Presenting probabilistic information in different formats: Effects of format preferences on comprehension and attitudes toward participation in a randomized clinical trial

Objective. To test whether a "match" or a "mismatch" between subjects' preferences for receiving probabilistic information in a particular format (i.e., in verbal or numeric terms) and the format in which they actually receive that information affects: (a) comprehension of the information provided in a hypothetical consent form designed to elicit informed consent to participate in a randomised, controlled, clinical trial (RCT); and (b) attitudes toward participating in the RCT.; Methods. A convenience sample of 228 subjects received pre-assembled, randomised packages containing a sham consent form about a hypothetical RCT. The consent form contained the standard information, with probabilistic information about risks and benefits of participation presented in either verbal or numeric format (the Intervention). Enclosed questionnaires were used to assess comprehension, format preferences, and other possible co-variates. The effect of format preference "match" or "mismatch" was subsequently tested in terms of differences in comprehension scores and attitudes toward participation in the RCT.; Only 33.0% of subjects achieved correct responses on all comprehension items and would have provided a fully informed consent or refusal to participate in the hypothetical RCT. The aggregated subgroups who received information in their preferred format ("match") did not differ in overall comprehension from the aggregated subgroups who did not ("mismatch"). However, comprehension was associated with the type of match/mismatch (p = <.01), and with the format received (p = <.01). Among those receiving the verbal format, comprehension scores were lower. Furthermore, the subgroup who preferred the verbal format and received a "match" scored lower than the subgroup who preferred the verbal format and received a "mismatch". No significant relationships emerged among received format, format preferences, and attitudes toward clinical trial entry. The co-variate of education level was significantly related to comprehension score (p = <.01).; Conclusion. Comprehension of quantitative probabilities may be affected by three interacting factors: information format, format preference, and level of education. The sources and stability of format preference, and the effects of information "match/mismatch", should be further investigated in other clinical contexts. Efforts to improve comprehension of probabilistic information during the RCT consent process need to be strengthened.