Abstract
Background:
In June 2019, the US Advisory Committee on Immunization Practices recommended shared clinical decision-making regarding potential HPV vaccination of mid-adults aged 27-45 years.
Objective:
To examine the incremental cost-effectiveness ratios (ICER) and number needed to vaccinate to prevent 1 HPV-related cancer (NNV) of expanding HPV vaccination to subgroups of mid-adults at increased risk of HPV-related diseases in the US.
Design:
Individual-based transmission-dynamic modeling of HPV transmission and associated diseases, using HPV-ADVISE.
Intervention and Target Population:
Expanding 9-valent HPV vaccination to all US men and women aged 27-45 years (mid-adults) and higher risk subgroups within this age range.
Outcome measures:
ICERs and NNVs.
Time horizon:
100 years.
Perspective:
Health care sector.
Results of Base-Case Analysis:
Expanding 9-valent HPV vaccination to all mid-adults, those with a higher number of lifetime partners and those who have just separated was projected to cost an additional $2,005,000, $763,000 and $1,164,000/QALY-gained, respectively. The NNVs to prevent 1 additional HPV-related cancer were 7,670, 3,190, and 5,150, respectively, in comparison to 223 when vaccinating persons aged 9-26 years (vs. no vaccination).
Results of Sensitivity Analysis:
The mid-adult strategy with the lowest ICER and NNV was vaccinating infrequently screened mid-adult women who have just separated and with higher number of lifetime sex partners (ICER=$86,000/QALY-gained; NNV=470).
Limitations:
Uncertainty about new sex partners and natural history of HPV among mid-adults.
Conclusion:
HPV vaccination of mid-adults is substantially less cost-effective and produces higher NNVs than for ages ≤26 years under all scenarios investigated. However, cost-effectiveness and NNV are projected to improve when vaccinating higher risk mid-adult subgroups, such as mid-adults with higher numbers of sex partners who have recently separated and women who are under screened.
Keywords: human papillomavirus, HPV, HPV vaccination, mid-adult vaccination, cost-effectiveness, number needed to vaccinate, cervical cancer, HPV-related cancer, transmission-dynamic modeling, infectious disease modeling
INTRODUCTION
In 2006, the United States Advisory Committee on Immunization Practices (ACIP) recommended routine human papillomavirus (HPV) vaccination for females aged 11-12 years (series can be started at age 9 years) with catch-up vaccination through age 26 years.(1) The recommendation was extended five years later to males aged 11-12 years (series can be started at age 9 years) and catch-up through age 21 years.(2) In June 2019, following the United States Food and Drug Administration’s expanded approval of the 9-valent HPV vaccine for use through age 45 years,(3) and after reviewing evidence related to HPV vaccination of adults, the ACIP recommended catch-up HPV vaccination for all females and males through age 26 years and shared clinical decision-making for HPV vaccination of women and men aged 27 through 45 years (“mid-adults”).(4) Evidence reviewed by the ACIP included mathematical models from 5 groups that suggested that, although vaccination is safe, the population level benefit of vaccinating mid-adults would be small.(5) However, at the individual level, mid-adults who are at higher risk of acquiring new infections (e.g., adults with multiple new partners, or who are separated after long-term relationships) or of developing HPV-related cancers (e.g., women who participate less in cervical cancer screening) could substantially benefit from HPV vaccination.(6) In this context, more information on the cost-effectiveness of vaccinating mid-adult subgroups at higher risk of HPV infections and the number needed to vaccinate (NNV) to prevent one additional HPV-related cancer can help inform shared clinical decision-making regarding potential HPV vaccination. The objective of the current study was to evaluate the incremental cost-effectiveness and NNV of extending HPV vaccination to subgroups of the mid-adult United States population that may be at higher risk of acquiring HPV.
METHODS
For model simulations we used the United States version of HPV-ADVISE (Agent-based Dynamic model for VaccInation and Screening Evaluation).(7, 8) HPV-ADVISE is a calibrated individual-based transmission-dynamic model that integrates six components: 1) demographic (open stable population 10 to 100 years of age), 2) sexual behavior & HPV transmission (18 HPV types modeled separately, including vaccine types 6/11/16/18/31/33/45/52/58), 3) type-specific natural history of disease (infection, anogenital warts, precancerous cervical lesions (3 grades), and cancer of the cervix (3 stages), and HPV-attributable cancers of the anus, oropharynx, penis, vagina & vulva), 4) vaccination, 5) screening & treatment, and 6) economic. The behavioral and biological parameter values were derived by calibrating (fitting) the model to highly stratified United States sexual behavior, HPV epidemiology, and screening data taken from the literature and population-based datasets.(7, 8) For model projections, we identified the 50 best-fitting parameter sets using data targets and least squares (see Technical Appendix). External predictive validation of the model has been carried out by comparing model projections to post-vaccination surveillance data.(9) Economic parameters (e.g. medical costs and quality-adjusted life years (QALY) weights) were derived from published United States-specific studies.(9) All costs were inflated to 2021 USD using the medical care component of the United States Consumer Price index.(10) In our base case, we assumed the price of the 9-valent vaccine was $262/dose (including administration fees) (Supplementary Table S1).
