Lay Summary
Background and objectives
The discovery of vaccines against Human Papilloma Virus (HPV) is an iconic example of pharmaceutical innovation [1]. Numerous cost-effectiveness studies agree in attributing a relevant social value to the selective vaccination of 12-year-old girls [2]. However, the incremental value of vaccination of male children remains controversial. This is due to the dominant value of selective vaccination, increased by the effect of acquired immunity. No evaluation has so far considered the impact of sexual behaviours and preferences that increase the risk of infection [3]. The objective of this study is to consider such behaviours as externalities and to evaluate their impact on the social value of the selective anti-HPV vaccination.
Methods
Social network analysis is the process of investigating social structures using networks and graph theory. It characterizes networked structures in terms of nodes and the links, edges, that connect them. Network analysis has been useful to simulate the dynamics of sexually transmitted diseases [4,5], although this methodology has never been previously applied to HPV infection. Two sexual networks were designed, with an identical number of partnerships: network A (without externalities) and network B (with externalities). The chosen perspective was that of British heterosexual men (“male chooser”) observed over their lifetime. Assortative random selection was used to match 6 lifetime partners to each man. Men with highest number of partners (5+ during their life) choose 80% of women from the same group and 20% of women with a lower number of partners (2-4). To estimate the number of women most likely to be chosen as partners over lifetime, we assumed that men would most likely select sexual partners from their own age group, the two groups younger and the two groups older, limiting the age difference of each partnership to a maximum of 29 years.
The two sexual networks included 57 men and 308 women (all ages). Network A included heterosexual British men and women only, replicating the predominant mating pattern in cost-effective studies in the literature. Network B included sexual behaviours and preferences that increase the risk of contagion, such as paid sex, concurrent relationships, foreign partners and occasional homosexual relationships [6]. Individuals who were exclusively homosexuals were not included in either cohort, as the risk associated with these relationships is known. The competitive risk (double counting) was edged with the random assignment (without replacement) of high risk partnerships. The protective effect of HPV vaccination was simulated as the “breaking” of the edge if one of the two partners was immune. At equilibrium, the edges left intact indicated the number of men still at risk of HPV infection after vaccination. In both networks, 80% of girls were vaccinated, simulating the coverage of the selective vaccination school program in UK. The effect of herd immunity was simulated by further reducing the number of unvaccinated girls by 50%. The relevant network analytics measured were (1) the density (number of edges over total possible links) and (2) the relative risk of HPV infection.
Results
The density (number of men with residual risk of HPV infection after vaccination) was 34.5% higher in cohort B than in A (n = 24 and n = 18, respectively out of 308 possible partnerships). The average risk of HPV infection in network B was 2.1 times higher that in network A (0.27 and 0.13, respectively).
Conclusions
Sexual behaviours and preferences should be systematically included in cost-effectiveness analysis of HPV vaccination strategies, to avoid over-estimating the protection offered by selective vaccination and herd immunity. At equilibrium, vaccinating women only can leave a residual number of unprotected relationships 34.5% higher than previously estimated. In unprotected relationships, «risky behaviours» may double the relative risk of HPV infection. The social value of adding boys to HPV vaccination should be re-assessed, including sexual preferences and behaviours in cost-effectiveness analysis
References
[1] The Nobel Assembly at Karolinska Institutet. Press Release 2008-10-06; 2008.
[2] Brisson M, Van de Velde N, Boily MC. Economic evaluation of human papillomavirus vaccination in developed countries. Public Health Genom 2009;12(5–6):343–51.
[3] Favato G et al. (2017) Ecological validity of cost-effectiveness models of universal HPV vaccination: a systematic literature review. Vaccine 35: 20. 2622-2632 May.
[4] Wylie JL et al. (2005) Identification of Networks of Sexually Transmitted Infection: A Molecular, Geographic, and Social Network Analysis. The Journal of Infectious Diseases 191:899–906
[5] Bearman P S, Moody J, Stovel K. (2004) Chains of Affection: The Structure of Adolescent Romantic and Sexual Networks. AJS 110 (1): 44–91
[6] Mercer C H et al. (2013) Changes in sexual attitudes and lifestyles in Britain through the life course and over time: findings from the National Surveys of Sexual Attitudes and Lifestyles (Natsal). Lancet 382: 1781–94
The discovery of vaccines against Human Papilloma Virus (HPV) is an iconic example of pharmaceutical innovation [1]. Numerous cost-effectiveness studies agree in attributing a relevant social value to the selective vaccination of 12-year-old girls [2]. However, the incremental value of vaccination of male children remains controversial. This is due to the dominant value of selective vaccination, increased by the effect of acquired immunity. No evaluation has so far considered the impact of sexual behaviours and preferences that increase the risk of infection [3]. The objective of this study is to consider such behaviours as externalities and to evaluate their impact on the social value of the selective anti-HPV vaccination.
