian gross domestic product. The cost of suffering of dollar 2970 million is 1.4% of the total burden of disease in Australia.     

Commentary: The battle of Louisville: money, power, politics, and publicity at an academic medical center

In 2009, the entire clinical faculty of the Department of Neurosurgery of the University of Louisville School of Medicine elected to become employees of a nearby community hospital. This took place in the context of the financial burden of caring for the indigent, declining reimbursement, clinical demands for neurosurgical coverage of a level 1 trauma center, rising salaries for neurosurgeons, and competitive pressure on hospitals. The author, who was dean of the school of medicine at the time, would not accept the abrupt withdrawal of these clinicians from the faculty practice plan, single-point contracting, and academic governance of clinical work assignments. Politicians, the press, and accreditation bodies quickly weighed in as the university, the school, and the public good were placed in jeopardy. The motivations for this event-the community hospital defending its market share and physician recruitment and retention pipeline, the dean defending principles of academic governance and the faculty practice plan-and the responses of the participants offer an instructive case study for academic medical management. The author concludes that one might view the protagonists of this episode not as defenders of principles but, rather, as pawns in a larger drama playing out related to a perfect storm of economic and social pressures in American medicine.     

A comprehensive space management model for facilitating programmatic research.

In FY04, the authors developed and implemented models to manage existing and incremental research space, and to facilitate programmatic research, at the University of Arizona College of Medicine. Benchmarks were set for recovery of total sponsored research dollars and for facilities and administrative (F&A) dollars/net square foot (nsf) of space, based on college-wide metrics. Benchmarks were applied to units (departments, centers), rather than to individual faculty. Performance relative to the benchmark was assessed using three-year moving averages, and applied to existing blocks of space. Space was recaptured or allocated, in all cases to programmatic themes, using uniform policies. F&A revenues were returned on the basis of performance relative to a benchmark. During the first two years after implementation of the model (FY05 and FY06), and for the 24 units occupying research space, median total sponsored research revenue/nsf increased from $393.96 to $474.46 (20.4%), and median F&A revenue/nsf increased from $57.42 to $91.86 (60.0%). These large increases in median values are driven primarily from redistribution and recapturing of space. Recruiting policies for unit heads were developed to facilitate joint hires among units. In combination, these policies created a comprehensive space management model for facilitating programmatic research. Although challenges remain in implementing the programmatic recruitment strategy, and selected modifications to the original policy were introduced later (e.g., research space for newly recruited junior faculty is now exempted from calculations for three years), overall, the models have created a climate of transparency that is now accepted and that allows efficient and equitable management of research space.     

An instrument for determining the amount of NIH support for clinical investigations at one academic health center.

PURPOSE: Although interest in supporting clinical investigators is increasing, information regarding the quantity, spectrum, and specific types of clinical research performed in academic health centers (AHCs) is generally not available. The authors report on an instrument to quantify the National Institutes of Health (NIH)-funded component of clinical research at one institution. METHOD: A systematic review of all NIH grants awarded to Massachusetts General Hospital (MGH) in fiscal year (FY) 1997-98 was performed using public information from two NIH Internet sources. Research abstracts from all 487 grants were reviewed and the percentage and type of clinical research activity within each was estimated and compared with estimates provided by a subset of principal investigators. RESULTS: During FY 1997-98, the MGH received $134 million in total NIH funding; $39.9 million (30%) supported the broadest definition of clinical research (that using human materials). When the definition of clinical research was narrowed to direct interaction between investigator and patient for investigative purposes (patient-oriented research), the total for clinical research was $18.2 million. These numbers significantly exceeded the institution's previous estimates of $.6 million for NIH-sponsored clinical trials and $2.2 million for population-based studies. CONCLUSIONS: Clinical investigation is an important component of AHCs' research portfolios from several perspectives, not the least of which is financial. Data on the clinical component of an institution's research effort should be collected prospectively and nationally to inform the optimal allocation of research resources and the alignment of the AHC's infrastructure.     

Using a productivity-based physician compensation program at an academic health center: a case study.

PURPOSE: With increased budget constraints, academic health centers (AHCs) have turned their focus on physician compensation. While many AHCs are concerned that compensation programs driven primarily by revenue generation will have a negative impact on their academic mission, little information is available to support this. The authors examined the effects on teaching and clinical productivity of an innovative compensation program for pediatrics primary care faculty at an AHC and related those effects to national standards for productivity. METHOD: A baseline productivity and compensation assessment was conducted for a group of 35 academic general pediatricians. The data were compared with Medical Group Management Association (MGMA) figures for general pediatricians. A productivity-based faculty compensation program using the work component of the relative-value unit (RVU) as the measure of productivity was designed and implemented. Productivity and compensation were measured after the first year of the program and compared with the baseline assessment. The numbers of hours precepting students and residents and the students' evaluations of their clinical experiences before and after implementation of the program were compared. RESULTS: The baseline assessment showed that over half of the faculty had productivity that fell below the MGMA 25th percentile, while the majority had compensation that exceeded this percentile. After implementation of the compensation program, 89% of the faculty increased their clinical productivity. The times faculty spent precepting and students' evaluations before and after program implementation were unchanged. CONCLUSIONS: Successful productivity-based physician compensation programs can be developed for AHCs.     

