I attended a lecture on severe heart failure this morning by the folks at University of Chicago and they mentioned this helpful website from the University of Washington to determine 1-, 2-, and 5-year mortality in patients with severe heart failure. Plug in things like blood test results, functional class, presence of devices, etc. and it estimates prognosis.
It does not include outcomes in patients with left ventricular assist devices or after transplant, but for doctors seeing severe heart failure patients without these destination therapies, it might patients understand the effect of their medications and devices on their survival.
-Wes
The Seattle Heart Failure model is a wonderful effort to show how prognostic aides might someday be helpful in clinical decision making. Somehow being able to integrate these tools into EMR's is the next step as well as seeing how frequent clinical use impacts medical decision making both in deciding to start new therapeutic adventures and when to consider trials of therapy discontinuation.
ReplyDeleteWes, given the years this tool has been around, why do you think cardiologists are not more aware of it? What are the primary barriers to implementing tools like these for cardiologists?
For me as a palliative care physician it is mostly about the appropriate timeline. I need prognostic tools to delineate mortality issues in the less than 1 year range.
Christian
"Wes, given the years this tool has been around, why do you think cardiologists are not more aware of it? What are the primary barriers to implementing tools like these for cardiologists?"
ReplyDeleteChristian-
First: There are a number of these tools out there, and while helpful, it is unclear how they reach their conclusions. It would be helpful if the reason (data) suporting why the mortality curve changes was denoted on the website, but as it is, we really aren't sure where the data for adjustments come from. As such, this leads to skepticism of these tools amongst cardiologists.
Secondly, there's the issue of "marketing" these sites. Not many folks get exposed to them.
Thirdly, there's the problem with keeping them up to date. As new data occurs, who will maintain the algorithms? How much weight should each study garner in the overall estimate? (I was interested to see that biV pacing at baseline did not affect mortality when CARE-HF demonstrated mortality improvement, for instance. Was this study included when the algorithm was made?)
Finally, there's the issue of seeming to play "God" and scaring a patient. Many are reluctant to want to have these discussions.
So there are limitations with these gizmos, but still, they can serve as a starting point when discussing heart failure prognosis with patients. In that respect, I think they're probably helpful.