Seven days in medicine: 3-9 May 2017BMJ 2017; 357 doi: https://doi.org/10.1136/bmj.j2263 (Published 11 May 2017) Cite this as: BMJ 2017;357:j2263
All rapid responses
We teach anaesthetists-in-training to always check the face validity of any drug calculations they make. 'Does the dose I've just worked out seem reasonable for a child of about one third/quarter/fifth the weight of a normal adult?' 'Does the amount of intravenous fluid I'm giving seem reasonable, in terms of teacup equivalents, for this frail elderly lady?' We encourage them to do the same when reading medical journals: 'Do the numbers presented seem to add up?
The short article headed, 'Quinine for muscle cramps may increase death risk' was eye-catching and quoted study data in some detail. Unfortunately not knowing the sample size of the non-quinine group (& not realising that it was quoted later in the journal) left us unable to assess the face validity of this unexpected and seemingly important finding. The numbers of course might be correct but we'd have to assume a much larger control group - and why would that be? We went to the JAMA abstract to seek clarification and found the number in the non-quinine group was indeed much larger, but yet again were presented with data that superficially appeared discordant - lacking in face validity. Patients in the study and control groups appeared to be broadly similar in most respects but were frequently reported as being statistically very significantly different ('p' often <0.001). Also, although more than 40% of patients were listed as hypertensive, the median number of prescriptions for diuretics, B-blockers, calcium channel blockers & statins was in each group and every case shown as zero; but with the differences between the two groups for all these medicines being reported as highly statistically significant. Possible - but odd. With no immediate access to the full text of the paper, we segued to the next article leaving the issue of quinine unresolved - and quite likely to be forgotten.
As the needs to keep up-to-date and also to guard against research fraud and 'fake news' become ever more important, perhaps in addition to encouraging readers to check face validity, we should also remind writers that its lack in abstracts, excerpts and articles may prove a stumbling block for an otherwise excellent paper.
Competing interests: No competing interests