Target population

The target population for routine vaccination with 4vHPV is school girls aged 11–12 years old. The target population for catch-up vaccination is girls and women aged 13–24 years old. The impact will be estimated for the entire population of Thailand

Scenarios

In the base case, the three scenarios being compared are: 1. Cervical cancer screening alone without vaccination (no vaccination), 2. Routine 4vHPV vaccination of girls aged 11–12 years old, along with the screening (routine 4vHPV vaccination), and 3. Routine 4vHPV vaccination of girls aged 11–12 years old plus catch-up vaccination of females 13–24 years old along with the screening (called “routine plus catch-up 4vHPV vaccination” in this article).

Demographic and sexual mixing model inputs

The model developed simulates aging, all-cause mortality over time and HPV transmission within the Thai population through sexual mixing. Data on population size, age and gender specific all-cause mortality rates were taken from the Health Information Group, while data on sexual mixing were adapted from Elbasha & Dasbach [12]. Details on the sexual behaviors of the population and the parameters used in the model calibration are provided in S1 Table.

Epidemiologic and clinical inputs

The age specific annual cervical cytology screening rates for Thailand were estimated based on Thai data from the 2009 Reproductive Health Survey [9], Center of Policy and Strategy of Ministry of Public Health (3rd National Survey on Thai Health 2004–2005) [13] and expert opinion (Table 1). In addition, based on the data from WHO/ICO Information Centre on HPV and Cervical Cancer 2010 [14], it was estimated that only 37.7% of Thai female population received adequate cervical cancer screening every year (defined as screening received once in 3 years).

Table 1

Screening and treatment parameters.

Routine cervical cytology screening, excluding those with hysterectomy (% per year)
Age group	Screening rate		Source
10–14 years	0.00		1) Center of Policy and Strategy, Ministry of Public Health; 3rd National Survey on Thai Health 2004–2005 [13] 2) The 2009 Reproductive Health Survey, National Statistical Office Ministry of Information and Communication Technology of Thailand [9] 3) Expert opinion
15–19 years	3.50
20–24 years	15.80
25–29 years	30.00
30–34 years	28.30
35–39 years	38.24
40–44 years	40.50
45–49 years	42.69
50–54 years	40.24
55–59 years	33.86
60–64 years	15.00
65–69 years	10.00
70–74 years	5.00
75–79 years	3.0
80–84 years	2.00
85 + years	0.00
Percent of female population that receives cervical cancer screening test at least once every 3 years
Women adequately screened (% per year)		37.7		WHO/ICO 2010 [14]
Percent of women with a follow-up screening test following an abnormal PAP result		68		Sirisamutr et al 2012 [19]
Recognize symptoms and seek treatment, % per year
Localized cervical cancer		2.6 (adjustable)		Elbasha & Dasbach 2010 [12]
Regional cervical cancer		10 (adjustable)		Elbasha & Dasbach 2010 [12]
Distant cervical cancer		90 (adjustable)		Elbasha & Dasbach 2010 [12]
Proportion of CIN/CIS is treated, by stage (%)
Percent of CIN 1 treated		40		Expert opinion
Percent of CIN 2 treated		75
Percent of CIN 3 treated		75
Percent of CIS treated		75
Proportion of genital warts cases treated (%)
Percent of genital warts treated		40		Expert opinion
Hysterectomy for non–HPV-related conditions rate (per year)
Age Groups		Rate		Source
15–24 years		0.02		Unpublished Data: Department of Obstetrics and Gynecology, Faculty of Medicine, Chulalongkorn University plus expert opinion
25–29 years		0.26
30–34 years		0.53
35–39 years		0.89
40–44 years		1.17
45–54 years		0.99
55+ years		0.36

* CIN, cervical intraepithelial neoplasia; CIS, carcinoma in situ

We utilized data based on expert opinion (Table 1) for the percentage of all diagnosed cases of CIN 1 including mild dysplasia and condyloma acuminata (anogenital warts), CIN 2 which includes moderate dysplasia, and CIN 3 including severe dysplasia and carcinoma in situ (CIS) that are treated in Thailand. Data on other clinical parameters, such as sensitivity and specificity of cervical cytology and colposcopy tests, CIN treatment cure rates, persistence of HPV infections after treatment were estimated based on data from the literature review done by Insinga et al. [15]. Annual Thai age specific hysterectomy rates were estimated based on unpublished data from Department of Obstetrics and Gynecology, Faculty of Medicine, Chulalongkorn University and expert opinion (Table 1). Consistent with the range of values examined in previous cost-effectiveness studies of HPV vaccines [16–18], in our reference case analysis, we assumed lifetime duration of vaccine protection against HPV types 6/11/16/18.

