In the news

Of 83 papers on molecular genetic analysis published in the four major medical journals in 1995, few satisfied all the criteria that should be demanded in epidemiological research [Bogardus, S. et a. (1999) JAMA 281, 1919 (Abstract)]. The papers were analysed with reference to 7 criteria: reproducibility of techniques, objectivity of interpretation, delineation of cases, adequacy of spectrum in case group, delineation of comparison group, adequacy of comparison group, and quantitative summary of results. The most common failings were lack of documented reproducibility and lack of objectivity (for example failure to blind experimenters to case and control groups). At least 70% of the papers passed on the other criteria, though the authors comment that they were "lenient" in the definition of these criteria.

Comment: It is worrying that the quality of papers on molecular genetic analysis is not higher in the four major medical journals, which presumably are publishing the cream of such studies. "Caveat lector" certainly seems an appropriate warning for similar types of papers published in many of the more specialist medical journals. An additional "health warning" concerns the nature of the quantitative summary of results that is presented in molecular genetic analyses. Bogardus et al checked whether the papers reported "both the magnitude of the difference and a statistical expression of numerical stability", such as odds ratios and 95% confidence intervals; in some cases, however, the numbers are so small and the confidence intervals so enormous that one wonders whether the reported result really is meaningful. Bogardus et al point out the importance, from a public health standpoint, of maintaining high standards in molecular genetic analyses: studies such as those reported might become the basis for widely-used genetic tests.