Page 19 Heart Failure Remote Monitoring Evidence From the Retrospective Evaluation of a Real World Remote Monitoring Program

on some sort of less intensive monitoring after discharge from the program) may be helpful. Less
intensive monitoring following a remote monitoring program like CCCP is appealing and feasible because

of the increasing availability and reducing costs of consumer-oriented monitoring devices that can be
easily used by patients [22,23]. Future prospective research to evaluate optimal program monitoring
duration, risk stratification to identify patients that may benefit for prolonged monitoring, and the
prospects of less intensive, long-term monitoring is needed.

Limitations

Apart from the retrospective nature of this study, it has a number of other limitations. The individual
matching done in this study did not include any measure of disease severity. To ensure that patients

and their matched controls were comparable at baseline, in terms of severity of disease, we evaluated
and found that the rates of hospitalization were similar in both groups 120 days prior to (the CCCP
patient and corresponding matched control) enrollment in the CCCP program. We also controlled for
measures of disease severity (ejection fraction and NYHA classification) in our analysis. Another

limitation is that any hospital admission occurring outside of the Partners’ electronic medical records
were not captured in these analyses. However, this effect is minimized in the CCCP group because they
were monitored daily and had more regular contact with their health care providers. On the other hand,
we were more likely to have missed hospitalizations in the control group who may have out-of-system

hospitalizations. We also cannot rule out unmeasured confounding from comorbidities and the fact that
control patients might have received other treatments including a remote monitoring program other
than CCCP. Additionally, given that the program was implemented in the setting of an academic medical
center, findings from this evaluation may be generalizable only to such settings.

Conclusions

Results from these analyses suggest that compared with usual care controls, this remote monitoring
program is associated with significantly lower hospitalization rates up to 90 days and significantly lower

pg. 19