And I also worry about the misapplication of AI in EM - computer says diagnosis is/isn't xyz. AI/LLM could have useful roles in EM, but as a decision aid in diagnosing AI is open to the same input/output traps as current decision aids

As departments are under increasing pressures to stream undifferentiated patients from triage, I worry that patients may lose out on seeing a specialist generalist, and are put too early into rigid boxes of medical/surgical, or PE rule-out path/low risk abdo pain etc.

I've found PE on CTPA when low Well's/negative D-dimer, ICH in people not meeting NICE HI criteria, and significant cardiac disease in the under 30s. All of these required thinking beyond the decision aid

Until and unless we have a political class willing and able to state these basic truths about social care, reform its provision so it doesn’t carry on destroying local govt, and actually raise the taxes needed to fund it properly, we are, not to put to fine a point on it, f***ed on this issue.

And, I think residents are much more likely to pick up the phone for advice if they know the consultant hasn't been on shift since 8am +/- got a clinic/another 24hr on call the next day. I certainly became more willing to "disturb the boss" when the EM on call changed from 24hr to 1700 start.

It is also much more clinically satisfying, and arguably less risky, for one clinician/team to complete a care episode than to hand over part way through (when discharge is predicted) for the sake of a time target.

Maybe one answer is that, yes the 4hr standard is a reasonable aim for many patients, but a 6hr standard would fit a whole bunch more folk who await diagnostics/definitive management before discharge.