Cost-effectiveness of 2-dose human papillomavirus vaccination for 12-year-old girls in Zhejiang Province: implications for China’s expanded program on immunization

ABSTRACT

Background

The high cost and insufficient supply of HPV vaccines have substantially slowed their implementation in lower-income countries. This study aimed to assess the incremental cost-effectiveness of two doses of human papillomavirus (HPV) vaccination (bivalent 16/18 vaccine; 2vHPV) compared to a no-vaccination scenario and a three-dose scenario in one province in China.

Methods

A static Markov model was used to model a lifetime cohort of 100,000 girls aged 12 years at the start of vaccination. A two-dose vaccination schedule was assumed to be non-inferior to a three-dose schedule in terms of vaccine efficacy, and both vaccination schemes were assumed to provide lifelong protection. Incremental costs, health effects and incremental cost-effectiveness ratios were used to measure the outcomes when comparing the different strategies.

Results

Compared to no vaccination, the incremental cost-effectiveness ratio (Chinese yuan per quality-adjusted life year) of the two-dose vaccination strategy is 12,472, and the 2-dose strategy is calculated to be cost saving relative to the 3-dose vaccination strategy.

Conclusions

Introducing the 2vHPV vaccine would be highly cost effective at a per-dose vaccine price of CNY 500, which has implications for cervical cancer control in China and other resource-limited countries.

KEYWORDS:

• Cervical cancer
• cost-effectiveness
• HPV vaccine
• bivalent
• China
• expanded program on immunization

Introduction

Persistent human papillomavirus (HPV) infection is a well-established cause of cervical cancer (CC) and other types of cancer, such as anus, vulva, vagina, penis, and head and neck cancer. CC is the fourth most common cancer among women globally, with almost 85% of cases occurring in developing countries.¹ China, the largest developing country, is facing an enormous disease burden, with 99,000 new CC cases in 2015, which accounted for 5.6% of all female cancer cases.² Two high-risk HPV types (type 16 and 18) were reported to be associated with 69% of CC in China on 10 December 2018 by the HPV information center.³

Screening is considered an effective method of early detection, while the most promising strategy to prevent CC is vaccination against HPV. The 16/18 bivalent HPV (2vHPV) vaccine has been proven to be able to protect against persistent HPV infection and CC in clinical trials^4,5 and has been evaluated for cost-effectiveness in many countries.^6,7

Some clinical trials have also suggested that a two-dose 2vHPV vaccination schedule is noninferior to a three-dose schedule.^8,9 The World Health Organization (WHO) has recommended that girls aged 9–14 years receive a 2-dose 2vHPV vaccination schedule,¹⁰ and countries such as the US,¹¹ Canada¹² and Switzerland¹³ have implemented two-dose vaccination schedules. The WHO has also recommended that a cost-effectiveness study be carried out before HPV vaccines are included in the expanded program on immunization (EPI) because of the high costs associated with the vaccine and its delivery.¹⁴

According to the WHO database, 85 countries had introduced the HPV vaccine into the EPI by February 2019.¹⁵ The current vaccination strategy for the HPV vaccine is voluntary and paid for out of pocket in China because of its high price and inadequate supply.¹⁶ This situation has substantially decelerated the pace of HPV vaccine implementation in China to eliminate CC, which was called for by the WHO in May 2018.¹⁷ Recently, a novel, Escherichia coli (E. coli)-produced 2vHPV 16/18 vaccine (Xiamen Innovax Biotech, Xiamen, China) was confirmed to be efficacious and safe in China¹⁸ and showed immunogenic noninferiority in girls aged 9–14 years who received 2 doses compared to women aged 18 to 26 years who received 3 doses.¹⁹ This candidate HPV vaccine has enabled vaccination in large populations in both China and other low-income countries at lower manufacturing costs. Furthermore, vaccination strategies vary among countries, which results in confusion regarding whether the 2vHPV vaccine should be integrated into the EPI.

