3 ms·

The statistical methods are explained in detail and the article was published in what appears to be a reputable peer-reviewed journal. What kind of validation a

• Cougher 7y ago
  I agree that reviewing the statistical methods before judging the soundness of the conclusions is a good idea, I disagree that being published in a reputable peer-reviewed journal is sufficient validation for many well-known reasons and others that may be less-known.
• jariel 7y ago
  It seems to me there's a huge range of factors they didn't and can't possibly control for. Particularly availability of black market opioids. And if I'm not misreading - the changing policies of opioid scrips over time, a lot of which is 'trendy' i.e. socially driven as much as policy driven (i.e. individual doctors making changes in their practice because they read something). More specifically, I'm always wary of 'top down' science that doesn't consider the myriad of real world considerations regarding their thesis; specifically they don't actually talk to people (or even take surveys) to get a feel for what is actually happening. What if they had a lot of conversations and did a few surveys and found that people tended to not medicate 'en masse' for specific interventions? I honestly don't understand how researchers can look at such an intensively social and behavioural subject without getting a personal feel for the prevalent behaviours. I'll bet doctors, social workers on the front line have pretty good insight into what's going on. So my question was pointed to anyone who understood this paper or issue in detail who could articulate in so many words 'why the correlation is valid'.