Aim: To assess a self-administered tampon specimen as an alternative method of detecting cytological abnormalities and its acceptability in comparison with a conventional Papanicolou (pap) smear.
Design: Comparative observational study.
Setting/population: Two hundred and seventeen women were recruited from the colposcopy clinic of an outer urban public teaching hospital and from sexual health clinics at suburban and major metropolitan hospital clinics.
Methods: Participants inserted and immediately withdrew a tampon, then placed it into a vial of ThinPrep PreservCyt fluid. This was analysed by a local private pathology laboratory. Results were compared to a pap smear performed the same day or within the previous 6 months. All women with an abnormal result (tampon or pap smear) underwent a colposcopy, with or without biopsy as necessary. Participants completed a questionnaire after performing the tampon test.
Outcome measures: Probabilities of tampon test detecting (i) a high grade abnormality (pHG), (ii) any cervical intraepithelial neoplasia (CIN) changes (pCINany), and (iii) any abnormalities (pabn) compared to the conventional pap smear and, if abnormal, compared to the biopsy taken at colposcopy. Acceptability of the tampon test and conventional pap smear were also measured.
Results: Probabilities of the tampon test compared to pap smear: pabn sensitivity 33%, specificity 89%, PPV 59%, NPV 73%; pCINany sensitivity 23%, specificity 97%, PPV 71%, NPV 79%; pHG sensitivity 19%, specificity 98%, PPV 63%, NPV 89%. Acceptability for tampon test was 91.21% and for pap smear, 45.85%.
Conclusions: Although the self-administered tampon ThinPrep method is a poor detector of cervical abnormalities compared to pap smear, it is highly acceptable to women. It has a relatively good negative predictive value (NPV). Our study suggests that if a more acceptable, sensitive method of cervical screening was found, which removed some of the existing barriers to conventional pap testing, screening rates for cervical cancer may improve.