OBJECTIVES: We evaluated the relative sensitivity for HPV detection of self-collection compared with practitioner-collected cervical specimens in the context of the Australian National Cervical Screening Program (NCSP).
STUDY DESIGN: 303 women aged ≥18 years attending a single tertiary referral centre took their own sample using a flocked-swab, and then had a practitioner-collected sample taken at colposcopy. All samples were tested at a single laboratory on the six PCR-based HPV assays which can be utilised in the NCSP; Roche cobas 4800 and cobas, Abbott RealTime, BD Onclarity, Cepheid Xpert, and Seegene Anyplex.
RESULTS: HPV16/18 results had high observed agreement between self- and practitioner-collected samples on all assays (range: 0.94-0.99), with good agreement for non-HPV16/18 oncogenic HPV types (range: 0.64-0.73).
CONCLUSIONS: Self-collection for HPV-based cervical screening shows good concordance and relative sensitivity when compared to practitionercollected samples across assays in the NCSP.
Methods: We assessed study-related records to determine the pace of data collection, response from potential participants, and feedback following data and sample collection. Overall and stratified measures of data and sample availability were summarised. Crude prevalence of key risk factors was examined.
Results: Approximately half (49.5%) of invited individuals consented to participate in this study, for a final sample size of 203 (161 adults and 42 children). Women were more likely to consent to participate compared with men, whereas children, young adults and individuals of Malay ethnicity were less likely to consent compared with older individuals or those of any other ethnicity. At least one biological sample (blood from all participants - finger-prick and venous [for serum, plasma and whole blood samples], hair or urine for adults only) was successfully collected from all participants, with blood test data available from over 90% of individuals. Among adults, urine samples were most commonly collected (97.5%), followed by any blood samples (91.9%) and hair samples (83.2%). Cardiometabolic risk factor burden was high (prevalence of elevated HbA1c among adults: 23.8%; of elevated triglycerides among adults: 38.1%; of elevated total cholesterol among children: 19.5%).
Conclusions: In this study, we show that it is feasible to create biodata resources using existing HDSS frameworks, and identify a potentially high burden of cardiometabolic risk factors that requires further evaluation in this population.