Health Outcome Evaluation: Accounting for Regression to the Mean
Submitted by TGreeneJSI on 13 February 2014 - 3:15pm
Wonderful presentation. I am wondering whether in determing health outcomes whether you were able to account for reductions in ER visits, hospitalizations, etc. that are likely to occur just over time rather than due to the intervention. For example, perhaps someone was enrolled because they had just had a severe asthma exacerbation and in six months time they may not have been expected to have another incident that serious.
jkrieger
13 February 2014 - 3:18pm
Permalink
I just posted the same
I just posted the same question - and also asked how convinving pre/post data are to decision makers. it it is sufficient, this makes it much easier to generate the data. Great work, Multnomah!
TGreeneJSI
13 February 2014 - 3:24pm
Permalink
Oh yes Jim, I see your post.
Oh yes Jim, I see your post. We were actually discussing today in Massachusetts whether we could apply results from your studies to possibly estimate regression to the mean. Because the populations are different we didn't think we could use it in a calculation, but could at least in the discussion provide estimates of the effect that are available through the literature. What do you think? What studies might we draw upon?
Terry