Quality of life measures

Quality-adjusted life years (QALYs) were estimated based on the following health utility measures: (a) Utilities for HPV disease states as reported for the U.S. model [11], and (b) Age and sex specific utility values for without HPV infection [20] (S1 Table).

HPV vaccination strategies and characteristics

We assumed that the 4vHPV vaccination strategy will be combined with current cervical cancer screening and HPV disease treatment practices in Thailand. We examined the health and economic impacts of three different scenarios: 1) no vaccination, 2) routine 4vHPV vaccination; and 3) routine plus catch-up 4vHPV vaccination. All scenarios were evaluated assuming that the current cervical cancer screening would remain at a constant rate throughout the time horizon of the evaluations. Based on the local experience with other vaccines in Thailand [21], we assumed that for the routine 4vHPV vaccination, 95% of all girls 11–12 years of age received the routine regimen. For the catch-up vaccination, it was assumed that 70% of females between 13–24 years were vaccinated in the first 5 years. Girls younger than 13 received 2 doses, while older females received three doses.

Economic data

We adopted a healthcare perspective to assess costs. The costs were broken down into 3 categories including costs of vaccination, cervical cancer screening and costs of treatment in Thailand. Cost data and corresponding references can be found in S1 Table. We assumed a cost of 500.00 THB per vaccine dose (≈ US$15) for each vaccine that was administered. The price in USD was estimated based on conversion rate of 1 USD /34.35 THB from http://finance.yahoo.com/currency-converter/#from=USD;to=THB;amt=1 accessed on April 3rd 2017. The cost per vaccine dose used was the cost proposed to pay in 2012 by the Ministry of Public Health of Thailand [22, 23]. Non-medical costs (e.g. transportation) and indirect costs (e.g. productivity loss) were not included in the analysis. Costs and health benefits (QALYs) were discounted by 3% [24].

Model outputs

We used a number of outcome measures to assess the epidemiologic impact and cost-effectiveness of each vaccination strategy. Epidemiologic outputs included CIN 2/3, CIN 1, invasive cervical cancer, genital warts cases for females and males, and cervical cancer related deaths. The economic outputs of interest from the model included total costs, quality-adjusted life years, and incremental cost per QALY. We measured the cost per QALY ratio as the incremental cost difference between the two strategies divided by the incremental QALY difference between the two strategies (i.e. Incremental Cost Effectiveness Ratio, ICER).

Model calibration

We assessed the predictive validity of the model by calibrating model predictions to the observed data of the incidence and mortality of cervical cancer in Thailand that was attributable to HPV 16/18. We also calibrated the model to the incidence rate of genital warts estimated from available data.

The total crude cervical cancer incidence rate in Thailand was reported as 29.2 per 100,000 [25] of which approximately 73.8% [14], i.e. ≈ 21.55 per 100,000, was assumed to be attributable to HPV 16/18. The corresponding crude cervical cancer mortality rate was reported as 12.7 per 100,000 [25] of which approximately 73.8% [14], i.e. 9.37 per 100,000, was assumed to be attributable to HPV 16/18. The model was calibrated using cervical cancer recognition and transmission probability parameters. At the beginning of the calibration process, all parameters were set to the values taken from Elbasha & Dasbach [12]. Then the cervical cancer recognition parameters were adjusted to obtain the observed cervical cancer incidence rate to cervical cancer mortality rate ratio (i.e. 21.55/9.37 ≈ 2.30). After that we adjusted the per sexual partnership transmission probability of HPV 16/18. Incidence rates for genital warts were 231 per 100,000 [8]. Approximately 90%, i.e. ≈ 208 per 100,000, were attributable to HPV 6/11.

Sensitivity analyses

Sensitivity analyses were conducted to examine how changes in the vaccination programs (routine vaccination and routine plus catch-up vaccination) or health-related benefit (health-related benefits, specifically HPV6/11-related genital lesions with and without HPV 6/11 related benefits) influenced the estimated cost-effectiveness ratio (Cost/QALY).