Despite a variety of previous cost-effectiveness studies of HPV vaccines, the generalizability of the results is not straightforward. Therefore, a cost-effectiveness analysis based on local parameters is preferred. There are several methodological innovations in this study. First, the costs of HPV-related diseases (such as cervical intraepithelial neoplasia and CC) are derived from a nationwide survey rather than experts’ opinions or estimations. Second, the cost of vaccination is based on a survey of the actual cost of the EPI in Zhejiang Province. Third, this study analyzes the candidate E. coli-produced 2vHPV vaccine with an assumed price and threshold analyses, since the vaccine will not be available on the market until 2020. Finally, this study evaluated the cost-effectiveness of a 2-dose vaccination schedule in a Chinese (Zhejiang Province) setting. Zhejiang is a relatively developed province in eastern China and plays a leading role in the EPI work in China, such as taking the lead in the EPI for 2-dose measles, mumps and rubella vaccines and the 2-dose inactivated poliomyelitis vaccine plus bivalent oral poliomyelitis vaccine. Considering the availability and large-scale production of HPV vaccines along with the current policy debate, this cost-effectiveness analysis aims to compare the two-dose and three-dose schedules for the 2vHPV vaccine.

Based on these considerations, a Markov model was used to simulate a lifetime female cohort of 100,000 people and estimate the impact and cost-effectiveness of 2-dose versus 3-dose 2vHPV vaccine schedules in Zhejiang Province. This study will provide a scientific basis for the rational and efficient use of HPV vaccines.

Materials and methods

Comparison of the strategies

To determine the best vaccination strategy, we compared 3 strategies: (1) no vaccine, which is regarded as no intervention because HPV vaccines are in short supply^16,17 and the vaccination rate is relatively low; (2) 2 doses of the 2vHPV vaccine; and (3) 3 doses of the 2vHPV vaccine. We assumed that both the 2-dose and 3-dose HPV vaccine schedules provide lifelong protection.

Modeling

A static Markov model was used to simulate the natural history of high-risk HPV and divide the natural history into a series of health states. These states consisted of healthy, HPV infection, low-grade squamous intraepithelial lesion (LSIL), high-grade squamous intraepithelial lesion (HSIL), and CC. The progression and relationship among the states are shown in Figure 1. The transition probability between the states is shown in Table 1: most probabilities assumed in the model progressed annually and would be stable lifelong. The model was developed using TreeAge Pro 2016 software (TreeAge software Inc. Williamstown, United States of America) to evaluate the long-term effects and cost-effectiveness of the 2vHPV vaccine in the abovementioned 3 different strategies/scenarios. The model ran in 1-year cycle, and a total of 88 cycles were performed to model a cohort of 100,000 girls from 12 years old to 100 years old according to the defined states and transition parameters. Costs and effects were discounted at 3% annually.

Figure 1. Markov model structure.

Figure 2. Univariate sensitivity analysis of main parameters for 2-dose schedule versus no vaccination.

Table 1. Model parameters.

	Age (years)	Values	Plausible ranges	References/sources
Transition probability
HPV prevalence				20
	12–24	0.1728	–
	25–29	0.1363	–
	30–34	0.1429	–
	35–39	0.1818	–
	40–44	0.1880	–
	45–49	0.1962	–
	50–59	0.1648	–
HPV back to health		0.6000	−20% – +20%	21
HPV to LSIL		0.1600	−20% – +20%	22
LSIL back to health		0.5000	−20% – +20%	23
LSIL to HSIL		0.0260	−20% – +20%	24,25
HSIL back to health		0.3150	−20% – +20%	22
HSIL to cervical cancer		0.1200	−20% – +20%	26
Cancer death rate		0.4280	−20% – +20%	3 0.3%/0.7% = 0.428
Cancer cure rate		0.2100	−20% – +20%	= 1 – (1 – A)^ (1/5)
A: 5-year survival rate for cervical cancer		0.7079	–	27
Female population mortality rate (‰)	12–14	0.1700	–	China Statistical Yearbook 2017
	15–19	0.0200	–
	20–24	0.2400	–
	25–29	0.1400	–
	30–34	0.2700	–
	35–39	0.6000	–
	40–44	1.0600	–
	45–49	1.2100	–
	50–54	1.8300	–
	55–59	3.6600	–
	60–64	5.4100	–
	65–69	10.1200	–
	70–74	17.0400	–
	75–79	32.0500	–
	80–84	60.6800	–
	85–89	89.0200	–
	90–100	155.8400	–
Proportion of women with HSILs receiving treatment		0.9500	30% – 100%	Assumed28
Proportion of women with HSILs treated by LEEP (loop electrosurgical excision procedure)		0.9000	−20% – +20%	29
Proportion of women with HSILs treated by hysterectomy		0.1000	–	29
Proportion of women with HSILs cured by LEEP		0.9000	−20% – +20%	Assumed29
Proportion of women with HSILs cured by hysterectomy		1.0000	–	Assumed28
Proportion of women with cervical cancer receiving treatment		1.0000	–	Assumed28
Vaccination parameters
Age at vaccination	12		–	Assumed28
Coverage		0.7000	30% – 100%	Assumed28
Vaccine efficacy		0.7678	40% – 100%	= (B × C) + ((100 − B) × D)
B: proportion of type 16/18		0.6900	–
C: protection rate for type 16/18		0.9780	–
D: protection rate for non-16/18 types		0.3000	–
Utility parameters (quality-adjusted life years, QALYs)
Healthy		1.0000	–
Diagnosed LSIL		0.9872	76% – 100%	28
Diagnosed HSIL		0.9872	70% – 100%	28
Cancer treatment		0.7270	−50% – +50%	28
Cancer cured		0.9380	42% – 100%	28

Abbreviations: HPV, human papillomavirus; LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion.

Costs

Costs related to the diagnoses and treatment of high-risk HPV-related diseases were collected by a nationwide survey in 2015 and published in 2018.²⁹ The survey was conducted through the Chinese National Health Industry Research Project and collected costs from 14 country-level hospitals and 9 provincial/municipal hospitals by stratified sampling. We assumed states of HPV infection and LSIL to have no treatment costs. We analyzed the costs from the government’s perspective; therefore, the costs include direct medical costs and direct nonmedical costs. The costs of vaccination were assumed because the candidate 2vHPV vaccine is not currently available in the market, and other costs of vaccination were obtained from an actual cost survey of the EPI carried out in Zhejiang Province,³⁰ which included human labor costs, daily maintenance costs, vaccination information system costs, and other expenses. The details are shown in Table 2.

Table 2. Costs of vaccination and treatment.

	CNY
	Parameters input	Plausible ranges	References/sources
Vaccination (per dose)
Vaccine (per dose)	500	50–1000	Market price of Cervarix^®
Other costs (per dose)	19.55	–	30
Costs of diseases
HSIL treated by hysterectomy (per patient per year)	17205.42	−20% – +20%	24,29
HSIL treated by LEEP (per patient per year)	3875.24	−20% – +20%	24,29
Cervical cancer (per patient per year)	51194.30	−50% – +50%	29
Direct medical cost (per patient per year)	45422.36	–	29
Direct nonmedical cost (per patient per year)	5771.95	–	29

Abbreviations: CNY, Chinese yuan; HSIL, high-grade squamous intraepithelial lesion; LEEP, loop electrosurgical excision procedure.

Effectiveness measurement

We measured effectiveness in terms of the numbers of CC cases and deaths, life years and quality-adjusted life years (QALYs). QALYs can assess a patient’s life years by combining morbidity and mortality in a single index, which is calculated by estimating on a cardinal scale from 0 to 1 (0 represents the worst health, while 1 represents the best health), thereby measuring effectiveness quantitatively (morbidity) and qualitatively (mortality). However, there are no published health state utilities or values of the abovementioned disease states in China. The index used in the model was from a literature review and is summarized in Table 1.

