Abstract
Objective. To assess the cost-effectiveness of using human papillomavirus testing (HPV triage) in the management of women with minor cytological abnormalities in Sweden. Design. An economic analysis based on a clinical trial, complemented with data from published meta-analyses on accuracy of HPV triage. The study takes perspective of the Swedish healthcare system. Setting. The Swedish population-based cervical cancer screening program. Methods. A decision analytic model was constructed to evaluate cost-effectiveness of HPV triage compared to repeat cytology and immediate colposcopy with biopsy, stratifying by index cytology (ASCUS = atypical squamous cells of undetermined significance, and LSIL = low-grade squamous intraepithelial lesion) and age (23–60 years, <30 years and ≥30 years). Main outcome measures. Costs, incremental cost, incremental effectiveness and incremental cost per additional high-grade lesion (CIN2+) detected. Results. For women with ASCUS ≥30 years, HPV triage is the least costly alternative, whereas immediate colposcopy with biopsy provides the most effective option at an incremental cost-effectiveness ratio (ICER) of SEK 2,056 per additional case of CIN2+ detected. For LSIL (all age groups) and ASCUS (23–60 years and <30 years), HPV triage is dominated by immediate colposcopy and biopsy. Model results were sensitive to HPV test cost changes. Conclusion. With improved HPV testing techniques at lower costs, HPV triage can become a cost-effective alternative for follow-up of minor cytological abnormalities. Today, immediate colposcopy with biopsy is a cost-effective alternative compared to HPV triage and repeat cytology.
Acknowledgements
We thank Marc Arbyn (Unit of Cancer Epidemiology, Scientific Institute of Public Health, Brussels, Belgium) for discussions on the study design and Fredrik Borgström (Karolinska Institutet, Stockholm, Sweden) for assistance with the model. Financial support was provided by grants from the Swedish Cancer Foundation (070623, CAN 2007/1044), the Swedish Research Council (521-2008-2899), the Medical Research Council, the Cancer Society in Stockholm, the Stockholm County Council, the Swedish Labour Market Insurance and KI Cancer Strategic Grants (5888/05-722).
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.