This would make doctors' lives a living hell and I don't think it would be very good for patients either.
No matter how bad a drug is, there are always going to be people who get better by coincidence when they're on it. No matter how good a drug is, there are always going to be people who weren't even diagnosed with the right disease so of course the drug isn't going to cure them. No matter how safe a drug is, there's always going to be someone who coincidentally got a rash or a heart attack or something at the same time ey started the drug.
All of these people post on Internet forums, and people who have the most extreme and atypical experiences post on Internet forums the most. This is already a big problem for doctors - they prescribe Drug X, the patient goes home, looks it up on the Internet, and finds a page by some crank who had a heart attack when on Drug X and is now leading a campaign against it, and the doctor has to try to communicate that there have been big studies which show the drug has no greater heart attack rate than in the general population to people who may not be fully aware of the difference between scientific and anecdotal evidence.
Evaluating a drug on anything other than the results of randomized clinical trials, if such trials are available, is a bad idea and gets people killed. When no randomized clinical trials are available, I suppose anecdotal evidence can be useful, but only if you're an exceptionally good rationalist who can filter out the chaff.
Also, I think there's a big difference between anecdotes from people you know and anecdotes from strangers. People you know are a random sample: if you saw Uncle Jim start trying Wellbutrin, and you observed the effects, he's essentially a scientific study with a sample of one. Getting anecdotes from people on the Internet self-selects for people who want to talk about their experience.
Maybe the people who get better on a certain drug become evangelists for it all over the Internet, while people who don't get better on it just consider it one more failed drug out of dozens and never talk about it. Maybe people who get a side effect are enraged at the drug company and talk about it every chance they get while people who do fine don't talk as much. Maybe people who are depressed don't have the energy to post on health forums, so if you ask whether an antidepressant works you're only going to get the people who aren't depressed anymore, who are all going to say that it did. Maybe 100% of people on the forum who got an emergency heart attack treatment will say it works, because the rest of them are all dead.
The human tendency to trust a guy named pRoZaCuSeR331 they met on the Internet over the results of meticulously-crafted meta-analyses is an amazing thing, and I'd be wary of doing anything that might further encourage it.
The only cases in which I think this is useful are cases like erowid and nootropics, where something is so uncommon or illegal that no one's bothered to do any formal investigations on it.
People you know are a random sample: if you saw Uncle Jim start trying Wellbutrin, and you observed the effects, he's essentially a scientific study with a sample of one.
If he's your blood-related uncle, he's not just random, he's better than random (when predicting the effects on you and not some arbitrary person).
That's a very good point. People on the internet are not representative. (This is why I doubt that this can replace conventional medical science.)
But I don't know how you can choose to go on a medication without getting some sense of what it does, from someone who isn't your doctor. It seems like jumping into the abyss. Especially with new medications or medications for things that aren't well understood (like the brain.)
There are a number of fora, but I think most routine drugs fall in a gap at this time, where people can get their doctor's advice and don't have serious enough problems to doubt it and go to fora. Birth control does seem like a good candidate for study, since it is so popular: there ought to be enough people who want more data.
The Quantified Self people are mainly about self-monitoring and self-experimentation, but they have dreams of aggregating the data. Moreover, it's run by journalists, so they like to talk about other groups, like Cure Together and Patients Like Me, which are doing "Big & Sloppy" studies. These two sites are structured around disease and while some of the complaints are side-effects of birth control, I'm not sure how one would get advice about other methods. The fora khafra suggests are more structured around drugs, but largely where it is difficult to get medical advice (whether because of legality or unpopularity).
Interesting suggestion, I look forward to knowing more about what you learn. In general, there's a serious problem of there being a difficulty of finding (relatively) unbiased information about a given politically charged topic. I'm interested in any discussion that might lead to better availability of information on such topics.
Maybe this exists already; maybe it would be drastically illegal; but it would be awesome.
I want a forum where people compare their experiences with pharmaceuticals. You add an entry that includes: what condition you have, what drug you took, what the side effects were, how well the drug worked. Some of these are pull-down menus and ratings on a built-in scale, but there's also room for your own commentary. The site runs stats on the drugs based on user data, and also provides a place to chat and vent about the success or failure of the medications.
I came to this idea from two perspectives. One, I'm a twenty-something woman. Everyone I know has a different experience with birth control; there are a variety of pills, and girls talk about a corresponding variety of side effects. But that's all anecdotal. What your doctor will tell you is based on the side effects in the clinical trials (and is calculated to reassure you.) My doctor told me, for instance, that if you don't get any side effects in the first three months, you never will -- but this happens to run counter to direct experience. I wish I could aggregate all the personal stories about birth control and other drugs. (Medications for mental illness also seem to be common among people my age, and there's a corresponding variety of stories about side effects and success rates, some very positive and some very negative. Wouldn't it be really useful to compare experiences with lots of people before making a decision about that?)
The other perspective I come from here is as someone with a passing interest in statistics and science methodology. Pharmaceuticals are tested in scientifically controlled but small studies. It would be useful, as a sanity check, to see if really large quantities of unscientific internet data come up with roughly the same results. We now have the opportunity to do what I think of as "Big & Sloppy" science -- it may be sloppy, but its very enormity might make it useful. Sergey Brin's search for a Parkinson's cure is based on self-reported internet data. It's not the way medical researchers conduct tests, but it takes advantage of the vast amount of anecdotal information that so far medicine doesn't have a great way to harness. People were experiencing health benefits from aspirin for decades before scientists observed a link to heart disease. I'm not as sure as Brin is that "Big & Sloppy" medical science can replace the traditional sort, but it certainly ought to generate insights and hypotheses.
I also suspect that the "Big & Sloppy" approach makes for good advice on choosing pharmaceuticals. In practice, a lot of us do make medical decisions at least partly based on anecdote (Uncle Jim tried Wellbutrin and it made him fat.) Anecdote, at least, has no agenda, compared to professional advice -- "Uncle Jim" really did get fat. But logically, if you would consider anecdotes from people you know, you should be more willing to consider aggregated anecdotal information from tens of thousands of people.
So: does this already exist? Is it illegal? And, out of curiosity, what would one need to do to build it? (to create a forum section, user identities, surveys, and statistics.)