Analysis

The cost-effectiveness analysis compares the incremental costs and consequences of the 2-dose strategy compared to the no vaccination and 3-dose strategies. The incremental cost-effectiveness ratio (ICER) was used to evaluate the results. It is calculated as the incremental cost divided by the QALYs gained per woman. A strategy is considered to be “very cost-effective” if its ICER is less than the per capita gross domestic product (GDP) and “cost-effective” if its ICER is less than three times the per capita GDP according to the WHO threshold. The per capita GDP was 92,100 CNY in Zhejiang in 2017.

To address the uncertainty of parameters in the analysis, a one-way sensitivity analysis was performed with several alternative assumptions: (1) vaccine efficacy (40% to 100%); (2) vaccine price (50 to 1000 CNY); (3) discount rates for both costs and effects (1% to 5%); and (4) the remaining cost parameters and probabilities, which varied within the ranges in Table 1.

Results

Health and economic impacts

In a cohort of 100,000 females (with 70% coverage), a total of 423 cancer cases, 116 cancer deaths and 3,627 QALYs would be saved by both two-dose and three-dose strategies compared to no vaccination. In addition, both the two-dose and three-dose vaccination schemes would avert 27,500,541 CNY in medical costs compared to no vaccination. With regard to the cost, if we assume that the price of this novel vaccine equals that of a currently available 2vHPV (Cervarix^®) at approximately 500 CNY per dose, then the two-dose vaccination strategy is expected to save 36,368,500 CNY in vaccination costs compared to the three-dose vaccination strategy, which shows that the 2-dose HPV vaccination scheme is cost saving compared to the three-dose vaccination scheme.

The incremental cost of the cohort is 45,236,459 CNY, and the incremental effects of that is the aforementioned 3,627 QALYs. Therefore, the two-dose vaccination versus no vaccination is clearly cost effective at an ICER of CNY 12,472 per QALY gained, and this figure is less than the per capita GDP.

Sensitivity analysis

In one-way sensitivity analysis of the two-dose vaccination strategy versus no vaccination (Figure 2), we found that the results were all less than the per capita GDP despite several variations. The most sensitive factor was vaccine cost, and the 2-dose strategy remained very cost effective even when the vaccine price was as high as 1000 CNY per dose. Similarly, even if the vaccine efficacy dropped to 30% due to problems with the system or the failure of cross-protection, the ICER would still be low enough for the vaccine to be very cost effective.

Discussion

To the best of our knowledge, this is the first HPV vaccine cost-effectiveness analysis to simulate 2-dose vaccination strategies in Zhejiang Province, China. The results showed that the ICER is less than the per capita GDP for the 2-dose vaccination strategy compared to no intervention in a cohort of 12-year-old girls, and the sensitivity analysis showed that this result was robust when we varied the main parameters.

Our finding that 2-dose vaccination with this novel 2vHPV is very cost effective and has implications for CC control in both China and other resource-limited countries. Despite the heterogeneity of vaccination strategies and HPV prevalence, numerous studies^31,32 have shown that introducing HPV vaccines is cost effective. Especially in some resource-limited countries such as South Africa, Malaysia and India, introducing 2vHPV vaccines without screening is cost effective because of the high rates of CC in those countries. Similarly, the 2-dose schedule for the 2vHPV vaccine is cost effective in these countries when compared to a 3-dose 2vHPV or 2-dose 4vHPV vaccine regimen.^33,34 Cost-effectiveness comparison studies of the 2-dose and 3-dose HPV vaccination strategies in the United Kingdom³⁵ and Canada³⁶ showed that the 2-dose vaccination strategy is the more cost effective of the two if protection lasts for more than 20 years or 30 years, respectively. However, the duration of the protection conferred by the 2-dose 2vHPV vaccine remains unknown.

Great importance should be attached to the influence of discounting in this study, which only focuses on vaccine efficacy and vaccine price. We used a discount rate of 3% for the base-case analysis, as most previous studies have done.^32,33 The costs of vaccine are immediate, while the benefits are delayed. Therefore, large discounting will make the vaccination strategy less attractive.

