OBJECTIVE: We compared informed consent forms of subjects participating in biomedical research with those of references texts in order to determine the factors that influence readability. METHODS: We assessed the readability of 73 informed consent forms of research protocols conducted in the clinical investigation centres in the Rhone-Alpes area, and then compared them with 33 reference texts corresponding to 5 French school grades (first year infant, primary school, GCS level, high school, and classics aggregation), using the Flesch test and Cordial" analyser. RESULTS: Median Flesch scores were 66 for the first year infant level, 62 for the primary school level, 58 for the GCS level, 42 for the high school level, and 43 for the aggregation level. It was 22 for the informed consent forms. Median Cordial scores were 86 for the first year infant level, 77 for the second, 74 for the third, 49 for the fourth, 43 for the fifth. It was 1 for the informed consent forms. No methodological factor correlated with Flesch and Cordial" results. CONCLUSION: The quantitative readability scores for informed consent forms for subjects participating in biomedical research are low, lower than those proposed to aggregation candidates, whatever the type of protocol. Some thought must be given to the impact of the reduced readability on patients' understanding, and steps should be taken to improve the readability of the forms.