Horizon

Analysis was performed for 100-year time horizon, with 5 years interval to allow for the complete health and economic vaccination impact, in alignment with Elbasha et al. [11].

Cost effectiveness analysis

To assess the cost effectiveness of the vaccination strategy in preventing the disease, the model estimated the total discounted costs and effects (i.e., QALYs) accrued over the 100 year time horizon for both vaccine strategies that were being evaluated (i.e. no vaccination scenario, the routine vaccination of 11–12 year old girls, and routine vaccination of 11–12 year old girls plus the catch-up vaccination of 13–24 year old females). Next, the model calculated the incremental cost incurred to achieve the incremental benefit from vaccination. This was then used to calculate the ICER (the ratio of the incremental costs to incremental QALYs gained).

Public health impact of the HPV vaccination strategies

Impact on the HPV6/11/16/18-related diseases

The routine 4vHPV vaccination and the routine plus catch-up 4vHPV vaccination strategies would cumulatively prevent cervical cancer for 434,130 cases (62.1%) and 472,502 cases (67.6%) over 100 years respectively when compared to no vaccination. The model predicted that HPV16/18 related deaths over 100 years would be reduced by 182,234 cases (59.6%) and 199,068 cases (65.1%) for the routine 4vHPV vaccination and the routine plus catch-up 4vHPV vaccination strategies, respectively when compared to no vaccination.

The routine 4vHPV vaccination would also cumulatively prevent 1,984,634 cases of HPV16/18 related CIN1, 4,045,772 cases of HPV16/18 related CIN2/3, 533,677 cases of HPV6/11 related CIN1 and 12,708,349 cases of HPV6/11 related genital warts among women and men over 100 years, compared with no vaccination. Case reductions for the routine plus catch-up 4vHPV vaccination were higher than that for routine 4vHPV vaccination. The routine plus catch-up vaccination would also cumulatively prevent 2,133,772 cases of HPV16/18 related CIN1, 4,346,356 cases of HPV16/18 related CIN2/3, 574,097 cases of HPV6/11 related CIN1 and 13,641,398 cases of HPV6/11 related genital warts among women and men compared with no vaccination. Table 2 shows the cumulative case and percent reduction of these outcomes over 0, 5, 25, and 100 years. (S1 Fig provides the graphical presentations of the estimated number of the disease events in the routine vaccination and the routine plus catch-up vaccination in a population of 100,000 over 100 years)

Table 2

Estimated cumulative reductions of the HPV-related disease outcomes at Thai population level over 100 years.

Health related outcome	Routine 4vHPV vaccination vs no vaccination				Routine and catch-up 4vHPV vaccination vs no vaccination
	5 yrs	25 yrs	50 yrs	100 yrs	5 yrs	25 yrs	50 yrs	100 yrs
Cumulative case reductiona
HPV16/18 related deaths	0	1932	38212	182234	1	5273	5237	199068
Cervical cancer	0	8060	102342	434130	16	18638	135675	472502
HPV16/18 CIN1	106	143844	706943	1984634	4262	254545	853427	2133722
CIN2/3	198	299419	1449621	4045772	81183	521514	1734445	4346356
HPV6/11 Genital warts
In female	4837	867965	2679925	6372998	106799	1324470	3156476	6849994
In male	1490	725258	2536286	6335351	51514	1137032	2990177	6791404
HPV6/11 related CIN1	78	60061	214666	533677	3980	97495	255033	574097
Cumulative % reductionb
HPV16/18 related deaths	0	2.5	25.0	59.6	0.0	6.9	34.3	65.1
Cervical cancer	0	4.6	29.3	62.1	0.0	10.7	38.8	67.6
HPV16/18 CIN1	0.1	22.0	54.2	76.1	3.3	39.0	65.4	81.8
CIN2/3	0.1	22.5	54.1	76.0	3.1	39.2	65.1	81.6
HPV6/11 Genital warts
In female	1.3	47.0	72.5	86.3	28.9	71.7	85.5	92.7
In male	0.4	38.1	66.7	83.3	13.4	59.8	78.6	89.3
HPV6/11 related CIN1	0.2	37.6	67.3	83.6	12.5	61.1	79.9	90.0

a. Cases rounded to nearest 1.

