Alex Thompson

NEW PAPER 👏👏 "Quantifying the impact of taking medicines for primary prevention: a time-trade off study to elicit direct treatment disutility in the UK" @kp170768 @gbrlrgrs @athomp_HE Please read full paper here: bmjopen.bmj.com/content/13/9/e…

My take on Trust in Algorithms in Harvard Data Science Review: "we need to be able to assess the trustworthiness of claims made both about an algorithm, and those made by an algorithm" hdsr.mitpress.mit.edu/pub/56lnenzj

NEW ARTICLE Going to do a review of methods? You might be interested in reading our open access article. TL;DR: default search engines (e.g. Pubmed, Scopus) can be poor in retrieving articles on used methods, recovery rates may vary over time jclinepi.com/article/S0895-…

"A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models." ML only performed better in studies with high risk of bias. Detailed review by @GSCollins @ESteyerberg and others jclinepi.com/article/S0895-…

Published! "Many Analysts, One Dataset: Making transparent how variations in analytical choices affect results" 65 of us led by Raphael Silberzahn demonstrate the contingencies of analytic decisions on observed outcomes. doi.org/10.1177/251524… (OA: psyarxiv.com/qkwst/)

Courage is a prerequisite for the practice of all other virtues otherwise one is virtuous only when virtue has no cost. #CSLewis