Hacker News new | past | comments | ask | show | jobs | submit login

Read the primary sources yourself. If you're a hacker you'll have no difficulty making sense of a research paper, even if you have no academic background.

Anecdotal evidence is particularly unhelpful in the case of depression because most cases resolve of their own accord. Spontaneous recovery is the most common outcome and antidepressants are the most common treatment, so it doesn't tell us a great deal if the two coincide quite often.

The big metastudies are fairly unanimous - the efficacy of antidepressants is marginal at best. Kirsch et al (2008) [1] showed that response to SSRIs was only better than placebo in the most severely depressed patients, due not to better response to the drugs but to lower response to placebo.

For me, the takeaway from the data isn't that drugs are useless, but that anything treats depression if you believe it to. All treatments have roughly the same efficacy - that of placebo. Do whatever makes you feel good and you'll feel better. I'd be reluctant to choose drugs over another placebo because they're an expensive and risky placebo, but if you believe in drugs, take the drugs.

The real thinking point is what this says about the nature of depression as a diagnosis. Personally, I think the disease model doesn't work and that we're inappropriately medicalising something because we have a cultural preference to see unhappiness as medical rather than social.

[1]http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fj...




> Anecdotal evidence is particularly unhelpful in the case of depression because most cases resolve of their own accord. Spontaneous recovery is the most common outcome and antidepressants are the most common treatment, so it doesn't tell us a great deal if the two coincide quite often.

I'm sorry, but huge CITATION NEEDED.

> The big metastudies are fairly unanimous - the efficacy of antidepressants is marginal at best. Kirsch et al (2008) [1] showed that response to SSRIs was only better than placebo in the most severely depressed patients, due not to better response to the drugs but to lower response to placebo.

Please cite the other meta-analyses that support this. I am somehow disinclined to believe you on this topic. Here are a few meta-analyses in patients with and without certain concomitant diseases that I found in the last five minutes published in 2011 that disagree with you:

http://www.ncbi.nlm.nih.gov/pubmed/21545782

http://www.ncbi.nlm.nih.gov/pubmed/21558287

http://www.ncbi.nlm.nih.gov/pubmed/21430793

http://www.ncbi.nlm.nih.gov/pubmed/21425505

http://www.ncbi.nlm.nih.gov/pubmed/21346483

>The real thinking point is what this says about the nature of depression as a diagnosis. Personally, I think the disease model doesn't work and that we're inappropriately medicalising something because we have a cultural preference to see unhappiness as medical rather than social.

Just because you think that the disease model doesn't work doesn't mean that it doesn't. I don't know if you have anyone close to you that suffers from depression, but it is most certainly not a purely social problem. There are demonstrable pathophysiological differences in depressed patients that are not present in non-depressed patients.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: