Integrating epidemiology, psychology, and economics to achieve HPV vaccination targets

Abstract

Human papillomavirus (HPV) vaccines provide an opportunity to reduce the incidence of cervical cancer. Optimization of cervical cancer prevention programs requires anticipation of the degree to which the public will adhere to vaccination recommendations. To compare vaccination levels driven by public perceptions with levels that are optimal for maximizing the community's overall utility, we develop an epidemiological game-theoretic model of HPV vaccination. The model is parameterized with survey data on actual perceptions regarding cervical cancer, genital warts, and HPV vaccination collected from parents of vaccine-eligible children in the United States. The results suggest that perceptions of survey respondents generate vaccination levels far lower than those that maximize overall health-related utility for the population. Vaccination goals may be achieved by addressing concerns about vaccine risk, particularly those related to sexual activity among adolescent vaccine recipients. In addition, cost subsidizations and shifts in federal coverage plans may compensate for perceived and real costs of HPV vaccination to achieve public health vaccination targets.

Fig. 1.

Flow diagram for the model. Four types of HPV infection (16, 18, other high-risk types, and low-risk types) were simulated among seven age classes, including the most common co-infections and type-specific immunity. See SI for equations and sexual activity patterns among males and females among age classes.

Fig. 2.

Percent cases averted through vaccination in a scenario with a 95% effective vaccine conferring lifelong immunity, without changes to screening or sexual risk behavior. Vaccine-preventable HPV infections (types 16, 18, 6, and 11) constitute 77% of cervical cancers, 39% of cervical intraepithelial neoplasias grade 1 (CIN1), 60% of CIN2/3, and 90% of genital warts cases at baseline (0% vaccination). Solid lines reflect female-only vaccination (red, CIN1; orange, CIN2/3; green, cancer; blue, warts); dashed lines reflect vaccination of both genders. The transmission dynamics of low-risk HPV types differs from that of the high-risk types, hence the impact of vaccination on warts produces a qualitatively different trajectory from vaccination's impact on cancer and precancerous lesions. Cases averted plateaus due to other strains not covered by the vaccine.

Fig. 3.

Varying the conversion ratio between dollars and QALYs. The social optimum vaccination level (blue) and Nash female vaccination level (orange) were compared while the willingness to pay for each QALY was varied. The difference in cancer incidence per 100,000 between the Nash and optimum levels (green) peaked after the utilitarian reached a maximum threshold of cancer cases averted (while higher socially optimal levels of vaccination conferred further benefits from reduced warts at higher levels of willingness to pay).

Fig. 4.

Nash vaccination levels are displayed against different perceived risks of vaccination. “Without non-HPV promiscuity risks” describes risks of other major STDs (HIV, herpes, gonorrhea, Chlamydia, syphilis) and teenage pregnancy under the perception that HPV vaccination would increase risky adolescent sexual behavior; “HPV promiscuity risks” similarly refer to HPV-related health consequences from promiscuity; “subsidizing financial costs” refers to full coverage of the uninsured by the VFC and vaccine-manufacturer programs, as well as coverage of non-vaccine costs to patients (patient time costs, administration fees, and doctors' visit costs).