Smartphone App with an Accelerometer Enhances Patients’ Physical Activity Following Elective Orthopedic Surgery: A Pilot Study

Abstract

Low physical activity (PA) levels are common in hospitalized patients. Digital health tools could be valuable in preventing the negative effects of inactivity. We therefore developed Hospital Fit; which is a smartphone application with an accelerometer, designed for hospitalized patients. It enables objective activity monitoring and provides patients with insights into their recovery progress and offers a tailored exercise program. The aim of this study was to investigate the potential of Hospital Fit to enhance PA levels and functional recovery following orthopedic surgery. PA was measured with an accelerometer postoperatively until discharge. The control group received standard physiotherapy, while the intervention group used Hospital Fit in addition to physiotherapy. The time spent active and functional recovery (modified Iowa Level of Assistance Scale) on postoperative day one (POD1) were measured. Ninety-seven patients undergoing total knee or hip arthroplasty were recruited. Hospital Fit use, corrected for age, resulted in patients standing and walking on POD1 for an average increase of 28.43 min (95% confidence interval (CI): 5.55–51.32). The odds of achieving functional recovery on POD1, corrected for the American Society of Anesthesiologists classification, were 3.08 times higher (95% CI: 1.14–8.31) with Hospital Fit use. A smartphone app combined with an accelerometer demonstrates the potential to enhance patients’ PA levels and functional recovery during hospitalization.

Optimization and Validation of a Classification Algorithm for Assessment of Physical Activity in Hospitalized Patients

“…Low amounts of physical activity (PA) and prolonged periods of uninterrupted sedentary activity are common in hospitalized patients. Patients spend between 87% and 100% of their day lying in bed or sitting in a chair [ 1 , 2 , 3 , 4 , 5 ]. Little time is spent being active, and bouts of standing and walking are usually short [ 6 , 7 ].…”

“…This sedentary behavior is found in all patient subpopulations. On average, surgical inpatients spend 10 to 71 min per day standing and walking [ 1 , 8 , 9 , 10 ], compared to 66 to 117 min for geriatric inpatients [ 7 , 11 , 12 , 13 ], 1 to 184 min for medical inpatients [ 14 , 15 , 16 , 17 , 18 , 19 , 20 ], 10 to 86 min for post-stroke inpatients [ 21 , 22 , 23 , 24 , 25 , 26 ], and 0 min for patients admitted to the intensive care unit [ 27 , 28 ].…”

“…Monitoring patients’ PA behavior during hospitalization is commonly performed using self-reported measures, behavior mapping, or wearable activity monitors [ 1 , 2 , 40 , 42 ]. Self-reported measures (e.g., surveys or diaries) are subjective and show low validity and reliability [ 40 , 42 , 43 ].…”

“…Moreover, it may under- or overestimate amounts of PA and periods of uninterrupted sedentary activity when observations are performed during daytime hours only, or when sampled at brief intervals (e.g., one minute in every ten minutes) [ 16 , 41 , 46 , 47 , 48 ]. As bouts of walking often last less than two minutes, they might not be recorded, resulting in an underestimation of the amount of PA. Wearable activity monitors, such as accelerometers, allow for objective, continuous quantification and classification of patients’ PA behavior over longer time periods, with minimal effort and invasiveness [ 1 , 2 , 6 , 41 , 49 , 50 ]. Despite all their advantages, accelerometers have not been widely integrated in clinical practice, due to issues relating to feasibility, reliability, and validity [ 2 , 41 , 50 ].…”

“…The performance metrics of an algorithm to measure patients’ PA behavior are influenced by patient characteristics (e.g., age, walking speed, gait pattern, and the use of a walking aid), sensor wear location, number of sensors used, and outcome parameters (e.g., classifying activities, step count, and intensity) [ 41 , 42 ]. Time spent in dynamic activities (e.g., walking, stair climbing) and the classification of postural transitions from sedentary to upright position are the most relevant outcome parameters for hospitalized patients, as they need to increase their amount of PA and interrupt prolonged periods of sedentary activity [ 1 , 35 , 45 ]. Most accelerometer algorithms are validated in healthy adults and lack the sensitivity to classify slow or impaired gait [ 58 , 59 ].…”