background: given that there are different ways of setting out clinical test results, this study is aimed at ascertaining whether the way in which these results are set out has any bearing on the intention to prescribed on the part of primary care physicians. methods: randomized, multi-center cross-study. following a random sampling of eight primary care teams in the province of ourense (spain), two groups of family physicians (na=45 and nb=51) were surveyed by means of a questionnaire on which the data from five published clinical tests (three on heart disease prevention, one on cognitive benefit in a dementia syndrome and another on preventing fractures among the elderly) was shown. one of the groups was shown the data as a relative lowering of the risk, and the other as the necessary number of individuals to be treated. age, gender and years of practice were also asked. following a three-week blanking period, the intervention was reversed. the intention of prescribing was gathered on a 0-10 likert scale. the c2, simple student t were used, or the pearson test for paired data and correlation, according to whether an error a de 0.05 were to be achieved. results: test a: ci95% (-1.55~0.17); b: ci95% (-2.75~-1.20); c: ci95% (-0.16~1.65); d: ci95% (-0.30~1.44); e: ci95% (-1.22~0.35). no difference with regard to gender, age or professional experience were found. regarding the overall effect, with the exception of one of the tests presented, no significant differences were found concerning the intention of prescribing the drugs involved in terms of the way in which the data was set out. conclusions: the way in which clinical test data is set out has no bearing in itself on the ntention to prescribe medications.