OBJECTIVES: In Germany, health economic studies are increasingly based on health insurance claims data analysis. Such data offer a wide range of scientific applications, especially when focusing on the assessment of resource utilization patterns and costs. The objective of our study was to estimate the direct health care costs of three frequent types of cancer (colorectal, breast, and prostate cancer) from a third-party payer perspective using longitudinal data from a German statutory health insurance fund and employing a matched pairs design. METHODS: Our analysis is based on administrative data of a German sickness fund covering a 5-year period (2005-2009). A total of 42,085 cancer patients were included. Disease-specific costs were estimated by matching cancer patients to counterparts without the particular condition and subsequently comparing the costs of the two groups. One-to-one matching was performed by application of the propensity score method to balance patient characteristics among the cancer groups and non-cancer controls. The cost categories considered in this study included prescription drug costs, outpatient visit costs, and hospitalization costs. RESULTS: The mean cancer-associated 5-year costs per patient amounted to €5,429 for colorectal cancer, €3,200 for breast cancer, and €5,350 for prostate cancer. The average disease-attributable costs of the first year following diagnosis were €8,750, €4,300, and €4,750 for colorectal, breast and prostate cancer, respectively. Corresponding excess costs of the last year of life were €15,900, €10,950, and €14,750. Costs associated with hospitalization accounted for a major part of the total disease-specific costs (up to 80%). CONCLUSIONS: This cost-of-illness study based on claims data analysis confirms the high economic burden of colorectal, breast, and prostate cancer. Most of the costs occurred in the initial and terminal treatment phases. Inpatient treatment was found to be the main cost driver.