Whilst the power calculation that Tim presents is accurate, and undeniably
daunting, I feel that we should not be deterred from seeking some sort of
answer to this question. Mike does not seem to be asking for class one
evidence, and even in these days of homage to the RCT deity it is often the
case that other types of study design are the only practical way of
addressing a particular question which, nevertheless, deserves some sort of
evidence-based answer.
It is interesting to note that none of the trauma-related "senior
involvement" papers so far identified by the list are RCTs, but I suspect
that most of us would agree that the evidence that is available would favour
senior involvement in major trauma. This is further supported by the
intuitive nature of such a finding, and it may be quite reasonable to assert
that the benefits of senior involvement are so obvious that research
evidence is not required. In some respects I would support such an
assertion, but we are currently contemplating radical changes in the way we
work (for example from a consultant-led to a consultant-based service), and
these will have such profound effects on us and our health budgets that it
would helpful to have some sort of evidence available. I suspect that this
is what Mike is really seeking, and in the context of a whole department's
activity is, I believe, achievable.
At the A&E research priority setting exercise held in Manchester three weeks
ago this question was widely recognised as being of considerable urgency and
importance. If we wish to expand consultant numbers (and retain our training
grades) then evidence that such an investment will prove worthwhile can only
support our cause.
Jonathan Benger.
SpR, Bristol.
----- Original Message -----
From: Timothy J Coats (SURG) 7728 <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, October 26, 2000 8:58 PM
Subject: Re: Senior involvement in acute medical care
> I think that we should probably not even be asking for such evidence.
>
> As an example. Our Department sees about 90,000 patients a year.
> Of these I estimate that about 5% (4500) are 'sick' (ie. outcome
> lived/died can be influenced by treatment seniority). Of these 4500 I
> estimate that 150 die in the first 48 hours (the time in which
> seniority of A&E doc will have most effect). As a proportion of the
> 'sick' this is 0.0334.
>
> Imagine that having a senior A&E Doc present prevents 20% of
> deaths!!!!! If so the number of deaths per year would be 120 (0.0267
> as a proportion).
>
> A sample size calculation on this (unrealistically large) difference
> tells us that with a Power (Beta value) of 80% and an Alpha value of
> 5% we would need 10500 'sick' patients to be seen by a consultant
> and 10500 'sick' patients to be seen by a junior to find this
> difference.
>
> As 'sick' patients are 5% of the workload, this study would require
> the randomisation of about half a million patients.
>
> If the ability of a senior A&E doctor to prevent death is less than
> 20%, then the number of patients required would go up substantially.
>
> My advice would be that we should not require Level 1 evidence in
> the debate about the effect of seniority of A&E care on mortality
> unless we are prepared to do this study!!
>
> Tim.
>
> > I'm not aware of any study looking at the benefits of having Senior
> > A&E doctors seeing acute medical patients. I would have thought it was
> > self evident that there would be an improvement, because you wouldn't
> > have the most inexperienced doctors seeing the sickest patients.
> > Probably preferable to saying the "thickest seeing the sickest".
> >
>
>
> Timothy J Coats MD FRCS FFAEM
> Senior Lecturer in Accident and Emergency / Pre-Hospital Care
> Royal London Hospital, UK.
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|