Featured Post

Is the new professionalism and ACP's new ethics really just about following guidelines?

The Charter ( Medical Professionalism in the New Millennium.A Physician's Charter) did not deal with just the important relationship of ...

Monday, February 05, 2018

Might meta analyses and meta data dreging gaslight everyone



Two dueling meta-analyses come to mind. There was a widely quoted meta analysis regarding Vitamin E which concluded it increased the risk of death. Soon after that article appeared in the Annals of Internal Medicine , letters to the editor claims that using a different statistical technique on the same data base, there was no increase in death risk.

Large data bases and multiple comparisons between outcomes and variables seem to always show that some factor becomes a risk factor for something often with relative risk values less than 2 and often less than 1.5.RR.Values in this range are unlikely per se to  unearth a causative factor.This activity is sometimes referred to as "data dredging"  or looking for a positive correlation to rush to print and press releases.

The prudent researcher usually emphasizes that these finding are "hypothesis generating" and not game changing breakthroughs although  the less prudent and the lay press and astro-turf propaganda outlets may suggest the latter. When apparently contradictory data then appears in the medical (and then the popular) press what is one to believe. This seems to be a cousin of "gas lighting" even though no one is plotting to make anyone doubt their ability to reason and analyze medical  data but  tons of data and sounding alarms on the findings of many  of them may have a similar effect of "not knowing what to believe".













No comments: