Optimising cervical cancer screening programs: how to evaluate incremental changes?

N. Wentzensen

DOI: https://doi.org/10.1111/1471-0528.14185

2017-02-01

Abstract:Cervical cancer screening programmes are currently undergoing important changes. The understanding of the causal role of HPV in cervical carcinogenesis has led to development of two important prevention approaches, HPV vaccines and HPV testing. Understanding of cervical carcinogenesis has also led to discovery of several biomarkers for cervical precancers (Wentzensen et al. J Clin Virol 2016; 76:S49–55). Deciding how best to use a growing number of tools for cervical cancer prevention is challenging. The article by Pedersen and colleagues describes a cost-effectiveness analysis of triage options for women between age 25 and 33 with minor cytological abnormalities in Norway. The authors demonstrate that several biomarker-based approaches are more cost-effective than currently recommended screening of young women in Norway. Importantly, the analysis focused only on a part of the overall screening population, women age 25–33. Typically, HPV prevalence peaks around age 20 and strongly declines thereafter; most of these HPV infections are transient and not related to precancer (Schiffman et al. J Natl Cancer Inst 2011;103:368–83). Therefore, it has been suggested that HPV-based screening should not start before age 30, requiring a different screening strategy for younger women. In the USA, HPV-cytology co-testing is recommended for women age 30 and older, whereas women age 21–29 should be screened with cytology (Saslow et al. CA Cancer J Clin 2012; 62:147–72). However, this age threshold is arbitrary and is based on the assumption that management of HPV-positive women is uniform across all ages. Recently, a new screening algorithm was approved in the USA including primary HPV testing starting at age 25 (Huh et al. Gynecol Oncol 2015;136:178–82). Despite the higher HPV prevalence in women age 25–30, a sufficiently specific triage test can limit colposcopy referral in these women. It would be good to see the costeffectiveness analysis by Pedersen et al. extended to evaluate primary HPV screening with various triage options starting at age 25. Importantly, the precision of predicting risk of cervical precancer using currently available tests needs to be balanced against the complexity of risk-stratified screening options. Already now, there is considerable confusion among providers and women about the best approach, which can result in more than necessary testing of some women and less than required testing in others. If age at first screening is 25, as in Norway, a simpler option would be to implement a single primary screening approach for the whole population. Management of positive screening results can be modified for younger women, e.g. by using a more specific triage test and more conservative treatment approaches. It is not feasible to conduct randomised trials for each of these questions. The analysis by Pedersen et al. demonstrates the importance of cost-effectiveness modelling to support decisions, particularly about incremental changes, in cervical screening programmes. Although the underlying evidence driving decisions about screening programs in different countries is basically the same, parameters for costeffectiveness modelling can vary substantially between settings and may lead to different recommendations. To optimise cervical screening programmes, integration of evidence from clinical trials, observational studies, and setting-specific costeffectiveness analyses is crucial.