ABSTRACT
OBJECTIVE: We investigated whether participants in a phase II randomized clinical trial of a candidate vaginal microbicide ever intentionally misled interviewers. STUDY DESIGN AND SETTING: We used audio computer-assisted self-interviews (ACASI) to ask the South African women (n=132) participating in the trial about the accuracy of self-reported data collected during face-to-face interviews. The trial protocol recommended that women use their assigned gel (active microbicide or placebo) with condoms during each vaginal sex act. RESULTS: Nearly four-fifths of participants (n=104, 79%) reported that they had misinformed trial interviewers at least once. Motivations included politeness (n=45, 34% of ACASI participants) to avoid criticism or seek praise (n=32, 24%), and embarrassment (n=24, 18%). Participants acknowledged misreporting eligibility characteristics to enroll (11%) and, during follow-up, exaggerating their enthusiasm for the study gel (13%), applicator (13%), and the effect of the gel on sexual pleasure (13%). In general, women who were untruthful had actually used the gel with condoms less and used the gel alone more than they had reported during the trial. Women overwhelmingly found the computer survey easy. CONCLUSION: Researchers cannot assume that participants always tell the truth about sensitive behaviors in face-to-face interviews. ACASI was efficient and acceptable in this population.
