Evidence-based medicine can seem like a grey amorphous sludge when
you’re in the middle of a consultation - small wonder then to discover
that
doctors find it hard to tow the line as Choudhry (1) and Sweeney (2) have
cleverly pointed out. The evidence may be there but we choose not to
believe
it, or perhaps we may believe something else based on our experience.
So what is it that helps us make decisions with our patients? If it
is
experience, then how reliable is a meta-analysis of anecdote, for is not
experience just a collection of anecdotes? I think it is a balance
between the
number needed to treat (NNT) and the number needed to poison (NNP) – the
former can never be more than a paper exercise because we never will know
what illness we have prevented – we will only know what has happened, that
to say, what is tangible.
The following case illustrates this well. My GP Registrar rattles off
the
conclusions of the latest European secondary stroke prevention trial. P +
A
does the business better than P or A alone, and the NNT is only 18. Some
time later an elderly patient we started on P + A comes back to the
surgery.
She slams the packet down on my desk and says, “I’m not f…ing taking
anymore of them”. It takes some time to placate her. Shortly afterwards I
get
a call from anxious relatives. I struggle to explain the concept of
prevention
within the context of what has happened. Indeed, I am barely convinced
myself because the number needed to treat may only be 18, but, this gives
us
18 separate opportunities to poison a patient.
It is no surprise that anecdote is so powerful, nor is it a great
surprise to
know that the pharmaceutical industry, whilst funding most of the large
trials
that inform evidence-based medicine, is not only aware of our
susceptibility
to anecdote, but also to the general publics’ susceptibility to anecdote
based
medicine.
(1) Choudry NK et al. Impact of adverse events on prescribing
warfarin in
patients with atrial fibrillation: matched pair analysis. BMJ 2006;332:
141-3
(2) Sweeney K. Personal knowledge. BMJ 2006; 332: 129-130.
Competing interests:
None declared
Competing interests:
No competing interests
26 January 2006
james n hardy
GP principal
Bethnal Green Health Centre, 60 Florida Street, London E2 6LL
Rapid Response:
Numbers needed to poison
Editor
Evidence-based medicine can seem like a grey amorphous sludge when
you’re in the middle of a consultation - small wonder then to discover
that
doctors find it hard to tow the line as Choudhry (1) and Sweeney (2) have
cleverly pointed out. The evidence may be there but we choose not to
believe
it, or perhaps we may believe something else based on our experience.
So what is it that helps us make decisions with our patients? If it
is
experience, then how reliable is a meta-analysis of anecdote, for is not
experience just a collection of anecdotes? I think it is a balance
between the
number needed to treat (NNT) and the number needed to poison (NNP) – the
former can never be more than a paper exercise because we never will know
what illness we have prevented – we will only know what has happened, that
to say, what is tangible.
The following case illustrates this well. My GP Registrar rattles off
the
conclusions of the latest European secondary stroke prevention trial. P +
A
does the business better than P or A alone, and the NNT is only 18. Some
time later an elderly patient we started on P + A comes back to the
surgery.
She slams the packet down on my desk and says, “I’m not f…ing taking
anymore of them”. It takes some time to placate her. Shortly afterwards I
get
a call from anxious relatives. I struggle to explain the concept of
prevention
within the context of what has happened. Indeed, I am barely convinced
myself because the number needed to treat may only be 18, but, this gives
us
18 separate opportunities to poison a patient.
It is no surprise that anecdote is so powerful, nor is it a great
surprise to
know that the pharmaceutical industry, whilst funding most of the large
trials
that inform evidence-based medicine, is not only aware of our
susceptibility
to anecdote, but also to the general publics’ susceptibility to anecdote
based
medicine.
(1) Choudry NK et al. Impact of adverse events on prescribing
warfarin in
patients with atrial fibrillation: matched pair analysis. BMJ 2006;332:
141-3
(2) Sweeney K. Personal knowledge. BMJ 2006; 332: 129-130.
Competing interests:
None declared
Competing interests: No competing interests