Comparison of type-specific human papillomavirus data from self and clinician directed sampling.

OBJECTIVE(S) To compare the type-specific human papillomavirus (HPV) recovery from physician and patient-collected samples. METHODS Three hundred thirty-four (334) women attending colposcopy clinics in three countries were enrolled in this cross-sectional study. Cervicovaginal samples were collected by patients and physicians and processed with polymerase chain reaction and reverse line blot genotyping. McNemar's Chi-squared tests and Kappa statistics were utilized to determine statistical associations between physician- versus patient-collected samples. RESULTS Oncogenic HPV infection was identified in 23.2% of patient-collected specimens compared to 34.9% of physician-collected specimens. Physician sampling detected significantly more infections with type 16 and 52 than did self-sampling and significantly more oncogenic HPV infection overall. For non-oncogenic HPV detection, there was no statistical difference between physician- and patient-collected samples. CONCLUSION(S) Patient sampling for HPV using a single vaginal brush does not identify all oncogenic HPV subtypes.