Does Telemonitoring of Health Failure Improve Outcomes?

Heart failure affects two million Americans, most of whom have poor quality of life and use significant amounts of health resources.  Finding ways to manage this condition effectively has been a focus of many disease management efforts and vendors.  Cost-savings are sought by preventing hospitalizations and emergency room use.  Increasingly technology has been brought to bear on the problem, primarily in the form of various methods of telemonitoring.  The New England Journal of Medicine reports on a major randomized trial examining one use of telemonitoring for these patients.  (NEJM Article)

The study covered about 1653 patients recently discharged from a hospitalization for heart failure, divided evenly into a usual care and a telemonitoring intervention group.  All patients got a scale and educational material.  The intervention group was trained in using an interactive voice response system to make daily calls to report weight and symptom information.  The data was downloaded and reviewed by a care team.  The primary endpoint was readmission for any reason or death from any cause, both within 180 days after enrollment and the secondary endpoints included analysis regarding hospital use.  No significant difference was found on the primary or secondary endpoints for either group.  An important fact to note is that although the patients were to use the system every day, 14% never used it and over 40% were not adhering to the protocol by the end of the study.

This research raises strong caution flags about assuming that use of telecommunications technology alone can result in significant benefit in terms of patient outcomes.  How the technology is integrated into an overall management scheme is critical.  At the same time, there are holes in this study that should lead to cautious interpretation of its results.  No analysis is given of the results for those patients who never or rarely used the system versus those who did.  It could be that those non-users had much worse outcomes, skewing the whole group.  It is not the technology’s fault if patients don’t use it; that is a program design and execution issue.  The article references doing satisfaction and quality of life surveys, but does not appear to report those results, which are very important outcomes.  Perhaps most importantly, there is no description of “usual care” other than that it was “guideline driven”.  Since there were 33 sites, usual care could have varied greatly and at some of the practices could have involved very significant management processes.   Weight may have been the wrong or too limited a biometric to focus on.  Greater autonomy without physician intervention may have needed to be given to the care team.  It would be a great mistake to read the results of this trial as remote telemonitoring of heart failure patients doesn’t work.  At a minimum it worked as well as usual care and it is unclear that it cost any more.