JMIR Publications

ABSTRACT

Background:

Preoperative assessment is crucial to prevent the risk of complications of surgical operations and is usually focused on functional capacity. The increasing availability of wearable devices (smartwatches, trackers, rings, etc.) can provide less intrusive assessment methods, cut costs, and improve accuracy.

Objective:

The aim of this paper is to present and evaluate the possibility of using commercial smartwatch data (Fitbit Inspire 2) to assess functional capacity before elective surgery and correlate such data with the current gold standard measure (Six-Minute Walk Test, or 6MWT, distance).

Methods:

During the hospital visit, patients were evaluated in terms of functional capacity with the 6MWT. Patients have been asked to wear the Fitbit Inspire 2 for 7 days and nights (with a variability of ± 2 days) after the hospital visit before their surgical operation. Resting heart rate (RHR) and daily steps data were retrieved directly from the smartwatch. Feature engineering techniques allowed to extract Heart Rate Over Steps (HROS) and a modified version of Non-Exercise Testing Cardiorespiratory Fitness (NET-F). All measures were correlated with the 6MWT. Heart Rate Variability (HRV) data were not available from the activity tracker.

Results:

The number of patients enrolled in the study is 31. The parameter that correlates best with the 6MWT is the NET-F index (r=0.68, P<.001). The average RHR over the whole acquisition period for each subject had a r of -0.39 (P=.028), even if some patients did not wear the device at night. The correlation of the 6MWT distance with the lower quantiles of HROS showed strong and significant for the 1% quantile, with a Pearson’s coefficient of -0.39 (P=.039) when an outlier is excluded. Fitbit step count had a fair correlation of 0.59 with the 6MWT (P<.001).

Conclusions:

Our study is a promising starting point for the adoption of wearable technology in the subject functional capacity evaluation, which was strongly correlated with the gold standard. The study also identified limitations in the availability of metrics, the variability of devices, accuracy and quality of data, and accessibility, as crucial areas of focus for future work.