,
7 tweets,
3 min read
Read on Twitter
1/ Thread: New study forthcoming on whether #gene-testing can help guide #antidepressant selection for #depressed patients who have failed to respond to 1 or (on average) 3-4 trials of medication. bit.ly/2LQU75I I have been skeptical: mostly, scant added value…
2/ & I found the reporting of this study annoying. On the primary endpoint, improvement in average #depression rating scores, the gene-testing failed. Patients whose MDs chose new meds w/o help of gene tests did just as well—& yet news reports suggested that tests can help…
3/ That's because the gene testing led to a choice of new meds that led to significantly more remissions (patient no longer depressed) and responses (halving of symptom burden). In general, in drug trials, response/remission shows antidepressants to better effect…
4/ Reason is that w attention from MDs, etc., patients on placebo (or, here, less effective meds) do somewhat better, making it hard to outperform on average. What antidepressants do, when they work, is make patients much better—but they also make some patients much worse…
5/ That's the argument I make in my book #OrdinarilyWell. We may think that if they had any sense the researchers would have made remission (rather than change in average symptom scores) their primary outcome. But there is a problem with this reasoning…
6/ which is that if gene testing works, it should protect against v-bad outcomes—so that when you get good remission rates you should also get strong improvement in average scores (since gene tests should stop MDs from prescribing meds that will cause patients to do poorly).
7/ So finally the outcomes here, improved remission rates but ho-hum improvement on average suggests that the gene tests are not providing the help that they are meant to offer.
