There are a few issues here.
Firstly if I am despatched to a patient as an emergency, rarely will I know the identity of the patient. Mostly we are given one out of male/female and another out of adult/child and a brief description from a member of the public about the injury which may well be incorrect. Many times we have gone to call given as an adult male in cardiac arrest to find the young girl sitting up and chatting but suffering from a sprained ankle, or perhaps vice versa. I exaggerate not. That is why every call is treated as an emergency until a reliable identifiable traceable person can give an assessment, which will usually be the first crew to arrive. We simply cannot rely on information from unqualified persons.
Secondly many of the patients are unknown to the callers (helpful members of the public) and therefore to us. Even if a person is found unwell in an address, no-one can be certain that they live there and are not their sister visiting for a weekend or are a burglar taken unwell "at work" (its happened), or are driving a car registered to someone else so until we start searching them for photo-id, with a recent reasonable photograph (and the hope that the scene is well lit, and there are no disfiguing facial injuries, changes of beard since the photo was taken etc) and then we can tie this back to some patient data (and lets leave identical twins out of it as well) then again what use are the records? I will do my best to identify the patient but if life saving interventions are required and there are no close friends or family nearby then I don't have time to do this before things start happening, by the time I found out anything relevant to emergency treatment from "the database" it might well be too late.
If we had the ability to "pull" the record in the street once we found out who the casualty was, rather than it being "pushed" to us from control when the call is despatched, then we'd also need a damn good data carrier. Large events tend to overload the cellular networks. I am sure people will suggest accolc, airwave and all sort of other things but to be frank when it all goes tits up then some of us still have a few handheld VHFs for resilience and "local control". One major service I know has retained 200 vehicle VHF sets. I've seen them (I wasn't supposed to). They're all in boxes at the back of a warehouse. Fantastic. I'd seriousl consider semaphore at events it the scouts still taught it. The idea of using my warning lights as some kind of Aldis lamp appeals but I realise here that I'm getting off topic.
As a few people have mentioned. it would be nice to assume that everyone with a known pre-existing medical condition wears some form of medic-alert bracelet. This simply does not happen, and there are quite a few situations where it is impractical to do so. You could say that this will be a form of Darwinian selection then, but in the same way as we don't refuse to treat the many many people who go to an all day music festival and "didn't think they'd need their inhalers" (for various respiratory conditions, which are often life threatening) we have to manage as best we can. This does make a reasonable argument for the proposal and I acknowledge the good intent.
Patient records can be wrong. A relative (and I appreciate this is a minor point but its the principle I am trying to illustrate) was incorrectly listed as a heavy smoker. She'd never smoked in her life and of this I am certain. A salive swab test (for life insurance) confirmed this insofar as such tests can. The GP accepted that the record was wrong but could not alter the record as he did not put that particular note on. The GP who had added this at a previous surgery where she had been registered had left that surgery and no-one else could remove that note. I pointed out the bit about Data Protection Act requiring information to be accurate and this cut no ice. An amendment could be put on but this was pages away from the summary which would be all we might have time to read. If the incorrect information had been for something more serious, well the consequences of treatment withheld or incorrectly give could be fatal.
There are many agencies involved in pre-hospital emergency care. Statutory ambulance services. Voluntary Aid Societies (Red Cross, St. John) who do 999 work for some statutory services or also cover events on their own, private services covering events or againg doing 999 work. BASICS Doctors. Local first responder schemes (some allow theirs to do considerably more than basic 1st aid, O2 and Defib). Would all of these have access? If so the papers will scream about the "untrained and unsupervised" people with access to your records. If not the papers will scream about the postcode lottery and the incomplete provision of services. If they do have access, will every first aider covering a small event (who will in effect supplement for the ambulance for some time until it arrives) have some sort of data terminal? You might say that first aiders don't need it but many are fully qualified ambulance personnnel, could they be held liable for a poorer standard of care given due to not having such information?
Who is going to pay for all these portable devices, the data contracts, the training, remote security management etc, and the replacements when the yoof of sarf lundun realise that some of the medics now go home from events in the evening with something more nickable and floggable than a couple of No. 4 Dressings. I'll protect my patients confidentiality but perhaps not at the point of a knife or looking down a tube with an inner diameter of 9mm or so.
It's a minefield of legal, technical and financial issues, however I think that the thing that will stop me worrying about it for some time is that for mobile users at least, you'll never get it to work properly anyway.