Med Patient Social Networks Are Better Scientific Institutions

When you’re suffer­ing from a life-chang­ing ill­ness, where do you find in­for­ma­tion about its likely pro­gres­sion? How do you de­cide among treat­ment op­tions?

You don’t want to rely on stud­ies in med­i­cal jour­nals be­cause their con­clu­sion-draw­ing method­olo­gies are hap­haz­ard. You’ll be bet­ter off get­ting your prog­no­sis and treat­ment de­ci­sions from a so­cial net­work­ing site: Pa­

Pa­ lets pa­tients with similar ill­nesses com­pare symp­toms, treat­ments and out­comes. As Jamie Hey­wood at TEDMED 2009 ex­plains, this rep­re­sents an enor­mous leap for­ward in the scope and method­ol­ogy of clini­cal tri­als. I highly recom­mend his ex­cel­lent talk, and I will para­phrase part of it be­low.

Here is a re­port in the Pro­ceed­ings of the US Na­tional Academy of Sciences (PNAS) about Lithium, which is a drug used to treat Bipo­lar di­s­or­der that a group in Italy found slowed ALS down in 16 pa­tients. When PNAS pub­lished this, 10% of the pa­tients in our sys­tem started tak­ing Lithium, based on 16 pa­tients’ data in a bad pub­li­ca­tion.

This one pa­tient, Hum­berto, said, “Can you help us an­swer these kinds of treat­ment ques­tions? I don’t want to wait for the next trial; I want to know now!”

So we launched some tools to help pa­tients track their med­i­cal data like blood lev­els, symp­toms, side effects… and share it.

Peo­ple said, “You can’t run a clini­cal trial like this. You don’t have blind­ing, you don’t have data, it doesn’t fol­low the sci­en­tific method—you can’t do it.”

So we said, OK, we can’t do a clini­cal trial? Let’s do some­thing even harder. Let’s use all this data to say whether Lithium is go­ing to work on Hum­berto.

We took all the pa­tients like Hum­berto and brought their data to­gether, bring­ing their his­to­ries into it, lin­ing up their timelines along mean­ingful points, and in­te­grat­ing ev­ery­thing we know about the pa­tient—full in­for­ma­tion about the en­tire course of their dis­ease. And we saw that this or­ange line, that’s what’s go­ing to hap­pen to Hum­berto.

And in fact he took Lithium, and he went down the line. This works al­most all the time—it’s scary.

So we couldn’t run a clini­cal trial, but we could see whether Lithium was go­ing to work for Hum­berto.

Here’s the mean de­cline curve for the most ded­i­cated Lithium pa­tients we had, the ones who stuck with it for at least a year be­cause they be­lieved it was work­ing. And even for this hard core sam­ple, we still have N = 4x the num­ber in the jour­nal study.

When we line up these pa­tients’ timelines, it’s clear that the ones who took Lithium didn’t do any bet­ter. And we had the power to de­tect an effect only 14 the strength of the one re­ported in the jour­nal. And we did this one year be­fore the time when the first clini­cal trial, funded with mil­lions of dol­lars by the NIH, an­nounced nega­tive re­sults last week.