Data entry is often unreliable. Has the accuracy of the data we are being compared with been validated? Could the differences between our data and others have arisen by chance? This is the most revealing question. Imagine a thought-experiment in which 4 equal doctors use different strategies for predicting whether a tossed coin will land heads or tails. One always chooses heads, one always chooses tails, and the other two alternate their choices out of synchrony with each other. When I did this experiment for a pre-decided 14 throws each (56 throws in total), the best doctor only had 2 errors, whereas the worse had 7 errors over 3 times the rate for post-operative deaths (or whatever). The public would demand that this doctor be retrained or struck off, and the General Medical Council might feel obliged to comply, simply to keep public confidence (it is under great pressure to do something). So must we all be prepared to be sacrificial lambs? The answer is Yes, but there are certain steps that can be taken to mitigate our own and our patients risk-exposure.
- When we encounter doctors who are clearly underperforming (eg due to addictions) we must speak out. This will encourage belief in the system.
- For statistical reasons any series with <16 failures might be best ignored. Such series simply do not have enough power to detect real effects.
- We must strive to be both kind and honest with our patients. The best response to I'd like a home-delivery might be to say I haven't done one for 5 yearsand that one went wrong: are you sure you want my services? rather than the UK perinatal death-rate is the same for home and hospital.
- It might be the case that, contrary to the GMC, we should not always be on the look-out for colleagues who might be underperforming so that we can report them to the proper authorities: rather we should be encouraging an atmosphere of mutual support and trustthe sort of environment in which doctors feel safe to say all my cases of X seem to be going wrongcan anyone think why? To stop this trust turning into cronyism we must be prepared to engage in, or be subjected to, audit (p 506). The alternative is for clinicians to develop into secret police, informers, and counter-informers. No one would benefit from this. We note that malicious informing is not an isolated occurrence in the UK; 80% of those suspended for presumed underperformance are exonerated, but few return to their previous job owing to the stresses enquiries always engender.
Advertisements
