Attrition of remote patient monitoring use for hypertension management in Medicare fee‐for‐service beneficiaries (2018–2021)

All authors report no conflict of interest. Figure S1. Distribution of individual posterior probabilities by group membership in the two-group trajectory model (final model selected). The figure shows distribution of posterior probabilities for belonging to the attributed trajectory group for all beneficiaries in our sample. We can see that the majority of beneficiaries have high probabilities (close to one) of belonging to the trajectory group they were assigned. In the low attrition group, however, there appears to be a small cluster of beneficiaries with probabilities of around 0.6, which indicates some uncertainty around classification of these beneficiaries. Table S1. Bayesian Information Criteria (BIC) for models with one, two, and three trajectories. The table shows BIC values for the model with one, two, and three trajectory groups, respectively. Higher values indicate better fit (the models were fit using proc traj in SAS). Here, three-group model has the best fit, but based on information in the next two figures, we decided that two-group trajectory model was the right model for the data. Figure S2. Three-group trajectory model of RPM use. The figure shows the three-group trajectory model with the overall probability and 95% confidence interval of RPM use by month for each trajectory group. We can see that the confidence interval for the high attrition trajectory group overlaps with the other two groups, indicating that the three-trajectory group model has high uncertainty. Figure S3. Distribution of individual posterior probabilities by group membership in the three-group trajectory model. The figure shows the distribution of probabilities of being assigned to a trajectory group for all beneficiaries in the analysis. We can see that in the intermediate attrition group, there is a big cluster of beneficiaries with probabilities of around 0.7 for being assigned to that group, suggesting more uncertainty for the three-group trajectory model. Figure S4. Multivariable logistic regression analysis for persistent use category (defined as at least 6 months of RPM use in the 12 months of follow-up). Percentage of patients with persistent use = 68.4%. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.