Primary care teaching physicians' losses of productivity and revenue at three ambulatory-care centers

This study reports two years of basic data concerning University of Illinois clerkship students, their teaching faculty, and their patients at three community health centers. Students from four classes (1985, 1986, 1987, and 1988) were studied in 1985 and 1986. The faculty were family physicians, internists, and pediatricians who provided 20% of the undergraduate medical education for the last 30 months of a four-year curriculum. The study's goal was to develop estimates of the primary care teaching physicians' productivity, to compare them with the productivity of physicians not involved in teaching, and to provide estimates of revenue shortfalls that occurred for the physicians who were teaching. The estimated productivity of the teaching physicians, working 29 hours a week in ambulatory-care settings, was lower by 30-40% when they were teaching medical students than the productivity of nonteaching physicians regionally and nationally. The average patient-care revenue loss for a full-time-equivalent faculty member per full-time-equivalent student for 1985 was estimated to be $27,531 (regional comparison) or $21,143 (national comparison). The corresponding figures for 1986 were $24,294 and $21,525, respectively. The study's results should be useful to those who are planning to establish ambulatory-care delivery systems and also to directors of existing ambulatory-care delivery systems who may be contemplating accepting medical students.     

Economic burden of schizophrenia in South Korea

This study estimates the treated prevalence of schizophrenia and the annual costs associated with the illness in Korea in 2005, from a societal perspective. Annual direct healthcare costs associated with schizophrenia were estimated from National Health Insurance and Medical Aid records. Annual direct non-healthcare costs were estimated for incarceration, transport, community mental health centers, and institutions related to schizophrenia. Annual indirect costs were estimated for the following components of productivity loss due to illness: unemployment, reduced productivity, premature mortality, and caregivers' productivity loss using a human capital approach based on market wages. All costs were adjusted to 2005 levels using the healthcare component of the Consumer Price Index. The treated prevalence of schizophrenia in 2005 was 0.4% of the Korean population. The overall cost of schizophrenia was estimated to be $ 3,174.8 million (3,251.0 billion Won), which included a direct healthcare cost of $ 418.7 million (428.6 billion Won). Total direct non-healthcare costs were estimated to be $ 121 million (123.9 billion Won), and total indirect costs were estimated at $ 2,635.1 million (2,698.3 billion Won). Unemployment was identified as the largest component of overall cost. These findings demonstrate that schizophrenia is not rare, and that represents a substantial economic burden.     

Sponsored-research funding by newly recruited assistant professors: can it be modeled as a sequential series of uncertain events?

PURPOSE: Recruitment of junior faculty with an investigative focus is essential to regenerate and expand the research mission of academic health centers. Predicting funding profiles for junior faculty is limited by variability in the timing, magnitude, and duration of projected research grant funding. The author demonstrated the validity of Monte Carlo simulation to predict sponsored-research revenues by newly recruited faculty. METHOD: Demographic characteristics and funding profiles were determined for assistant professors recruited to Yale University School of Medicine in four separate fiscal years (1992-93, 1993-94, 1996-97, 1997-98). These data were applied to develop and assess the simulation model. RESULTS: Only when assistant professors were subcategorized by type of research was it possible to accurately predict recovery of both direct research costs and facilities and administrative costs. Simulations illustrated both the high degree of variability among individual faculty and also the advantage of a prediction tool that displays the range and probability of all possible outcomes. CONCLUSION: Sponsored-research funding by newly recruited assistant professors can be modeled as a sequential series of uncertain events and used to predict consequences of imminent changes in federal funding for biomedical research.     

Measuring the direct costs of graduate medical education training in Minnesota.

PURPOSE: To demonstrate the usefulness of self-reported cost-accounting data from the sponsors of training programs for estimating the direct costs of graduate medical education (GME). The study also assesses the relative contributions of resident, faculty, and administrative costs to primary care, surgery, and the combined programs of radiology, emergency medicine, anesthesiology, and pathology (REAP). METHOD: The data were the FY97 direct costs of clinical education reported to Minnesota's Department of Health by eight sponsors of 117 accredited medical education programs, representing 394 sites of training (both hospital- and community-based) and 2,084 full-time-equivalent trainees (both residents and fellows). Average costs of clinical training were calculated as residency, faculty, and administrative costs. Preliminary analysis showed average costs by type of training programs, comparing the cost components for surgery, primary care, and REAP. RESULTS: The average direct cost of clinical training in FY97 was $130,843. Faculty costs were 52%, resident costs were 26%, and administrative costs were 20% of the total. Primary care programs' average costs were lower than were those of either surgery or REAP programs, but proportionally they included more administrative costs. CONCLUSIONS: As policymakers assess government subsidies for GME, more detailed cost information will be required. Self-reported data are more cost-effective and efficient than are the more detailed and costly time-and-motion studies. This data-collection study also revealed that faculty costs, driven by faculty hours and base salaries, represent a higher proportion of direct costs of GME than studies have shown in the past.