[Descriptive study of statistical methods in the original article published on the cigarette smoking habit in four Spanish medical journals (1985-1996)].

BACKGROUND

Being the tobacco use a high-priority subject of investigation and having itself increased the utilization of statistical techniques in biomedical publication the used statistical techniques are described and the statistical accessibility is quantified in the original articles on tobacco use published in four Spanish medical journals.

METHODS

Retrospective descriptive study of 154 original articles on the cigarette smoking habit published in 1985-1996 in the journals Atención Primaria, Medicina Clínica (Barcelona), Revista Española de Salud Pública and Revista Clínica Española. An only observer codified the statistic techniques in 14 categories in agreement with the classification processed by Carré et al (1995) from the classification settled down by Emerson and Colditz (1983). The knowledge of bivariable techniques, to simple lineal regression, was stablished as the reference for the study of statistical accessibility.

RESULTS

81.8% original articles used inferential statistics. The most frequently used categories were "Contingency tables" (37.0%), "Descriptive statistics" (18.2%) and "Life tables and analysis of survival" (9.7%). A reader familiarized with bivariable techniques has statistical access to 96.0% for the originals of Revista Española de Salud Pública, 86.2% of Atención Primaria, 66.7% of Medicina Clínia (Barcelona) and 33.3% of Revista Clínica Española. The same reader had statistical access to 100% for the originals published from 1985 to 1987 and 68.1% from 1994 to 1996.

CONCLUSIONS

The use of statistical methods depends on the investigation subject and design, the journal and the year of the publication. The decrease of the statistical accessibility recommends to identify the profile of the standard reader in Spain, to adjust his knowledge to the current biomedical literature demand.