Breast and prostate cancer patient's reliability of treatment reporting.

Purpose: To compare patient reported cancer treatments with data obtained by trained abstractors in the Metropolitan Detroit Cancer Surveillance System (MDCSS) to assess the reliability of patient interviews as a source for cancer treatments.

Methods: Patients (n=492 breast patients and n=291 prostate patients) were identified from the MDCSS and interviewed approximately 6 months after initial diagnosis for receipt of cancer treatment. Kappa statistics compared agreement between patient's report of their treatments and the MDCSS.

Results: Breast cancer patients had moderate levels of agreement regarding receipt of chemotherapy and radiation and excellent agreement for surgery. In contrast, prostate cancer patients and registry reports achieved excellent concordance for radiation therapy, very good agreement for surgery, and moderate levels of agreement for hormone therapy. Sensitivity of chemotherapy, surgery, and radiation reporting exceeded 90% for both patient cohorts. Overall, patients reported more treatment than was recorded i n M DCSS.

Conclusion: Patients can be reliabledata sources for medical information pertaining to cancer therapies, although recall may vary by treatment type and time since treatment. Protocols involving patient interviews may wish to consider these encounters as timely, reliable data source options.