Methods
Social network analysis is the process of investigating social structures using networks and graph theory. It characterizes networked structures in terms of nodes and the links, edges, that connect them. Network analysis has been useful to simulate the dynamics of sexually transmitted diseases [4,5], although this methodology has never been previously applied to HPV infection. Two sexual networks were designed, with an identical number of partnerships: network A (without externalities) and network B (with externalities). The chosen perspective was that of British heterosexual men (“male chooser”) observed over their lifetime. Assortative random selection was used to match 6 lifetime partners to each man. Men with highest number of partners (5+ during their life) choose 80% of women from the same group and 20% of women with a lower number of partners (2-4). To estimate the number of women most likely to be chosen as partners over lifetime, we assumed that men would most likely select sexual partners from their own age group, the two groups younger and the two groups older, limiting the age difference of each partnership to a maximum of 29 years.
The two sexual networks included 57 men and 308 women (all ages). Network A included heterosexual British men and women only, replicating the predominant mating pattern in cost-effective studies in the literature. Network B included sexual behaviours and preferences that increase the risk of contagion, such as paid sex, concurrent relationships, foreign partners and occasional homosexual relationships [6]. Individuals who were exclusively homosexuals were not included in either cohort, as the risk associated with these relationships is known. The competitive risk (double counting) was edged with the random assignment (without replacement) of high risk partnerships. The protective effect of HPV vaccination was simulated as the “breaking” of the edge if one of the two partners was immune. At equilibrium, the edges left intact indicated the number of men still at risk of HPV infection after vaccination. In both networks, 80% of girls were vaccinated, simulating the coverage of the selective vaccination school program in UK. The effect of herd immunity was simulated by further reducing the number of unvaccinated girls by 50%. The relevant network analytics measured were (1) the density (number of edges over total possible links) and (2) the relative risk of HPV infection.
Results
The density (number of men with residual risk of HPV infection after vaccination) was 34.5% higher in cohort B than in A (n = 24 and n = 18, respectively out of 308 possible partnerships). The average risk of HPV infection in network B was 2.1 times higher that in network A (0.27 and 0.13, respectively).
Conclusions
Sexual behaviours and preferences should be systematically included in cost-effectiveness analysis of HPV vaccination strategies, to avoid over-estimating the protection offered by selective vaccination and herd immunity. At equilibrium, vaccinating women only can leave a residual number of unprotected relationships 34.5% higher than previously estimated. In unprotected relationships, «risky behaviours» may double the relative risk of HPV infection. The social value of adding boys to HPV vaccination should be re-assessed, including sexual preferences and behaviours in cost-effectiveness analysis
References
[1] The Nobel Assembly at Karolinska Institutet. Press Release 2008-10-06; 2008.
[2] Brisson M, Van de Velde N, Boily MC. Economic evaluation of human papillomavirus vaccination in developed countries. Public Health Genom 2009;12(5–6):343–51.
[3] Favato G et al. (2017) Ecological validity of cost-effectiveness models of universal HPV vaccination: a systematic literature review. Vaccine 35: 20. 2622-2632 May.
[4] Wylie JL et al. (2005) Identification of Networks of Sexually Transmitted Infection: A Molecular, Geographic, and Social Network Analysis. The Journal of Infectious Diseases 191:899–906
[5] Bearman P S, Moody J, Stovel K. (2004) Chains of Affection: The Structure of Adolescent Romantic and Sexual Networks. AJS 110 (1): 44–91
[6] Mercer C H et al. (2013) Changes in sexual attitudes and lifestyles in Britain through the life course and over time: findings from the National Surveys of Sexual Attitudes and Lifestyles (Natsal). Lancet 382: 1781–94
| Date made available | 2018 |
|---|---|
| Publisher | F1000 Research Limited |
Cite this
- DataSetCite