Memory problems around the menopause are an interesting area for study.1 However, it is hard to see where the paper in the June issue of BJGP fits in, and what it has to offer. There is very little data on the women invited to take part: all we are told is that 400 women were randomly selected from the general practice patient list — what is this, a primary care trust-held list? There is no detail as to the number of women in te age bands, apart from that they were women aged 25-64 years. The response rate is extremely disappointing and really it is very hard to see how a study with a reported response rate of 50.5% can give any information on the condition under study. In fact, this is actually compounded by the fact that of the 202 women who did respond, 30 failed to meet the inclusion criteria. So, really, the overall response rate from the 400 women invited was only 43% (although presumably there would be other women in the random group who also did not meet the inclusion criteria).
We do not pretend to have any great understanding of the statistics involved, but looking at the low response rate and the small numbers in each group, it is hard to see how any statistical significance can be achieved. There is also no explanation of why 400 women were invited. Was there a power calculation on which this number was based? Common-sense suggests that no conclusions of any sort could be drawn from this study; although obviously disappointing for the authors, this might also demonstrate the difficulty of undertaking a study such as this.
Of course, it is tempting to conclude that the women with the real memory problems simply forgot to return the questionnaire!
- © British Journal of General Practice, 2004.