2023 ISAKOS Biennial Congress ePoster
Using Fitbit To Monitor Our Patients In The Post-Operative Phase
Martinique Vella-Baldacchino, MD, MRCS, PG CERT (HEALTH LEADERSHIP), IOC DIp Spo, London, United Kingdom UNITED KINGDOM
Sulaiman Sajjad, MD, London UNITED KINGDOM
Shahaab Shameem, MD, London UNITED KINGDOM
Muhammad Beg, MD, London UNITED KINGDOM
Alex D Liddle, MBBS BSc DPhil FRCS(Orth), London UNITED KINGDOM
MSK Lab, Imperial College, London, UNITED KINGDOM
FDA Status Not Applicable
Summary
Caution must be advised when using wearable technology data due to inaccuracies.
ePosters will be available shortly before Congress
Abstract
Objectively measuring physical activity using wearable technology such as smartwatches or smartphones is increasing worldwide. However, its accuracy and relevance towards orthopaedic surgery remains unclear. A survey conducted by AO has shown that 21 % of surgeons employ wearable systems as part of their clinical work using a combination of smartphones and accelerometry in order to analyse patient’s activity such as step count and kinematics.
Consumer wearable physical activity trackers are low cost devices when compared to £20,000 laboratory standard gait analysis treadmill. Despite the advantages, the validity of many consumer devices have not been evaluated and thus it is no surprise that 18% of the AO survey respondents are unsure what data to use.
The aim of this study was to determine if a wearable smartwatch such as the Fibit Luxe is equivalent to the gait analysis treadmill (H/p/cosmos quasar). A second aim was to determine if there is a difference when wearing a wearable sensor on the dominant compared to the non-dominant hand.
Method
21 healthy adults were recruited to this study, their basic demographic profile was recorded. Participants were positioned on the gait analysis treadmill and provided with two Fibit Luxe devices. The FitBit devices were worn on the subject’s dominant and non-dominant hand. Data collection began for a total of 6 minutes when the treadmill reached a speed of 5km/h. The number of steps from the gait lab treadmill were deduced from the computer-based treadmill software. The number of steps from the Fitbit device which corresponded to the same minute intervals as the treadmill was recorded from Fitabase, a software platform which allows researchers to collect data from wearable technology. A non inferiority test was performed using statistical software statgraphics19.
Results
The mean step count for fitbit was 679 steps (SD 38) during the same period on the gait treadmill was 648 (SD 41). Based from previous studies, 5% was taken into account as an acceptable equivalence margin from the gold standard which was determined as 32 steps. A two sided equivalence test was performed and the results did not demonstrate equivalence P= 0.45. A two tailed T test was conducted which showed there was no difference when wearing the sensor on the dominant vs non dominant hand (p=0.225).
Conclusion
With the help of technological giants our ability to collect large amounts of data continues to improve. Despite the excitement of being able to analyse large swathes of data in terms of clinical care, such as in the post-operative phase following trauma or arthroplasty, caution should be observed in terms of accuracy of the data. It is worrying that only 10% of AO survey respondents were worried about validity. In the future, perhaps a more accurate way of objectively measuring recovery would be energy expenditure.