Tag Archives: Systematic overviews

Recommended: A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models

This is one of those articles where you have to restrain yourself. Its message, that good old statistical tools like logistic regression can perform as well as these new fangled machine learning approaches that you haven’t taken the time to learn, is quite tempting. But I’d be cautious here. Maybe logistic regression is still competitive, but maybe the systematic overview got a bunch of biased studies. It’s worthwhile to cite this whenever someone makes an overly strong claim about machine learning models, but don’t use this as an excuse to keep from learning the new stuff yourself. This article is stuck behind a paywall. Sorry! Continue reading

Recommended: Published methodological quality of randomized controlled trials does not reflect the actual quality assessed in protocols

When evaluating a series of research articles, you often have to assess the quality of the individual papers based on the type of blinding, for example. What do you do if the paper does not discuss these items? I have usually advocated a “no news is bad news policy.” If a paper does not mention blinding, assume that no blinding was done. It seems reasonable, but the paper by Mhaskar et al provides empirical evidence that sometimes authors leave out information that would strengthen the credibility of their study. A similar paper is at https://www.ncbi.nlm.nih.gov/pubmed/22424985 Continue reading

Recommended: In search of justification for the unpredictability paradox

This is a commentary on a 2011 Cochrane Review that found substantial differences between studies that were adequately randomized and those that were not adequately randomized. The direction of the difference was not predictable, however, meaning that there was not a consistent bias on average towards overstating the treatment effect or a consistent bias on average towards understating the treatment effect. This leads the authors of the Cochrane review to conclude that “the unpredictability of random allocation is the best protection against the unpredictability of the extent to which non-randomised studies may be biased.” The authors of the commentary provide a critique of this conclusion on several grounds. Continue reading