JMIR Publications

ABSTRACT

Background:

The increasing use of activity trackers in mobile health studies to passively collect physical data has shown promise in addressing patient adherence issues.

Objective:

The goal of this study was to develop machine learning models to classify patient-reported outcome (PRO) scores using Fitbit data from a cohort of patients with rheumatoid arthritis (RA).

Methods:

Two different models were built to classify PRO scores; the random forest (RF) Classifier model treated each week of observations independently when making weekly predictions of PRO scores, while the hidden Markov model (HMM) additionally took correlations between successive weeks into account. Analyses compared model evaluation metrics for: 1) a binary task of distinguishing a normal PRO score from a severe PRO score, and 2) a multiclass task of classifying a PRO score state for a given week.

Results:

For both the binary and multiclass tasks, the HMM significantly (p < 0.05) outperformed the RF for a majority of PRO scores, and the highest AUC, Pearson’s Correlation coefficient, and Cohen’s Kappa coefficient were 0.751, 0.458, and 0.450 respectively.

Conclusions:

While further research on this topic is needed to validate our results and findings for use in a real-world setting, this study demonstrates the ability of physical activity tracker data to classify health status over time in patients with RA and enables the possibility of scheduling preventive clinical interventions as needed. If patient outcomes can be monitored in real time, there is potential to improve clinical care for patients with other chronic conditions as well. Clinical Trial: NA