Introduction: Home time, defined as the number of days spent away from home following hospital discharge, is used as a proxy measure of functional recovery in stroke studies. Data on home time may be obtained from clinical or administrative data, or by direct patient interview. There is little information on the accuracy of self-reported home time. Methods: We measured the accuracy of self-reported home time using data from the Michigan Stroke Transitions Trial (MISTT) - a randomized trial of case management to improve outcomes in acute stroke patients who returned home after hospitalization. We calculated 90-day home time using information on readmissions to hospitals or rehab facilities obtained from the eMRs of the original admitting hospital and other external reports. Patients or their proxies who completed a 90-day telephone interview reported the number of nights they spent away from home due to medical reasons since first returning home. We compared the agreement between calculated and self-reported home time using kappa statistics. Results: Of 265 enrolled subjects, 214 (81%) completed a 90-day interview and of these, self-reported home time data was available on 195 (91%). Based on calculated home time, 21.5% (42/195) of patients spent at least one night away from home during the 90-day period. The mean and median calculated home time was 88 and 90 days, respectively. When compared to self-reported home time, the kappa statistic for the binary outcome of 90-day home time versus <90days was 0.64 (95%CI = 0.51-0.77). When home time was classified into 3 categories (Table), the observed agreement was 85% and the weighted kappa 0.59 (95%CI = 0.45-0.72) indicating good agreement beyond chance. Women and patients first discharged to a rehab facility were more likely to report discrepancies of >= 1 week in home time. Conclusions: Self-reported home time data has acceptable reliability for use as an outcome measure in stroke studies.