Nice trial – pity about the ethics committee?

Many trials comparing different teaching methods are of dubious quality (e.g. here), but this one (click here) in last week’s British Journal of Obstetrics and Gynaecology (BJOG) looked like it was going to be good – perhaps until the research ethics committee got involved.

Passing a vaginal speculum to take smears, fit coils etc., is a tricky skill to learn. Many medical schools let students practice on plastic models, which work reasonably well but are not very realistic. Others employ healthy women volunteers, who not only let students practice on them but teach, feed back on the student’s technique, and often become part of the teaching faculty.

Such gynaecological teaching associates (GTAs) are wonderful women, but do they really make for more skilful doctors? It would be good to have unbiased evidence. Unfortunately no firm conclusions could be drawn from the three poor quality randomised trials done to date. Hence the new trial.

It was prospectively registered here, with a planned sample size of 101 students (94 achieved) randomised in a 1.4: 1 (GTA: control) ratio, to maximise the use of available GTAs. The primary endpoint was the objective structured clinical exam (OSCE) score (range 0-54, high = good). The sample size was sufficient to show a shift in mean scores of about half a standard deviation. The OSCE was conducted with the student examining a model pelvis and scored by a GTA and a trained gynaecologist working together, neither of whom were aware of the student’s group. The result was negative. GTA training  made no difference. Median control group score 43, GTA group score 44, P=0.26. Oh dear!

But assessment on a plastic pelvis is surely not the best way to judge a student’s skill. Students trained on models may do fine in exams on models, but go to pieces when faced with the real thing, and only the person being examined can really judge if a doctor is being gentle. Why didn’t the researchers measure the student’s skill passing a speculum on a real woman, i.e ask a GTA to act as the exam model, as well as helping score student’s performance?

Perhaps they wanted to. There is a rumour that the researchers had planned to evaluate the students properly but were dissuaded by the research ethics committee, squeamish about something or other.

If true, this is outrageous – the trial may have given a false negative result, will need to be repeated, and more women will suffer avoidable discomfort and embarrassment – but it’s not easy to check. Queen Mary’s research ethics committee (click here) don’t publish their deliberations. Someone might speak up, but I’m not optimistic. No researcher wants to go on record about these sort of shenanigans; ethics committees have almost unlimited power to delay or meddle with projects.

Let’s hope I was misinformed.

Jim Thornton