Profile avatar
danlewer.bsky.social
Consultant in Public Health at Bradford Institute for Health Research. Interested in quantitative research methods and mental health
120 posts 233 followers 177 following
Regular Contributor
Active Commenter
Posts 13 Comments 36

The only function of conference posters is to prop up the conference business model

submitted 14 days ago • 1 comment

In science, the fact no one has done something before is a good reason to question whether that thing is really a good idea.

submitted 14 days ago • 0 comments

So many people have a visceral reaction to pie charts. Is it really just that the segments can be a bit harder to compare than in a bar chart? Why are you so passionately against pie charts? Look inside yourself, what have pie charts done to you?

submitted 21 days ago • 0 comments

✨ ✨ WE ARE HIRING! ✨ ✨ We need two new Senior Research Fellows! Have you: 🟢 Experience in statistical and epidemiological analysis 🔵 Used quasi-experimental methods 🟢 An understanding of epidemiological methods Then we need you! borninbradford.nhs.uk/careers/seni...

submitted 30 days ago • 0 comments

Is it possible to calculate a p-value for a p-value? Like the effect was significant at the level of p < 0.05 (p < 0.05)

submitted 36 days ago • 2 comments

Please expand the scope of your research including data you didn't collect while reducing word count

submitted 37 days ago • 0 comments

I do research into health inequalities. People often say "when are we going to stop describing the problem and do something about it?" In answer to the first part of the question, I'm not going to stop describing the problem.

submitted 41 days ago • 3 comments

www.bmj.com/content/308/... From 30 years ago …

submitted 47 days ago • 1 comment

I'm starting to think that interrupted time series in a single population (however large) are not strong designs, as they're so sensitive to modelling decisions, eg.slope/step changes, how you treat seasonality, etc. For better evidence you need an intervention rolled out across multiple populations

submitted 49 days ago • 1 comment

For anyone excited about research using AI and NHS data, how is AI going to help with: - Research questions? - Confounding? - Collider bias? - Immortal time? - Estimands? - Data protection? - Uncertainty? - Generalisability? - Measurement? - Missing data? - Triangulation? - PPI?

submitted 51 days ago • 2 comments

Today’s plan MAINLINES ARTIFICIAL INTELLIGENCE INTO THE VEINS of this enterprising nation www.gov.uk/government/n...

submitted 51 days ago • 1 comment

Published estimates suggesting a survival advantage from giving polyclonal intravenous immunoglobulin in streptococcal toxic shock syndrome are explained, at least in part, by immortal time bias

submitted 54 days ago • 0 comments

NEW PREPRINT! Enjoyed working with @tom-parks.bsky.social and @petedodd24.bsky.social on immortal time bias In its simplest form, ITB results from misallocation of person time, i.e. pre treatment time in intervention group being counted as time on treatment www.medrxiv.org/content/10.1... [1/n]

submitted 54 days ago • 10 comments