Comparison of self-reported initial treatment with medical records: Results from the prostate cancer outcomes study

Medical records are generally accepted as the most accurate source of infor mation documenting cancer treatments. However, as the health care system be comes more decentralized and more cancer care is delivered in outpatient se ttings, it is increasingly difficult and expensive to review records from t he many surgeons and medical/radiation oncologists who administer cancer th erapies in the community setting. Using 1994-1995 data, the authors compare d initial treatment for prostate cancer self-reported (from a mailed questi onnaire or telephone/in-person interview) by 3,196 US men in the population -based Prostate Cancer Outcomes Study with information obtained from medica l records. Agreement between self-reports and medical records varied by typ e of treatment. Generally, agreement was excellent for more invasive proced ures such as prostatectomy or radiation (kappa values > 0.8), with decreasi ng agreement for hormone shots and pills (kappa values < 0.7). If the medic al record abstract is assumed to be the "gold standard," the estimated sens itivity was generally high (>80%) for prostatectomy and radiation but low ( 68%) for hormone pills, although the estimated specificity was 90% or great er for all treatments. These results can serve as a useful guide to researc hers contemplating the use of surveys as an alternative to medical record a bstraction to ascertain treatment in studies of patient outcomes.