Patient Recall of Health Care Events and Time to Diagnose a Suspected Ovarian Cancer

Abstract Introduction Patient recall is often used by clinicians to create a history of care leading to consultation with a gynecologic oncologist. Although patient recall may be an efficient method to explore the context of the patient's concerns, the accuracy of recall and its potential impact on care are unknown. This study sought to explore the consistency of patient recall compared with data found in health care records. Patients and Methods This study enrolled 105 eligible patients who were referred to a gynecologic oncologist for suspected ovarian cancer. Ninety-one of these patients were interviewed regarding symptoms, health care events, and the dates leading to diagnosis. The medical records of these patients from all previous providers were obtained and data were abstracted. The intraclass correlation coefficient (intraclass correlation coefficient, ICC[3,1]) was used to examine correspondence between recall and medical record data. Results There was low correspondence between patient recall and the medical record for time to diagnosis (ICC = 0.12; 95% confidence interval [CI] = −0.09 to 0.33; P = .12) and health care events (0.15; 95% CI, −0.05 to 0.348; P = .008). Conclusions There should be limited confidence in information obtained from patient recall given the inconsistency between recall and actual dates and events.