Morbidity after intracranial tumor surgery: sensitivity and specificity of retrospective review of medical records compared with patient-reported outcomes at 30 days

OBJECT: Published outcome reports in neurosurgical literature frequently rely on data from retrospective review of hospital records at discharge, but the sensitivity and specificity of retrospective assessments of surgical morbidity is not known. The aim of this study was to elucidate the sensitivity and specificity of retrospective assessment of morbidity after intracranial tumor surgery by comparing it to patient-reported outcomes at 30 days. METHODS: In 191 patients who underwent surgery for the treatment of intracranial tumors, we evaluated newly acquired neurological deficits within the motor, language, and cognitive domains. Traditional retrospective discharge data were collected by review of hospital records. Patient-reported data were obtained by structured phone interviews at 30 days after surgery. Data on perioperative medical and surgical complications were obtained from both hospital records and patient interviews conducted 30 days postoperatively. RESULTS: Sensitivity values for retrospective review of hospital records as compared with patient-reported outcomes were 0.52 for motor deficits, 0.4 for language deficits, and 0.07 for cognitive deficits. According to medical records, 158 patients were discharged with no new or worsened deficits, but only 117 (74%) of these patients confirmed this at 30 days after surgery. Specificity values were high (0.97-0.99), indicating that new deficits were unlikely to be found by retrospective review of hospital records at discharge when the patients did not report any at 30 days. Major perioperative complications were all identified through retrospective review of hospital records. CONCLUSIONS: Retrospective assessment of medical records at discharge from hospital may greatly underestimate the incidence of new neurological deficits after brain tumor surgery when compared with patient-reported outcomes after 30 days.