Professor Jonathan Shepherd CBE is Professor Emeritus of Oral and Maxillofacial Surgery at Cardiff University. Early in his career, he found that most assaults treated in hospital were never reported to the police. He built the Cardiff Model to address this gap by combining anonymised A&E data with police data. This approach has helped cut violence by up to 60% across England and Wales. More recently, Professor Shepherd has focused on building the evidence ecosystem. He helped to create the What Works Centres, the College of Policing, and the Chartered College of Teaching.
Three takeaways from me from this episode were:
* The value of a small team: Professor Shepherd discussed the importance of having fewer people in the room, each of whom could take action and get things done.
* The evidence ecosystem - a phrase he coined. His goal is to put everything professionals need around them: evidence generation, professional standard-setting bodies, and regulators.
* The consistency that one person can bring when public servants are coming and going all around: Professor Shepherd did that at Violence Prevention Board; receptionists do it on the ground.
We also discussed, how a randomised trial of pint glasses changed what the whole pub industry bought, turning A&E data into action to cut violence, how Theresa May added the last piece of the violence reduction puzzle, how the colleges of teaching and policing match up to the royal colleges of medicine, and how Richard Trevithick influenced his approach.
Professor Shepherd’s decades’ involvement in turning evidence into action make this a particularly fascinating interview.
You can listen to the interview on Spotify, or read the transcript, lightly edited for clarity, below.
I’d love to start with pint glasses, because it seems like a great example of the kind of work you do.How did you get from a finding about the role of glass in assaults to a change in what the whole pub industry was buying?
This was a planned programme of advocacy and experiments. We discovered that glasses were a weapon in violence that hadn’t been identified before. The next question was: what kind of glass is used as a weapon most frequently? We did a survey on a busy bank holiday weekend across England and Wales, and found that it was the one-pint lager glass – straight-sided, with a bulge at the top, known as a nonic.
I thought: what’s the breakage resistance of these glasses? We set up an experiment in the dental materials lab and got samples from all the manufacturers of one-pint glasses. We found that one of the glass types from one particular manufacturer was almost impossible to break using our impactometer. It turned out that it had been toughened – tempered – in the manufacturing process.
These things are always a series of steps, both for experimentation and for advocacy and achieving impact. The next step was to carry out a randomised trial of toughened glasses in 57 pubs in the West Midlands and South Wales. All the pint stocks in those pubs were switched either to toughened or non-toughened glass. The result was a 60% reduction in injury risk associated with the tougher glasses.
The next step was to get that published in a journal which is widely covered by the media. We chose the British Medical Journal’s specialist journal, Injury Prevention. That produced a substantial amount of media interest: I remember the yellow stickers from each media outlet plastering the door of my study in the hospital.
Did you have all these steps clear in your mind at the start?
It wasn’t clear to start with. Just asking, from the original finding, the bank holiday research, what’s the next step? That led us to the materials laboratory work, then the randomised trial, and then the advocacy after that.
I believe that led to a substantial fall in injuries caused by glass nationwide?
It did. Home Office analysts estimated that this reduced the number of glass assaults by around 40,000 a year in England and Wales. That was the lower bound of their estimate – it was not absolutely precise, but it was clear that this was associated with a step reduction.
There was a lot of advocacy: I spent a lot of time with the media, the pub trade and the glass-manufacturing business. The important moment was the announcement in 1999 in The Licensee, the trade magazine for the pub trade. The headline was, “Industry Switches to Toughened Glass.”
The starting point for the Cardiff Model of Violence Prevention was your discovery, as part of your PhD, that most violence treated in hospitals wasn’t known to the police – because victims were scared, didn’t know what had happened, or were unwilling to face scrutiny of their own behaviour.
How were you able to share this data with the police while maintaining patients’ trust and confidentiality?
The Cardiff Model does not depend on the sharing of personal information. All the information is anonymised before it leaves the hospital. In 30 years, we’ve not had one complaint, or even a question about confidentiality. The Information Commissioner has reviewed this twice, and it has obtained ethics committee approval.
When the patrons of pubs in Cardiff were asked, “How do you feel about the local A&E monitoring which licensed premises violence happens in, and then taking action to make those premises safer?”, there was widespread support. People want to know where it’s safe to go out in the city centre.
In 2000, you published a list of the 10 most violent pubs in Cardiff. What happened next?
That involved a two-page article in the Western Mail, the main newspaper in Wales. Around the edges of the article were the quotes from the interviews I’ve just mentioned with the pub patrons. There was a very helpful result: all of a sudden, the licensees realised that their premises were being scrutinised. They got together and established a licensees’ forum in Cardiff, to make sure that their voice was heard. That was a very positive move, and it was supported by South Wales Police.
We then invited the chair of the licensees’ forum, a prominent licensee in the centre of Cardiff, to join the violence prevention group – not for the purpose of creating policy, but to make sure that information about prevention was relayed from the group to the licensees in an organised way. That relationship has been very good ever since, and continues to this day.
The board brings together agencies including health, the police and the local authority.
Tell us about the first meeting: who was in the room, what did you ask them to do, and how did they respond?
The first meeting was in July 1996. I convened it because I thought we needed a mechanism for turning this data – A&E and police data combined – into practical prevention. I thought the best way to do that – and so it proved - was to get executives from the police, local government, health and, at that time, the national charity Victim Support to meet. I posed the question, “Do you think we could reduce violence more effectively using A&E and police data combined than we can with what is obviously very incomplete police data?” That became Cardiff’s Violence Prevention Group, which has been meeting ever since. I chaired it for two decades, until 2018. An A&E consultant colleague of mine and the head of the Violence Prevention Unit for Wales have co-chaired it since.
The group to turn the evidence into practical prevention is one of the two new ingredients that the Cardiff Model uses. Secondly, the data from A&E. These are particular data items:
* The specific location – a named school, park, street, pub or licensed premises
* What weapon was used
* How many assailants there were, because we realised that more than one assailant is often a signal of gang violence
* The relationship between the victim and the person who harmed them – we used the same categories as the Crime Survey for England and Wales
That information is collected by the emergency department receptionist and shared with the police.
Who does the collating?
The data is recorded b…
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit
www.thingsdonewell.uk