I was fortunate enough to hear Professor Jonathan Shepherd talk about this recently at the RSM, and most people will have seen his
leading article in the BMJ last week.  (BMJ 2000;321:1481-2) I am quite surprised that there has not yet been a debate about his proposal
amongst A&E physicians.

Professor Shepherd in essence wants A&E departments to give the Police data on assaults sustained on a regular (perhaps monthly or
quarterly) basis. He points out something we all know well; many assaulted persons that present to A&E departments never inform the
Police so Police statistics are inadequate. He suggests that the location of the assault, the weapon if any, the type of incident, the
relationship of the attacker to the attacked should all be forwarded - in a non-confidential, aggregate fashion.

He has formed the Cardiff violence prevention group, and shown that particular licensed premises are hotspots for violence. Police
activity has been concentrated on these, and a consequent reduction in assaults has been claimed.

We need to debate this because the Government is suggesting that its plans to tackle community violence will be based not just on
information about offences but also in our injury data. (Simmons J. Review of crime statistics: a discussion document. London: Home
Office, 2000)

Am I the only person to have misgivings about this approach? The aim of reducing community violence is of course laudable, but is this
approach wise? Shepherd says that there are 6 reasons that the data should be provided:

1. Injury data provides a measure of serious violence which is independent of Police measures- which are often inaccurate or incomplete.

2. Recording injuries treated in A&E has the potential for largely complete coverage of serious community violence.

3. Recording provides local information of importance to the Police that will help them respond.

4. Injury data would provide a new performance indicator of policing at police force level.

5. Injury data can provide outcome information on the injured victim, which is currently lacking in Police reports.

6. Injury data provide a set of measures which are compatible with other data sources - in this case the British crime survey.

I ask you to bear in mind that Shepherd is a Professor of Maxillo-facial surgery. What he does not have is the responsibility for ensuring
that the community continues to have faith in its emergency services and that it perceives the hospital, and particularly A&E, as
independent from the Police. One can see the worst consequences of a partisan view in Ulster, where people often seek backstreet
medical help rather than attend a hospital. If people think that we are going to be providing 'some' data to the Police, they will
automatically assume that we are providing evidence that can be traced to them personally. A fortuitous arrest and release on bail and
they appear to take it out on the A&E staff who they think 'snitched' on them. There is far too much violence directed at A&E staff
already without this additional hazard. What he also does not have is a perspective on the gamut of violence. If the brutes do not take it
out on each other in the street what is going to happen? They aren't suddenly going to become pansy-clutching Walters. The least
unpleasant scenario is that they will simply switch venue, and fight outside a different pub. Worse, and perhaps more insidious, they will
go home and take it out on their partner or family. Again, domestic violence is not going away.

Let us examine each of  Shepherd's justifications in turn.

It can be agreed that injury data is more complete than Police data, but this has been shown repeatedly. Teanby (BMJ 1992;304:422) even
found pedestrian fatalities that the Police were unaware of!  His second point is just the first turned on its head so may be disregarded.
His third proposal is perhaps the one that is most worrying. If the data is monthly or quarterly it will be in arrears and Shepherd does not
make it clear how it would help the Police to respond unless he is proposing stationing officers at or near known troublespots. Is the
intelligence around Cardiff so poor that the Police had no knowledge of the local hotspots before this data appeared? Even if their data
was incomplete, statistically it should still be distributed in the same way - in other words their data would be equally bad about
everywhere, so the same hotspots would appear. Although I am lucky enough to work in an area with comparatively little violence, I still
know exactly which pubs not to stand outside at after 11pm and the Police know about them too. I don't understand his fourth point; it
may represent some sort of value for money measure. His fifth point is splendid if only for its complete irrelevance - why does a Police
report need outcome data? Finally, the data is claimed to be compatible with something called the British crime survey. Shepherd can't
have it both ways; either the Police haven't got enough data and want ours or they have got enough to provide a reasoned evaluation of
the data he wants us to provide.

I think that Jonathan Shepherd's aims are entirely laudable. We all want to see less violence; we all want to see less unnecessary work.
However I am uneasy on two counts. Firstly I am very uneasy that this is not a workable solution to the problem; it is simplistic and
fraught with potential risk, including to ourselves and our staff. Secondly, and this is possibly an even greater worry - is it the thin end
of the wedge? Will we be required to breathalyse all drivers attending after road crashes and report them to the Police if they fail? Will
we be required to inform the Police of all positive urine tests for illicit substances? Will we have to arrest the lads who appear with
wounds clearly acquired breaking into someone's house? This is as good a forum as any for the debate.

/Rowley./