Cost-effectiveness of human papillomavirus (HPV) vaccination in Burkina Faso: a modelling study

We estimate that the current HPV vaccination programme in Burkina Faso, using GARDASIL-4®, will prevent a substantial number of cervical cancer cases and deaths. Our assumptions about cross-protection were influential. Without cross-protection, CECOLIN® is likely to be the preferred product, generating lower net costs and similar benefits to both GARDASIL-4® and CERVARIX®. With cross-protection, CECOLIN® also had the most favourable cost-effectiveness, but because CERVARIX® generated substantially more health benefits than CECOLIN® (66% versus 48% vaccine impact), and only slightly less favourable incremental cost-effectiveness (0.28 versus 0.19 times the national GDP per capita), this option should be given serious consideration if affordable. Our findings also suggest that GARDASIL-9® is unlikely to be a viable option unless the current (assumed) price per dose is substantially reduced.

Burkina Faso introduced GARDASIL-4® based on the recommendations of the NITAG after considering the epidemiology of cervical cancer and the availability and affordability of the different products. We assumed a societal perspective including costs borne by both the Government and Gavi. Under this scenario, we find there could be advantages in switching from GARDASIL-4® to either CECOLIN®, or, budget permitting, CERVARIX®. This assumes Burkina Faso will still have access to the vaccine prices assumed in this analysis when it graduates from Gavi support. It is important to emphasise the decision-making perspective considered in our analysis. If we had assumed a narrower payer perspective (excluding any vaccine costs borne by Gavi), then CERVARIX® would have been the most cost-effective option because it was estimated to have the highest health impact. This is because the government co-pay would be the same for any of the three Gavi-supported vaccines. However, irrespective of the decision-maker perspective (with or without costs borne by Gavi) our analysis suggests that switching the current HPV vaccine (GARDASIL-4®) could improve cost-effectiveness. Our analysis did not account for the switching costs involved in replacing GARDASIL-4®, such as health worker training, communication, and monitoring and evaluation. We also did not account for the potential impact of GARDASIL-4® on genital warts, which is one important distinguishing feature of this vaccine. In addition, there is substantial uncertainty around the efficacy and duration of protection of each vaccine product on each HPV type, which makes it difficult to definitively favour one product over another.

In our base case scenario we used HPV type distribution data from a cross-sectional study in Burkina Faso, rather than international database estimates for the African continent. Our cost-effectiveness estimates were substantially more favourable when we re-ran our analysis using the HPV type distribution estimated for the African continent because types 16/18 represented a far higher share of total cervical cancer cases (67% versus 39%). Updated estimates of the HPV type distribution in Burkina Faso would help to clarify the most prominent types in circulation and inform the optimal choice of vaccine product. Our base case assumptions on cross-protection were also uncertain. We included cross-protection against HPV types 52 and 56 but a study by Wheeler et al. [37] has suggested this effect might be due to chance observations. However, both types combined represent < 5% of the total cervical cancer cases in our analysis, so this assumption had a minimal effect on our results.

In our analysis, the vaccine with the most favourable cost-effectiveness (CECOLIN®) reported a cost per DALY averted in the range of 0.19–0.26 times the national GDP per capita in Burkina Faso. Based on historical thresholds of cost-effectiveness (e.g., 1 times national GDP per capita) this would be considered a highly cost-effective intervention. However more stringent and/or context-specific WTP thresholds are now recommended for low- and middle-income countries (LMICs) [41]. For example, Ochalek et al. [21] have proposed a WTP threshold for Burkina Faso in the range of 0.09–0.29 times the national GDP per capita. It is reassuring that CECOLIN® (with or with cross-protection) and CERVARIX® (with cross-protection) could both potentially represent reasonable value for money (< 0.29 times the national GDP per capita) despite our analysis assuming the full cost of the vaccine. In Burkina Faso, households make an important contribution to the healthcare costs associated with cervical cancer, and Gavi make an important contribution to the costs of vaccination. We therefore included both in our estimates of cost-effectiveness. Had we considered a strict ‘Government perspective’ fewer healthcare costs would have been averted (only those associated with radiation therapy), but the cost-effectiveness ratios would have been far more favourable because the Government only pays a small contribution towards the costs of vaccination e.g. $0.20 versus $2.90 per dose of CECOLIN. As the Government will eventually be expected to contribute the full cost of the vaccine, the results for this perspective could be misinterpreted, and the value of HPV vaccination over-stated.

