Skip navigation

On Quality (editorial)



One of the sub-plots of 1997 for Bandolier was the issue of quality. People we talked to sometimes mused on quality past - things used to be better, we don't appreciate quality so much, and, from time to time, told us some frankly horrifying stories of where quality was overlooked or ignored. Yet in twelve issues we managed to fill pages based predominantly on some really high quality stuff, like the soaring practical applications of evidence and knowledge like getting pain relief after day case surgery ( Bandolier 40 ). And the increase in good systematic reviews and meta-analyses is astounding. Whereas a few years ago they had to be mined from the literature, it seems these days impossible to move without tripping over half a dozen.

Quality in EBH

Evidence-based healthcare is all about quality. Quality in searching to find the studies addressing a particular question, in applying appropriate quality filters to ensure only unbiased studies are included, in distilling the information into knowledge and making the knowledge understandable and useful with NNTs or likelihood ratios. There is another quality step as well, when the knowledge is combined with a practitioner's education and experience, knowledge of a patient, and the values of people and society to make sensible decisions.

But quality comes in different forms. Quality assurance is making sure not only that the right things get done, but that wrong things don't. In clinical practice, audit is a shield of quality.

Diagnostic imagining

Mistakes matter (like this one from the ITV commentary on the Oxford versus Cambridge rugby match), and will always happen. In clinical biochemistry, where quality assurance has become part of everyday working, there is a rich literature on mistakes. So this month we summarise some evidence on how often these mistakes, or blunders , can occur. Mostly they are simple errors of mislabelling inside or outside the laboratory, and about 3 results in a 1000 may be blunders.

Analytical errors occur much less frequently, at about 0.4 in 1000. But there is an important lesson from Australia, in which a survey showed that on average 11% (or 110 in 1000) of external quality assurance samples were outside the norm. How so? Probably because there is no constraint on laboratories there to be inside the norm. No carrot and no stick.

But finding good criteria and sticking to them can make dramatic improvements in performance. Bandolier revisits the histological diagnosis of melanoma , and finds reports of great improvements in agreement between pathologists when criteria are agreed, and then used.

New for '98

Bandolier is trying to take up the quality theme in 1998 for its Internet pages . These have been the print pages written in hyper-text. But some of the many thousands of users in the UK and overseas (200,000 visits a month) think that we could do better.

So we are using some of our precious resources to try to do better. It will take some months to organise, but we hope to carry reviews with the clinical bottom line at the top, and with simple summaries of the evidence from reviews in a reasonably standard format. The Internet version is free, so potential sponsors (public or private) might like to get in touch.

Risky business

Some favourable comments on the method of risk presentation that appeared in Bandolier 45 , with requests for more of the same. But this is tricky territory and more mining of good quality risk data is needed. One thought is for an international risk data exchange on the Internet from Bandolier's home pages . Is there anyone out there who has or knows of information, or where to get it? And again, because this is new territory, anyone know of how this might be funded?

And finally

There have been request for Bandolier to make a slide set available. It is possible for that to be done as a downloadable PowerPoint file from the Internet. But again it's work. So would anyone be interested?



next story in this issue