Predicting Prognosis

In yesterday's post about prognostic indexes reviewed in JAMA, I mentioned in passing that researchers had put all the 16 existing indicators into one online resource, eprognosis.org. While it's intended for doctors, it's freely available to anyone.

Paula Span, editor of the New York Times' terrific New Old Age blog didn't hold back. She pointed her readers directly to eprognosis.org. In the comments section, some of the readers said they were happy to have this tool, some found it distasteful, and several others said they couldn't figure out how to use it.

I, too, initially had found it confusing. But since there seems to be a lot of interest, here's a quick explainer.

First, a screen grab of the site's home page:

Each of those 16 blue bubbles is a separate index. Now, look in the upper right corner, "settings filter." Select the setting that matches your situation. (You will have to go to the website, I unfortunately can't capture interactivity here).

Are you hospitalized ("Inpatient")? Or living at home? ("Community"). Select only one. Then some of the blue bubbles will disappear. Only the indexes which were designed for people in the hospital or people living at home, depending on what you chose, will remain.

Next select just one of the blue bubbles. A series of questions will appear. Every index first asks if you are a healthcare professional. You must select "yes" to fill out the index. The site does not require any validation.

For the purpose of this exercise, I selected age 80-84, male, and then answered "no" for every question.

From here you will get an estimated risk of mortality over the next three or five or ten years, depending on the index you have chosen.

There's a big caveat here. Remember that the researchers said all of these 16 indexes were imperfect, in one way or another. They stress that this site should only be used as one tool. In fact, at the top right of the eprognosis.org home page are three "usefulness" bubbles. The researchers who created this site are asking doctors to let them know how useful the index is at the time they use it. With more information about usefulness, the indexes can, hopefully, be improved. For now, as people use it, the bubbles will get larger or smaller, depending on whether doctors found it useful or not.

Again, the goal is improved decision-making in elderly patients. These indexes are not crystal balls. That's where the "usefulness" comes in. Did the index help guide decision-making in an effective way?

Back at the New Old Age blog, some of the readers who were able to use the tool thought the information could have helped them make better medical decisions, like this person:

I do think that this is highly useful -- for instance, when my mother in law had a catastrophic stroke at the age of 82 it might have been nice to know that all the hyper rigorous efforts that were then being made to control her diabetes were likely too little too late, and it would have been okay to let her enjoy a more varied diet without feeling like we were killing her.