NEW JOB KLAXON

As of 11am today I am the Bennett Professor of Evidence Based Medicine at the University of Oxford and Director of the Bennett Institute for Applied Data Science

phc.ox.ac.uk/news/oxford-la…

We are Bennett.ox.ac.uk

We will do many, very great things together Heartfelt thanks to Peter Bennett and the Peter Bennett Foundation for recognising and backing our work at @ebmdatalab, to the Oxford teams in dept, division and university, and of course my mighty glorious team. This will allow us to grow, and share, more than ever.

BIG NEWS
theDataLab.org is growing: we will become the Bennett Institute for Applied Data Science Bennett.ox.ac.uk following a generous donation from the Peter Bennett Foundation. Proud of DL team, very grateful to Peter Bennett & team. On!
ox.ac.uk/news/2021-09-1… This is absolutely huge. It brings stability to our core team. It brings us the freedom to innovate quickly. And it is a huge boost to our core ethos: that work with data should drive practical positive impact on lives; and that it should be open code shared to all.

A new independent Citizens’ Jury, commissioned for NHSx and the National Data Guardian, reports that OpenSAFELY is by far the most strongly and consistently supported of all NHS COVID data projects examined.

Why? MINI-THREAD!

thedatalab.org/blog/190/opens…

arc-gm.nihr.ac.uk/projects/Citiz… People often look at "comms" to persuade the public that their use of NHS data is good. Good comms is vital -and we must be open, and celebrate the benefits - but comms is not enough on its own. With @OpenSAFELY we wanted to do more than just SAY that our work was useful and safe

NEW OpenSAFELY PAPER

Prevalence of "Long Covid" in 58 MILLION patients' primary care records.

You read that correctly, this study was conducted in 58 MILLION patients full GP records, analysed in situ.

Now, a puzzle…

doi.org/10.1101/2021.0… The rate of Long Covid, looking at diagnostic codes in primary care, is really low: 40 cases per 100,000 people. This seems much lower than other estimates. So what is happening? Patients not presenting? Missed diagnoses? A high diagnostic threshold? Low awareness of the codes?

The NHS in England has rapidly delivered mass vaccination. Here we share an OpenSAFELY analysis that gives insights on exactly which patients received the COVID-19 vaccine in the earliest parts of the vaccination campaign. medrxiv.org/content/10.110… The data comes from OpenSAFELY-TPP, a secure analytics platform created by our team in huge collaboration now running across the full, linked, pseudonymised electronic health records of over 23 million people in England for urgent COVID-19 research (e.g. nature.com/articles/s4158… )

BING BONG NEW PAPER

NHS hospital medicines data is being collected, but then needlessly withheld from analysts.

This means patients are exposed to sub-optimal treatment. And NHS money is being wasted.

The barriers are NOT technical. They are "people".

bmj.com/content/370/bm… THREAD

Hospitals are where the most expensive drugs are used, and where new drugs are used.

Hospitals are therefore the place where you will find the biggest variation in quality, safety and cost effectiveness of prescribing.

So the biggest opportunities to do good with data!

BING BONG NEW PAPER

We pulled together a huge team to look at how we can bring NHS data analysis into the 21st century.

THREAD!

journals.sagepub.com/doi/10.1177/01… Academics love to run regression analyses and publish academic papers as PDFs in obscure academic journals, all driven by the magical belief that this will change practice (lol). But that's only one small niche use-case for NHS and health data...

journals.sagepub.com/doi/10.1177/01…

Here's a post on the value and hazards of scientific preprints during Covid, from Science Media Centre.

sciencemediacentre.org/what-should-pr…

I HAVE THOUGHTS. brief thread... During the peak of the first wave we urgently preprinted the world's largest ever study on factors associated with death from Covid-19, in over 17 million patients. This is still the world's largest ever study. 2 months later, it's soon to be published.

https://twitter.com/bengoldacre/status/1258372975004389379?lang=en

A serious, constructive thought about regulating tech. No hair triggers, just discussion.

MHRA today sent out a reminder that all Covid apps should have a CE mark.

gov.uk/government/pub…

Regulation is good.

However CE marks for apps seem to do very little. For example: Here is a systematic review in the BMJ describing the woeful evidence for AI skin cancer apps with CE marks.
(It had a small error, corrected).

bmj.com/content/368/bm…

Here is our BMJ editorial accompanying that paper.

bmj.com/content/368/bm…

I've pulled out the key quotes.

This Surgisphere silliness may never have happened if all journals always expected all researchers to share their analytic code. It's easy to scribble out a 3,000 word article and a few tables. It's a huge amount of work to produce all code for all data management and analysis.

https://twitter.com/bengoldacre/status/1197482152151175168

That tweet won't get the RTs it deserves but if I still had time to write columns I'd be writing on this. It's common for medical researchers to not share their analysis code for their papers. Let me tell you what I think when I see this. Personal, general views only.

