Book club
- Thread starter trinker
- Start date
Possible article to review, listed below. It was made available ahead-of-print this week and has not been formatted yet. The article is on Mendelian Randomization (MR) and how close it can adequately serve as a proxy to a randomized clinical trial, which for some instead of clinical trials you all may use A/B testing or natural experiments. I haven't read the article yet, but will probably later today.
It may be beneficial to all based on its review of logic in trying to establish causal/strong associations between exposures and outcomes. I also don't think the article is too technical or jargon heavy.
So my general description of a MR study, it is a study design approach where nature randomizes people to treatment/exposure group and uses that as an instrument. So a basic example is lactose intolerance, lets say it is random, so if you wanted to study the effects of lactose consumption on obesity or heart disease, people in the lactose intolerant group will not consume it, so they are unexposed - but may land in any setting as those having comparable background characteristic (covariate balance). Will see about putting content on Slack, which I have only gotten onto once.
Epidemiology. 2017 Jun 5. doi: 10.1097/EDE.0000000000000699. [Epub ahead of print]
Nature as a trialist? Deconstructing the analogy between Mendelian Randomization and randomized trials.
Swanson SA1, Tiemeier H, Ikram MA, Hernán MA.
https://www.ncbi.nlm.nih.gov/pubmed...Mendelian+Randomization+and+randomized+trials.
It may be beneficial to all based on its review of logic in trying to establish causal/strong associations between exposures and outcomes. I also don't think the article is too technical or jargon heavy.
So my general description of a MR study, it is a study design approach where nature randomizes people to treatment/exposure group and uses that as an instrument. So a basic example is lactose intolerance, lets say it is random, so if you wanted to study the effects of lactose consumption on obesity or heart disease, people in the lactose intolerant group will not consume it, so they are unexposed - but may land in any setting as those having comparable background characteristic (covariate balance). Will see about putting content on Slack, which I have only gotten onto once.
Epidemiology. 2017 Jun 5. doi: 10.1097/EDE.0000000000000699. [Epub ahead of print]
Nature as a trialist? Deconstructing the analogy between Mendelian Randomization and randomized trials.
Swanson SA1, Tiemeier H, Ikram MA, Hernán MA.
https://www.ncbi.nlm.nih.gov/pubmed...Mendelian+Randomization+and+randomized+trials.
Nother possible one on the battle between machine learning vs. traditional stats: http://projecteuclid.org/download/pdf_1/euclid.ss/1009213726 (~1700 cites)
Since Slack is threaded, I think an easy way to manage this club is to simply make a channel for the club. Initial post can share the article. Everyone who wants to join the discussion on that article can participate in a thread off of it. Some will get attention, others will get none. We'll see where our interests end up!