b. Percentages round to nearest 0.1

Impact on healthcare cost

At population level, the routine vaccination and the routine plus catch-up vaccination strategies were projected to avoid the direct medical costs of the HPV-related diseases for 57,471,932,088 THB and 69,487,905,607 THB over 100 years when compared to no vaccination, respectively. Approximately three-fourths of the treatment costs avoided were from HPV16/18-related treatment costs (77.2% for routine vaccination and 77.0% for routine plus catch-up vaccination) and one-fourth were HPV6/11-related treatment costs (22.8% and 23.0%, correspondingly). (S2A and S2B Fig present the estimated healthcare costs avoided over 100 years by HPV genotypes at population levels of the routine vaccination and the routine plus catch-up vaccination when compared to no vaccination.) The estimated vaccination costs over 100 years were 17,108,393,827 THB and 24,943,313,705 THB for the routine vaccination and the routine plus catch-up vaccination respectively. (Table 3 and S2 Fig)

Table 3

Estimated costs over 100 years at population level.

Cost	Estimated cumulative costs over 100 years (THB)
	No vaccination (screening only)	Routine 4vHPV vaccination	Routine and catch-up 4vHPV vaccination
Direct medical costs of HPV-related diseases treatment
Cervical cancer	77,212,923,754	52,312,413,664	47,143,360,147
HPV16/18 CIN1	1,774,039,845	908,082,629	728,196,733
CIN2/3	38,293,570,197	19,706,623,671	15,922,378,541
HPV6/11 Genital warts
In male	10,283,135,581	4,020,431,254	2,692,431,841
In female	9,986,580,109	3,366,656,208	1,864,840,212
HPV6/11 related CIN1	396,580,852	160,732,517	107,751,083
Total direct medical costs of HPV-related diseases treatment	137,946,830,338	80,474,939,944	68,458,958,556
HPV-related disease costs avoided*		57,471,932,088	69,487,905,607
Vaccination cost	0	17,108,393,827	24,943,313,705

* Net when compared to no vaccination.

Cost-effectiveness of HPV vaccination strategy

Base case analysis

Under base-case analysis, the estimated incremental costs per person by adding the routine 4vHPV vaccination and the routine plus catch-up 4vHPV vaccination, compared with no vaccination, were 337.28 THB and 476.52 THB, respectively and the estimated QALY gained were 0.040 and 0.049, respectively. The estimated cost per QALY gained by adding the routine 4vHPV vaccination and the routine plus catch-up vaccination to the existing cervical cancer screening were 8,370 and 9,650 THB, respectively (Table 4).

Table 4

Estimated cost per QALY gained by adding routine 4vHPV vaccination or routine plus catch-up 4vHPV vaccination to cervical cancer screening in Thailand.

	Discounted Total		Incremental
	Costs/Person (THB)	QALYs/Person	Costs/Person (THB)	QALYs/Person	Costs/QALYs (THB)
Base case
Screening only (no vaccination)	30,183.6	28.53312	--	--	--
Routine 4vHPV vaccination	30,520.9	28.59342	337.28	0.04029	8,370
Routine plus catch-up 4vHPV vaccination	30,660.1	28.60250	476.52	0.04938	9,650
Sensitivity analysis
Base case with excluding HPV 6/11 health related disease
Screening only (no vaccination)	29,836.8	28.55760	--	--	--
Routine 4vHPV vaccination	30,404.1	28.59511	540.28	0.03751	14,420
Routine plus catch-up 4vHPV vaccination	30,587.9	28.60458	724.13	0.04599	15,747

Costs rounded to 0.01, QALYs rounded to 0.00001, and Costs/QALYs rounded to 1

Comparing to screening only strategy.

ICER = $$ \frac{CostsinthevaccinationStrategy-Costinscreeningonly}{QALYinthevaccinationStrategy-QALYinscreeningonly}$$

Limitations

The results of this analysis should be interpreted with the following limitations. Firstly, there is a limitation in the availability of Thai specific data in a key number of parameters, including data on the incidence of genital warts, CIN1, and CIN2/3 in the general population, therefore we could not calibrate the model to these outcomes. The model was calibrated specifically to cervical cancer incidences rates, mortality rates and genital warts and hence is conservative as no other cancers (such as vulvar, vaginal, anal, head and neck and penile cancers) were taken into consideration. Secondly, only direct medical costs were included. This potentially underestimates the benefits of vaccination. In this case, we used the default model parameters. In addition, the computation of ICER (Cost/QALY gained) is based on direct cost medical variables. Indirect medical cost and other non-medical economic variables are not used.