Inter-rater reliability of activity measure for post-acute care '6-Clicks' inpatient mobility short form in the intensive care unit

Objective Examine the inter-rater reliability of the activity measure for post-acute care (AM-PAC) inpatient mobility short form (IMSF) when completed by physical therapists (PTs), during routine clinical practice, in a variety of patients with critical illness. Methods A prospective observational evaluation at single, large academic hospital in the United States. Patients (n = 76) in surgical, medical and neurological intensive care units (ICUs) were evaluated as part of routine clinical practice using the AM-PAC IMSF administered by eight PTs with at least 6 months of experience using this tool. One of two reference rater PTs observed the physical therapy session, and simultaneously scored the AM-PAC IMSF. The reference rater and clinical PTs were blinded to each other's scores with a minimum of 10 assessments completed by each clinical PT. Bland-Altman plots were constructed and intra-class correlation coefficients (ICC) were computed using a random intercept (physical therapy session) model. Results Eighty one assessments (five patients assessed twice) were scored by both a clinical PT and reference rater PT (total assessments = 162). Bland-Altman plots revealed a mean difference in AM-PAC IMSF scoring of 0.0 (95% limits of agreement: -3.0 to +3.0), with an ICC (95% confidence interval) of 0.957 (0.947-0.964). The ICC (95% confidence interval) for patients in surgical, medical and neurological ICUs was very similar: 0.949 (0.927-0.959), 0.963 (0.946-0.971) and 0.936 (0.886-0.955), respectively. Conclusions The AM-PAC IMSF demonstrates excellent reliability compared with reference rater PTs when performed by PTs during clinical care across surgical, medical and neurological ICUs.