Estimated health and economic impact of quadrivalent HPV (types 6/11/16/18) vaccination in Brazil using a transmission dynamic model

Every year, nearly 530,000 women develop cervical cancer and 275,000 die from the disease worldwide [1]. About 88% of deaths from cervical cancer occur in developing countries [1]. Human papillomavirus (HPV) is the primary cause of cervical cancer, with HPV-types 16 and 18 causing approximately 70% of cases globally [2]. A quadrivalent HPV vaccine against HPV-types 6, 11, 16, and 18 (Gardasil® by Merck & Co., Inc.) was recently developed and is widely available around the world [3‐5].

Cervical cancer is the second most common cancer among women in Brazil. The cervical cancer incidence rate is two to three times higher in Brazil (19.2 per 100,000 women every year) than in North America and Europe [6]. The quadrivalent HPV vaccine was approved for use in Brazil to prevent cervical, vulvar, and vaginal cancer, precancerous lesions, and genital warts caused by HPV types 6, 11, 16, and 18. However, HPV vaccination has not been introduced as a national program in Brazil. Understanding the health and economic impacts of HPV vaccination is essential for policy makers to make decisions regarding the introduction of national vaccination programs. Previous cost-effectiveness studies of HPV vaccination in Brazil and other middle-income countries have used a range of different methods [7‐11]. Because the mathematical transmission dynamic model is required to describe the transmission of HPV infections and diseases, we used this method to examine the effectiveness of HPV vaccination.

The purpose of our study was to examine the long-term health and economic impacts of quadrivalent HPV vaccination in Brazil using a transmission dynamic model. We evaluated the following vaccination strategies: routine vaccination of 12-year-old girls and routine vaccination of 12-year-old girls in combined with a catch-up vaccination of 12 to 26-year-old women.

We adapted a previously developed transmission dynamic model to evaluate the health and economic impacts of HPV vaccination in Brazil [12]. The transmission dynamic model incorporates the direct and indirect benefits (herd immunity) of vaccination. Details of the model have been previously published [12‐14]. Here, we briefly described the model structure and methods for adapting the model to Brazil.

Our study demonstrates that quadrivalent HPV female vaccination can substantially reduce the burden of cervical diseases and genital warts in Brazil. The major advantage of our analytic approach is the use of transmission dynamic model that incorporates the direct and indirect benefits of vaccination while evaluating effectiveness at the population level over time. We found that the routine vaccination of 12-year-old girls in combination with a catch-up vaccination of 12 to 26-year-old women can be a cost-effective strategy that can achieve earlier and greater reduction in HPV-related diseases than the routine vaccination. Routine vaccination in combination with a catch-up vaccination could prevent approximately 163,000 cases of cervical cancer, 48,000 deaths from cervical cancer, 2.3 million cases of CIN2/3, and 11.4 million genital warts in the next 50 years.

The World Health Organization considers an intervention to be "very cost-effective" when its incremental cost-effectiveness ratio is below GDP per capita [31]. We found that the incremental cost-effectiveness ratios for HPV vaccination strategies ranged from US$219 to US$450 per QALY gained, which fell below a GDP per capita (US$10,710 in Brazil). In order to address uncertainty of the parameters, we conducted extensive one-way sensitivity analyses. For an example, duration of protection remains uncertain. Even when we decreased the duration of vaccine protection to 20 years, we found that the HPV vaccination was cost-effective especially the vaccination strategy that included a catch-up vaccination. Four prior studies examined the cost-effectiveness of HPV vaccination in Brazil [7‐10]. Goldie et al. (2007) used individual-based stochastic models and showed that routine vaccination would cost I$120 to I$820 per year of life saved [7]. Kim et al. (2007) also found vaccination to be cost-effective using dynamic models [8]. Using the similar individual-based dynamic model by Kim et al., Vanni et al. (2012) recently found that the ICER for the quadrivalent HPV vaccination that incorporated the benefits of preventing genital warts to be US$255/QALY assuming similar cost of vaccination as ours ($55 for a total cost of vaccinating woman) with discounting at 5% [10]. Colantonio et al. (2009) used Markov models and found the ICER to be US$10,200 per QALY assuming US$210 for a total cost for vaccinating woman [9]. Higher ICER results may be because Colantonio et al. (2009) assumed higher vaccine cost than ours and used cohort model that did not take into account of herd immunity. In spite of differences in the model structure and assumptions about model parameters regarding natural history of HPV disease, vaccine property, health utilities, and costs, all studies consistently found HPV vaccination of females to be cost-effective in Brazil.

