Dietary treatment of hyperlipidaemia. Diets were poorly evaluated.

EDITOR, - It is ironic that Douglas G Altman's editorial castigating poor medical research1 should appear one week after the paper by Angela A Rivellese and colleagues, which concluded that two test diets “are suitable for treatment of hypercholesterolaemia.” 2 There are many problems with this study. The study is indeed controlled in the sense that its design permitted a valid comparison of the low total fat and high unsaturated fat diets. The statistical power was not, however, sufficient to exclude a meaningful difference between the diets, as the authors maintain. The confidence intervals indicate that the effect of the two diets on total cholesterol concentration, for example, could easily differ by as much as 7%. To put this in context, the efficacy claimed for the diets when compared with the “control” phase was only 8-9%. The authors' main message, however, had nothing to do with the comparison of the diets but focused on the supposed efficacy …