This study estimates the costs of clinical follow-up for women with early stage breast cancer and evaluates the effects of patient and provider characteristics on follow-up intensity. Claims data were cumulated from 1/1/89 through 4/30/93 for a comprehensive set of follow-up tests (office visits, radiologic, and laboratory) ordered for 222 women diagnosed at a university hospital between 1/1/89 and 12/31/91. Aggregated measures of the volume and costs of follow-up over 6 month intervals were expressed in terms of Medicare's 1993 relative value units (RVUs) and their 1993 cost equivalents. Excluding the first 6 months, women received on average, 11.7 RVUs of follow-up in the first year (equivalent to a cost of $362), and 9.5 RVUs in the second year ($297). In the first year, chest x-rays, clinical chemistry tests, automated hemograms, and bone scans, accounted for 36% of the cost of follow-up, while computerized tomographic scans and magnetic resonance imaging studies accounted for 30%. Multiple regression analysis of the first year data showed that women who received radiation and/or were followed by oncology, were more likely to receive intensive follow-up. Age, race, socioeconomic status, insurance, stage, and treatment did not impact follow-up. Costs of follow-up for breast cancer are substantial, though much lower than suggested by others. Additionally, wide variations in practice are largely unexplained by patient and/or provider characteristics. In light of recent evidence questioning the benefit of intensive surveillance, this study supports the need for an accepted set of follow-up guidelines for breast cancer.