Thursday, May 03, 2007

Smoke, Mirrors & Percentages

Nic Fleming’s article in The Daily Telegraph, reporting the work of Prof. Dennis Black, illustrates some important numerical confusions. I would draw attention to the second to last paragraph where the “throw away” comment says that patients in the trial had 3 times the risk of having a heart rhythm abnormality. If one makes a mathematical approximation to simplify these numbers for the purposes of illustration the expectations of this treatment may not be so readily recommended by NICE. If one assumes an incidence of 100 per population each for hip fractures, spine fractures and the abnormal heart rhythm then the percentage changes Prof Dennis work reportedly shows would mean the following. 41 less hip fractures, 70 less spine fractures and 200 more abnormal heart rhythms. My readers will appreciate that I have deliberately invented the incidence rate to illustrate a purpose and that these conditions' incidences are not all the same. I do however think it is an important point to make about the use of percentages, relative risk ratios and multiplier comparisons in the reporting of health matters. I personally would support a wholly more responsible level of reporting that at least attempted to reveal the truth.