Cancer screening by primary care physicians: A comparison of rates obtained from physician self-report, patient survey, and chart audit

Objectives. This study measured the cancer screening rates of family physicians and compared the measures obtained through physician self-reports, chart audits, and patient surveys. Methods. A cancer screening survey was sent to 50% of the members of the Washington Academy of Family Physicians, with 326 family physicians (74% response rate) completing the survey. Sixty physicians were recruited for the patient survey and chart audit phase, with a 90% participation rate. Patient surveys were conducted with about 350 patients per physician, and chart audits were conducted on a subset of about 50 patients per physician. Each physician's rate of providing each service was computed from the self-report, the patient survey, and the chart audit. Results. Physicians provided many of these services at rates different from those commonly recommended. Large discrepancies were found between the rates measured by different methods. Correlations between rates derived from chart audits and patient surveys were high; however, correlations between rates from physician self-report and either patient survey or chart audit were much lower for all services. Conclusions. Studies of physicians' provision of cancer prevention services should not rely on physician self-report, but should obtain the rates through patient surveys or chart audits.