The potential cost-effectiveness of prophylactic human papillomavirus vaccines in Canada

Abstract

Aim

Clinical trials have shown prophylactic human papillomavirus (HPV) vaccines to be effective against infection and disease. We examined whether HPV vaccination has the potential to be cost-effective.

Methods

A cohort model of the natural history of HPV was developed, which fits simultaneously Canadian age and type-specific data for infection, cervical intraepithelial neoplasia, cervical cancer (CC) and genital warts (GW). Quality-Adjusted Life-Years (QALYs) lost and costs were estimated using data from the literature.

Results

Vaccinating 12-year-old girls (efficacy = 95%, no waning, cost/course = CAN$ 400) against HPV-16/18 and HPV-6/11/16/18 is estimated to cost the health provider CAN$ 31,000 (80%CrI: 15,000–55,000) and CAN$ 21,000 (80%CrI: 11,000–33,000) per QALY-gained, respectively. Results were most sensitive to age at vaccination, duration of vaccine protection, vaccine cost and QALY-lost due to GW, and were least sensitive to the medical costs.

Conclusion

Vaccinating adolescent girls against HPV is likely to be cost-effective. The main benefit of vaccination will be in reducing CC mortality. However, unless screening is modified, the treatment costs saved through vaccination will be insignificant compared to the cost of HPV immunization.

Introduction

Human papillomavirus (HPV) causes cervical intraepithelial neoplasia (CIN), cervical [1] and other anogenital cancers (vulva, vaginal, anus, penile) [2], [3], head and neck cancers [4], genital warts [5], [6], and recurrent respiratory papillomatoses [7], [8]. In Canada and other developed countries, cervical cancer screening programmes have substantially reduced the incidence and mortality of cervical cancer over the past 50 years [9]. However, the marked declines seen until the 1990s have been slowing in recent years [9]. In Canada, cervical cancer is currently the third most common cancer in women aged 20–49 and, each year, there are approximately 1400 new cases and 400 deaths from the disease [10]. High risk types HPV-16/18 account for approximately 70% of all cervical cancers [11], [12], [13]. Low oncogenic risk types HPV-6/11 are responsible for approximately 90% of genital warts [6], [14].

Two HPV prophylactic vaccines, which target HPV-16/18 (Cervarix^®) and HPV-6/11/16/18 (Gardasil^®), have been shown to be highly effective in clinical trials [15], [16], [17], [18]. With promising safety and efficacy results from these trials and the licensure of Gardasil^® in Canada, the US and many other countries [19], [20], [21], policymakers will be asked to make recommendations and decisions regarding the introduction of HPV vaccines. The main criteria considered in such decisions include safety, effectiveness, cost-effectiveness, affordability, programmatic feasibility, equity, public preferences, and the political consequences of decisions [22], [23]. In this study, we focus on examining the potential cost-effectiveness of prophylactic HPV vaccination in Canada. The goal of cost-effectiveness analysis is to compare the health and economic impact of different interventions in order to identify which interventions maximize the health of the population, in a context of limited resources. The specific policy questions that we examine in this study are: What is the cost-effectiveness of introducing HPV vaccination, under the current conventional cytology-based screening programs in Canada? What is the relative cost-effectiveness of a quadrivalent vaccine (HPV-6/11/16/18) compared to a bivalent vaccine (HPV-16/18)? What is the impact of age at vaccination on the cost-effectiveness results?

Because many of the benefits of prophylactic HPV vaccines occur in the medium to long term, mathematical models are needed to project the impact of vaccination beyond the time horizon of clinical trials. The development of models are based on assumptions, which necessarily introduce uncertainty regarding the conclusions that can be drawn from their results [24]. It is therefore important to examine the uncertainty of model predictions to provide policy makers with the necessary information to make appropriate decisions. In the case of HPV, it is particularly important to quantify uncertainty due to the complex natural history of HPV infection (encompasses numerous stages of disease which depend on HPV-type, screening and treatment) and the limited data on age and type-specific HPV natural history. Given these considerations, an additional aim of this study is to quantify the uncertainty around model predictions.