Importantly, vaccine price is the most influential factor according to the sensitivity analysis. Because the new E. coli-produced 2vHPV vaccine can be manufactured on a large scale at low cost, the availability of an inexpensive vaccine will be favorable for resource-limited countries. If the vaccine price is dramatically reduced, then vaccination may be a very affordable policy option compared to no intervention. The new 2vHPV vaccine will soon be available on the market, and only Chinese-produced vaccines are deemed to be a reliable supply for the EPI.³⁷ The price of a vaccine is usually expected to decrease by 50–80% from the market price if the vaccine is included in the EPI. Introducing the HPV vaccine into China’s EPI would be an effective strategy for CC elimination in both urban and rural areas.³⁸ We did not take screening into consideration because of the low screening coverage and disparities in health care access in Chinese settings. A review of modeling studies in developed countries³⁹ found that delaying the starting age of screening and extending the screening interval are important ways of increasing the cost effectiveness of screening by either cytology or HPV testing for vaccinated females.

The strength of our study was that we used the locally established costs of vaccination, epidemiological data and costs of treating HSIL and CC. Since EPI and non-EPI vaccines are regulated differently in China, it is best to use the actual EPI cost in Zhejiang³⁰ to estimate the vaccination costs other than the vaccine itself. The costs of treating HSIL and CC were underestimated by previous studies that relied on expert panels or hospital-based studies without nonmedical costs.^28,40 With more reliable data and a robust sensitivity analysis, the result is less uncertain and sheds light on current questions pertaining to introducing HPV vaccines into the EPI. Our study evaluated only 2vHPV vaccines because of the applicability of current policies and the availability of vaccine resources. In our study, we analyzed only the cost-effectiveness of 2vHPV vaccine for girls prior to HPV exposure. Therefore, the use of a static model to stimulate costs and effects in a female cohort in a simple way can yield satisfactory results.³⁹ However, if we consider catch-up vaccination and boys’ vaccination in the future, this model is less likely to contribute reliable results. We considered the cross-protection provided by the 2vHPV vaccine in terms of vaccine efficacy and performed a sensitivity analysis in case of problems with the system or failure of cross-protection.

There are several limitations in our study. First, unlike some studies that applied dynamic transmission models to evaluate CC prevention strategies based on herd immunity among boys,^41–44 we merely used a static Markov model for the economic evaluation, and we analyzed only the direct effects of this 2vHPV vaccine. Second, the model failed to consider other diseases caused by HPV, such as vagina, head and neck cancers, and thus may have underestimated the effects of this vaccine. Finally, because the model is based on a one-year cycle, some transition parameters are not available and were calculated roughly. For example, the case fatality rate is not used for chronic diseases such as CC. We calculated a 0.428 rate of cancer death per annum for cancer patients, which is the cumulative risk of CC incidence (0.7%) divided by the cumulative risk of CC mortality (0.3%) at 75 years old.³ This method underestimated the number of CC deaths to a certain degree because some people who live beyond 75 years old are not considered.

In conclusion, introducing 2-dose 2vHPV vaccines is very cost effective at a per-dose vaccine price of CNY 500. The 2-dose vaccination schedule is a strictly dominant strategy if the per-dose vaccine price is CNY 50. The novel two-dose HPV vaccination has implications for CC control in both China and other resource-limited countries.

Disclosure of potential conflicts of interest

All authors declare that they have no conflicts of interest.

Author contributions

Zhiping Chen and Ting Wu designed the study; Yan Luo and Xuewen Tang reviewed the literature; Yan Luo and Hanqing He developed the model and performed the data analysis; Shenyu Wang and Jun Zhang estimated the validity of the model; and Yan Luo wrote the manuscript. The corresponding authors are the guarantors for the data and have full access to all data. All authors approved the final version of the manuscript.

Ethical approval

The study protocol was approved by the ethics committee of the ZJCDC.

Acknowledgments

We gratefully acknowledge Professor Ya Fang and Mr. Mingliang Luo for their technical support. Editing was performed by a specialized company.

Additional information

Funding

This work was supported by the Medical Health Science and Technology Project of Zhejiang Province, China under Grants [number 2019KY052 and 2017KY295].