Comparing patient-reported hospital adverse events with medical record review: do patients know something that hospitals do not?

BACKGROUND: Hospitals routinely survey patients about the quality of care they receive, but little is known about whether patient interviews can detect adverse events that medical record reviews do not.
OBJECTIVE: To compare adverse events reported in postdischarge patient interviews with adverse events detected by medical record review.
DESIGN: Random sample survey.
SETTING: Massachusetts, 2003. PATIENTS: Recently hospitalized adults. MEASUREMENTS: By using parallel methods, physicians reviewed postdischarge interviews and medical records to classify hospital adverse events.
RESULTS: Among 998 study patients, 23% had at least 1 adverse event detected by an interview and 11% had at least 1 adverse event identified by record review. The kappa statistic showed relatively poor agreement between interviews and medical records for occurrence of any type of adverse event (kappa = 0.20 [95% CI, 0.03 to 0.27]) and somewhat better agreement between interviews and medical records for life-threatening or serious events (kappa = 0.33 [CI, 0.20 to 0.45]). Record review identified 11 serious, preventable events (1.1% of patients). Interviews identified an additional 21 serious and preventable events that were not documented in the medical record, including 12 predischarge events and 9 postdischarge events, in which symptoms occurred after the patient left the hospital. LIMITATIONS: Patients had to be healthy enough to be interviewed. Delay in reaching patients (6 to 12 months after discharge) may have resulted in poor recall of events during the hospital stay.
CONCLUSION: Patients report many events that are not documented in the medical record; some are serious and preventable. Hospitals should consider monitoring patient safety by adding questions about adverse events to postdischarge interviews.