Several studies have found HPV vaccination to be a cost-effective intervention in LMICs [42‐44]. A recent meta-regression analysis in 195 countries by Rosettie et al. [43] found that the mean predicted Incremental Cost-Effectiveness Ratio (ICER) for HPV vaccination was US$ 4,217 per DALY averted (95% uncertainty interval [UI]): US$ 773–13,448) worldwide, and sub-Saharan Africa and South Asia had the lowest predicted ICERs, with a population-weighted mean ICER across 46 countries of US$ 706 per DALY averted (95% UI: US$ 130–2,245). However, thresholds based on the national average GDP per capita are likely to mask large sub-national inequalities in the cost and benefits of vaccination. Cost-effectiveness is just one of the decision criteria that should be considered by policy makers. Other criteria include equity, budget impact, feasibility, and acceptability [45].

The cost of the current annual vaccination strategy with GARDASIL-4® ranges from US$ 2.3 million (in 2022) to US$ 4.5 million (in 2031). This annual cost in 2022 represents 31% of the Government’s share of the total EPI budget in 2022, and 3.4% of the total EPI budget (including donor contributions) in 2022. In Ghana, the projected average annual undiscounted costs of the HPV vaccine programme represented 11–15% of the total immunization costs for 2022 [46]. We find that a single dose strategy could further reduce costs and improve cost-effectiveness. WHO’s Strategic Advisory Group of Experts on Immunization recently recommended the use of either one or two doses of CERVARIX®, GARDASIL-4®, and GARDASIL-9® [47]. The option to use a single dose is based on available evidence and therefore may not apply to CECOLIN® at this time.

Burkina Faso launched a nationwide HPV vaccination programme for 9-year-old girls in May 2022 without a catch-up campaign for girls aged 10–14 years. Scenarios with a catch-up campaign had a higher cost per DALY averted, but the cost-effectiveness ratios were still relatively favourable. This strategy should therefore be considered to protect girls currently aged 10–14 years who were not eligible for routine HPV vaccination at age 9 years. If costs are prohibitive, then a single dose strategy could be considered to reduce the cost. A further enhancement to the programme would be to include HPV vaccination for boys, but evaluation of this would require the use of a more complicated model, and was therefore outside the scope of our current study.

Our analysis was conservative (unfavourable to HPV vaccination) for several reasons. First, we assumed the full cost of the vaccines rather than simply assuming the proportion contributed by the government in the initial phase of Gavi support. Second, we assumed a limited societal perspective, excluding lost wages of patients and their families, and any health system healthcare costs borne by the government. Third, we probably underestimated the burden of cervical cancer because some sick women will not be included in official statistics. Fourth, we used a simple static cohort model, and have therefore not captured any additional indirect (herd immunity) benefits associated with vaccination. In contrast, our analysis may have favoured HPV vaccination by assuming relatively low incremental health system costs despite very high vaccine programme coverage for two doses (97.5% from year 3). Despite more than 87% of 9-year-old girls attending school in Burkina Faso [16], this optimistic level of programme coverage may still require higher incremental health system costs per dose than we have assumed in our analysis. Our estimates of the incremental health system costs also have excluded the cost of activities that were covered by a Gavi vaccine introduction grant. This was estimated to be US $1,033,385 in the first year of introduction and represents US$ 3.15 per vaccinated child, distributed between Gavi support (US$ 2.4 per vaccinated child), other partners (US$ 0.34 per vaccinated child), and the government (US$ 0.41 per vaccinated child) [30]. However, due to delays, programmatic challenges and exchange rate fluctuations, the governmental cost contribution was higher than planned. Finally, we assumed the counterfactual (without vaccination) rate of cervical cancer cases and deaths would remain constant with time.