Do physiotherapy staff record treatment time accurately? An observational study

Objective: To assess the reliability of duration of treatment time measured by physiotherapy staff in early-stage stroke patients. Design: Comparison of physiotherapy staff's recording of treatment sessions and video recording. Setting: Rehabilitation stroke unit in a general hospital. Subjects: Thirty-nine stroke patients without trunk control or who were unable to stand with an erect trunk without the support of two therapists recruited to a randomized trial evaluating the Oswestry Standing Frame. Twenty-six physiotherapy staff who were involved in patient treatment. Main measures: Contemporaneous recording by physiotherapy staff of treatment time (in minutes) compared with video recording. Statistical analysis: Intraclass correlation with 95% confidence interval and the Bland and Altman method for assessing agreement by calculating the mean difference (standard deviation; 95% confidence interval), reliability coefficient and 95% limits of agreement for the differences between the measurements. Results: The mean duration (standard deviation, SD)of treatment time recorded by physiotherapy staff was 32 (11) minutes compared with 25 (9) minutes as evidenced in the video recording. The mean difference (SD) was -6 (9) minutes (95% confidence interval (CI) -9 to -3). The reliability coefficient was 18 minutes and the 95% limits of agreement were -24 to 12 minutes. Intraclass correlation coefficient for agreement between the two methods was 0.50 (95% CI 0.12 to 0.73). Conclusions: Physiotherapy staff's recording of duration of treatment time was not reliable and was systematically greater than the video recording.