(P471) Comparison of computer-assisted self-interview (CASI) versus clinician-interview (CI) for self-reported vulvovaginal symptoms

Objective: Patients may feel uncomfortable reporting vulvovaginal symptoms. Computer-assisted self-interview (CASI) may be less susceptible to under reporting than clinician-interview (CI). Sparse data exists on how survey modalities affect report of vulvovaginal symptoms. Two published studies have found higher reporting of any current STI symptom (nonspecific) to CI than CASI. Therefore, we compared CASI versus CI for eight self-reported vulvovaginal symptoms.

Methods: This is a secondary analysis of N=101 participants enrolled in a longitudinal observational study to assess the vaginal microenvironment. Primary analyses were restricted to the first clinical visit in which a participant reported a vulvovaginal symptom to CASI, clinician or both (N=43 participants). Participants were emailed a CASI daily which queried various vulvovaginal symptoms (REDCap survey, Yes/No on itching, odor, discharge, dryness, irritation, burning, dysuria and dyspareunia). The CI administered the same survey prior to pelvic examination at three clinical visits. Symptoms were assessed for concordance when reported on CASI and CI on the same date. For each symptom, concordance was tested with kappa statistics, which accounts for agreement by chance. Sensitivity analyses were conducted to evaluate whether various study design factors affected concordance.

Results: Among 101 participants, 43% reported vulvovaginal symptoms. There were no demographic differences between those who did and did not report symptoms. Among those who reported symptoms (N=43), concordance between CI and CASI was 70-95% (κ=0.14-0.50). Itching (k=0.50) and odor (k=0.49) had the highest agreement. Of those who reported symptoms, 11 were diagnosed with bacterial vaginosis and 5 with vulvovaginal candidiasis. Because most participants did not report symptoms, concordance was higher between CASI and CI when the full cohort was assessed (N=101, range: 95-99%, κ=0.31-0.74). Utilizing the CASI from the day prior to the CI did not affect the results. Similarly, concordance was not affected if CASI was submitted late.

Conclusion: Among those with symptoms, concordance was moderate for report of vulvovaginal symptoms between CASI and CI. These data do not demonstrate superiority of CI or CASI for the ascertainment of self-reported vulvovaginal symptoms. Larger, longitudinal studies designed to test CASI reliability could explore these survey modalities further.