Misreporting of contraceptive hormone use in clinical research participants

OBJECTIVE: Researchers traditionally rely on participant self-report for contraceptive use. We hypothesized that self-reported contraceptive use by clinical research participants may disagree with objectively measured hormonal status. STUDY DESIGN: We enrolled women in Harare, Zimbabwe, aged 18–34, who by self-report had not used hormonal or intrauterine contraception for >30 days, or depot medroxyprogesterone acetate for >10 months, into a study designed to assess biologic changes with contraceptive initiation and use. Blood samples obtained at enrollment and each follow-up visit (N=1630 from 447 participants) were evaluated by mass spectrometry for exogenous hormones. We individually interviewed a subset of participants (n=20) with discrepant self-reported and measured serum hormones to better understand nondisclosure of contraceptive use. RESULTS: Discrepant with self-reported nonuse of hormonal contraception, synthetic progestogens were detectable in 120/447 (27%, 95% confidence interval 23%–31%) enrolled women. Measured exogenous hormones consistent with use of contraceptive pills (n=102), injectables (n=20) and implants (n=3) were detected at enrollment, with 7 women likely using >1 contraceptive. In-depth interviews revealed that participants understood the requirement to be hormone free at enrollment (100%). Most (85%) cited partner noncooperation with condoms/withdrawal and/or pregnancy concerns as major reasons for nondisclosed contraceptive use. All interviewed women (100%) cited access to health care as a primary motivation for study participation. Of participants who accurately reported nonuse of hormonal contraception at enrollment, 41/327 (12.5%) had objective evidence of nonstudy progestin use at follow-up that disagreed with self-reported nonuse. CONCLUSIONS: Women joining contraceptive research studies may misrepresent their use of nonstudy contraceptive hormones at baseline and follow-up. Objective measures of hormone use are needed to ensure that study population exposures are accurately categorized. IMPLICATIONS STATEMENT: Among Zimbabwean women participating in a contraceptive research study, 27% had objective evidence of use of nonstudy contraceptives at enrollment that disagreed with self-report. Studies that rely on self-report to identify contraceptive hormone exposure could suffer from significant misclassification.