Busy day. I will be back online at 15:30 UK time to talk more about how we built an unprecedented, open source, highly secure analytics platform running across 24 million patients' full pseudonymised health records, in wide collaboration between academia and the EHR industry... For now our paper is here:

medrxiv.org/content/10.110…

Press release here:

opensafely.org/press-releases…

OUR NEW PAPER.

Factors associated with COVID-19-related hospital death in the linked electronic health records of 17 MILLION adult NHS patients.

Largest study of its kind ever, anywhere.

The power of UK / NHS data, realised.

opensafely.org/outputs/2020/0…

medrxiv.org/cgi/content/sh… HERE IS THE PRESS RELEASE FOR OUR PAPER.

This is the largest study to date, analysing NHS health data from 17.4 MILLION UK adults between 1 Feb and 25th April 2020.

It gives the strongest evidence to date on risk factors associated with COVID-19 death.

opensafely.org/press-releases…

The Covid pandemic poses a very interesting evidence-based policymaking challenge. Sorry, very busy, barely on twitter, but here's a mini splurge. There are legitimate questions about the strengths and weaknesses of evidence for the impact of each option. How many lives will really be saved by “Stay At Home”? What is the medium- and long-term economic impact? Nobody can be absolutely sure on either of these numbers.

BING BONG NEW PAPER

@NDevito1 @sebbacon and me in the LANCET today.

We definitively assess compliance with US law FDAAA 2007, which requires clinical trials to report results within 12 months of completion.

The answer:

59% of trials breached the law

thelancet.com/journals/lance… @NDevito1 @sebbacon This was a truly epic piece of work. We parsed the law. We converted it into unambiguous logic. We built software to download all US trial data and apply that logic. We stored snapshots for time trends. We built regression models to help spot patterns. Because of that we can say:

** NEW EPIC PAPER from our DATALAB **

We ship tools to >100,000 users a year.

We constantly find the NHS breaching best practice around open data.

THIS BLOCKS INNOVATION.

We've now summarised all these barriers in one huge paper. 1/n

jmir.org/2020/1/e15603/ This paper is hard work, but rewarding. You HAVE to read it.

The NHS aspires to be a data driven organisation. The proof is delivery. In @EBMDataLab we have up-close experience of working on NHS data. We hit barriers at every step: those barriers show what needs to change. 2/n

* BING BONG *

A new BMJ EDITORIAL from my team.

WHY RESEARCHERS SHOULD SHARE THEIR ANALYTIC CODE

Non-paywall link below.

NOW this is VERY IMPORTANT stuff for safety, efficiency, reproducibility and quality in research, data science, and evidence…

bmj.com/content/367/bm… Our BMJ editorial is pegged to a recently retracted trial in @JAMA_current

The researchers found, to their horror, that they had made a catatrophic error in their code - their data analysis scripts - for processing and analysing the data in their RCT...

(2/n)

Our fabulous new paper is out in BMJ! We have used VAST datasets to explore DIFFUSION OF INNOVATION in the NHS on an UNPRECEDENTED scale, using a data science approach from another field, adapted with vast effort into a fun new library for medical research bmj.com/content/367/bm… Let me tell you briefly why this paper is awesome, in terms of (1) the specific findings; (2) how this new data science method can be used by us and others in future work; then (3) the power of modern OPEN approaches, sharing code and (gasp) libraries...

The true genius of Brexit is to politicise a personality trait that splits humanity into two basic groups. "Do you want to fix your own complex problems, and take responsibility for them? Or do you want to blame other people for them in an unwarranted but very rewarding fashion?" You see this most clearly in the narratives around schools and housing. Immigrants pay more tax per head than British born people. When you can't get a GP appointment, a council flat or a school place, do you think "my government is failing" or do you think "outsiders are bad"?

I'm in a restaurant in Oxford. Nearly empty. Four very loud American evangelical Christians are very loudly expressing some very, very offensive homophobic views. What do I do? It's horrible. Ok. I asked the waiter. Not great English but lovely. He says, in his view best not to raise it. They were very hostile to the waiting staff when they arrived.

So. We are complicit. The homophobe Christians win. I'm hopeless at violence, no skills.

I am very pleased to say that TWO NEW PAPERS are out TODAY from my group.

They represent one of the most challenging, interesting, and exciting projects I have ever been involved in, across my entire 44 years of nerdy life-course.

Ladies and gentlemen: @COMPare_Trials is out.. @COMPare_Trials We sent correction letters on every misreported clinical trial in the world's biggest medical journals, to see how the editors - and the trialists - would react.

What happened next?

trialsjournal.biomedcentral.com/articles/10.11…

trialsjournal.biomedcentral.com/articles/10.11…

...

Out of stock medicines are in the news today.

bbc.co.uk/news/health-46…

On our OpenPrescribing.net we have data for every individual one of 8,000 GP practices and CCGs showing whether you're affected, and what changes you can make.

THREAD...

ebmdatalab.net/price-concessi… The first thing to note is how quickly the costs from out of stock medicines in the NHS are rising. Here's our dashboard for the whole of England:

openprescribing.net/all-england/co…

Note that the costs are rising very substantially, and very rapidly. Now...