Improving patient selection for use of consumer grade physical activity monitors in the hospital

Background: Bedrest is toxic for inpatients and consumer grade physical activity monitors offer an economical solution to monitor patient ambulation. But these devices may not be accurate in debilitated hospitalized patients who frequently ambulate very slowly. Objective: To determine whether measures of physical capacity can help identify inpatients for whom wearable physical activity monitors may accurately measure step count. Methods: Prospective observational study of 54 adult inpatients with acute neurological diagnoses. Patients were assessed using 2 physical capacity assessments (Activity Measure for Post-Acute Care Inpatient Mobility Short Form [AM-PAC IMSF] and Katz Activities of Daily Living [ADL] scale). They also completed a 2-minute walk test (2MWT) wearing a consumer grade physical activity monitor. Results: The wearable activity monitor recorded steps (initiated) in 33 (61%) of the inpatients, and for 94% of inpatients with gait speeds >0.43 m/s. Physical capacity assessments correlated well with gait speed, AM-PAC IMSF r = 0.7, and Katz ADL r = 0.6, p < 0.05. When the physical activity monitor initiated, the mean absolute percent error (SD) comparing device calculated steps to observed steps, was 10% (13). AM-PAC IMSF (T-score >45) and Katz ADL (>5) cutoff scores identified inpatients for whom physical activity monitors initiated with a sensitivity of 94 and 91%, respectively. Conclusions: Physical capacity assessments, such as AM-PAC, and Katz ADL, may be a useful and feasible screening strategy to help identify inpatients where wearable physical activity monitors can measure their mobility.