Contrary to most previous studies, we incorporated the potential impact of vaccination on HPV6/11-related genital warts. The quadrivalent HPV vaccine was projected to reduce the incidence of genital warts in a short period of time. This is consistent with a rapid decline in the incidence of genital warts observed among young women in Australia where vaccination has been already implemented [32, 33]. Although numerous studies from North America and Europe have shown that HPV female vaccination is generally cost-effective, our incremental cost-effectiveness ratios in Brazil were lower than that in those countries [34‐36]. Despite the low cost of treatment for HPV-related diseases, quadrivalent HPV vaccination in Brazil can be a cost-effective intervention, because it can prevent substantial burden of cervical cancer and genital warts.

Our model projected that HPV vaccination could prevent approximately 163,000 cases of cervical cancer, 48,000 deaths from cervical cancer and 11.4 million genital warts in the next 50 years. Many women who suffer from cervical cancer are young and actively caring for their families and it could have devastating consequences to their children and families who lose their mothers. Because we cannot incorporate such effects in cost-effectiveness analysis, our results likely underestimate the potential societal benefits of vaccination. Moreover, our model projected the substantial impact of HPV vaccination on genital warts, which may have long-term psychological and/or physical consequences and profoundly affect patient's quality of life [29, 37]. For example, even after treatment, women and men may experience anxiety of recurrence or persistence of genital warts [29]. Reducing cases of genital warts will not only reduce healthcare utilization but will also free up resources for diagnosis and treatment of other diseases.

Our study has several limitations. Although the model was built on available current knowledge on HPV diseases, more studies are needed to understand the natural history of HPV. Detailed data specific to Brazil, such as healthcare seeking behaviors and age- and stage-specific mortality rates from cervical cancer, were limited. Because health utility data were not available from Brazil, we used the health utility values from the U.S. Our model does not incorporate demographic changes in the future, such as population growth. Our model did not account for any temporal changes in screening practice or possible introduction of new screening methods such as HPV DNA testing. As some researchers have examined, it is important to consider screening changes in future research [38, 39]. Although the coverage rate of HPV vaccination is unknown, we assumed high coverage because Brazil has a strong national immunization program that has achieved high (>90%) immunization coverage in many currently scheduled vaccines [40].

We only assessed fit of the adapted model by assessing overall incidence of cervical cancer and did not assess age-specific data. We did not employ calibration method that identifies parameters that best predict observed data. Moreover, because of the complexity of the model, it was not feasible to conduct probabilistic sensitivity analysis.

Our projected incidence rates of HPV6/11-related genital warts were comparable to the previously reported incidence rates in other parts of the world [28‐30]. A survey conducted by the Ministry of Health in Brazil found that 5.7% of pregnant women reported having a history of clinical diagnosis of genital warts [25]. However, data regarding the incidence of genital warts in the general population in Brazil were not available.

HPV vaccine may have cross-protection against non-vaccine, oncogenic HPV-types; however, the duration and efficacy of the cross-protection remains uncertain [41]. If this additional benefit of vaccine was considered, we would have more favorable cost-effectiveness. We did not consider the potential benefits of HPV vaccination on other HPV-related diseases such as vulvar, vaginal, anal, head and neck, and penile cancers, and recurrent respiratory papillomatoses. Previous studies in the U.S. included these diseases and showed an improvement in incremental cost-effectiveness ratios [12, 42]. The HPV vaccination of males could also be important because it may prevent some of these cancers and genital warts.

In conclusion, our study demonstrates that quadrivalent HPV vaccination can substantially reduce the burden of cervical diseases and genital warts in Brazil. Our model results show that HPV vaccination of females, particularly routine vaccination of 12-year-old girls in combination of a catch-up vaccination of 12 to 26-year-old women, can be a cost-effective intervention in Brazil.