Given that HPV-ADVISE is an individual-based model, it allows for the specific modeling of vaccination scenarios within subgroups of the modeled population based on individual risk factors such as sexual behavior. The comparator scenario is based on the existing 9-valent HPV vaccination program of persons aged 9-26 years in the United States. In our base-case analysis, we examined the expansion of this program to include the vaccination of all mid-adults or selectively include 2 specific subgroups of this population that may be at higher risk of HPV infection. The subgroups vaccinated are 1) mid-adults with higher sexual activity and 2) mid-adults who have just separated. Within the HPV-ADVISE model, the individuals upon entry in the modeled population are attributed 1 of 4 levels of sexual activity based on the number of lifetime partners that they are expected to have (from low (level L0) to high (level L3) - see Technical Appendix). For the scenario involving vaccination of mid-adults with higher sexual activity, mid-adults from the higher levels L2 and L3 in the modeled population were vaccinated, corresponding to those who are expected to have over 10 lifetime partners. For the scenario of vaccination of mid-adults who have just separated, vaccination occurred immediately upon the termination of a stable partnership within the modeled population. This was possible because the HPV-ADVISE model tracks each individual’s history, so we can know exactly when they end a stable partnership (see “Partnership formation and separation process” in the Technical Appendix).
We modeled historical HPV vaccination uptake in the United States by age and year based on the same assumptions described in Laprise et al. 2019,(9) but included more recent HPV vaccination coverage data from the National Immunization Survey (NIS)–Teen from 2017 to 2020.(11) Briefly, we reproduced 2-dose and 3-dose coverage from data obtained from the NIS-Teen for 13 to 17-year-olds and assumed 18-year-olds have the same uptake rates as 17 year-olds. For vaccination coverage among young adults aged 19 through 26 years, we used the estimates from Chesson et al. 2018(12) (see Supplementary Figure S1 for projected vaccination coverage of females and males over time). We assumed 100% vaccination uptake among unvaccinated mid-adults of from 2021 onwards to maximize the incremental benefits of mid-adult vaccination and reduce variability. For persons aged 9 to 17 years and for adults, we made the simplifying assumption that maximal vaccine efficacy occurs at the 2nd dose and 3rd dose, respectively. Vaccine efficacy is assumed to be prophylactic and to be 95% against HPV vaccine-types with lifelong duration of protection.
We performed a cost-utility analysis from the health care payer perspective including only direct medical costs (see Supplementary Table S4 for the Impact inventory for the health care perspective). The main outcomes were costs per QALY gained (incremental cost-effectiveness ratios, “ICER”). We discounted future costs and benefits at 3% per annum and used a 100-year time horizon from the start of HPV vaccination. QALYs and medical costs were attributed to clinical outcomes over time. In addition, we calculated the additional number of persons needed to be vaccinated to prevent one additional HPV-related cancer (NNV). In sensitivity analysis, we utilized the individual-based feature of the HPV-ADVISE model to vary the vaccinated subgroups. These variations were based on individual risk factors integrated in HPV-ADVISE including age, gender, sexual activity level, and cervical cancer screening behavior level (comprising 5 screening levels based on frequency from annual to never screened) (see Technical Appendix). We also examined the impact of natural history assumptions (e.g., progression rates from infection to disease, higher incidence of vulvar, anal, and oropharyngeal cancers), and economic parameters (e.g., health care costs, disease burden (QALY weights & case fatality), vaccine price, discount rate, time horizon). Model projections are presented using the median, and 10th and 90th percentiles (“80% UI”) of results from the 50 best-fitting parameter sets identified during model calibration, to consider variability due to uncertainty in sexual behavior, HPV infection and progression of HPV-related diseases.
Role of the Funding Source
Coauthors from the Centers for Disease Control and Prevention (CDC) contributed to the design of the study, interpretation of the findings, and editing of the manuscript.
RESULTS
Under base-case assumptions, expanding an HPV vaccination program for persons aged 9-26 years to include all mid-adults is projected to cost $2,005,000 (80% UI: 1,227,000 to >10 million) per QALY gained (Table 1, Figure 1A). When limited to persons with higher sexual activity and who have just separated, mid-adult HPV vaccination is projected to cost $763,000 (80% UI: 447,000 to 2,946,000) and $1,164,000 (80% UI: 525,000 to 4,627,000) per QALY gained, respectively (Table 1, Figure 1B). The model also projected that 7,670 (5,550 to 12,120), 3,190 (80% UI: 2,230 to 4,750) and 5,150 (80% UI: 3,800 to 8,250) mid-adults would need to be vaccinated to prevent one additional HPV-related cancer (NNV) when vaccinating all mid-adults, and when limiting vaccination to mid-adults with higher sexual activity and who have just separated, respectively (Table 2, Supplementary Table S2). By comparison, 223 (80% UI: 210 to 232) individuals aged 9-26 years would need to be vaccinated to prevent one additional HPV-related cancer versus no vaccination (Supplementary Table S3).
Table 1.
Cost-effectiveness of vaccinating mid-adults in the context of an ongoing HPV vaccination program for persons aged 9-26 years (Base-case* analysis)
| Additional costs ($ million) |
QALYs gained (1,000 QALYs) |
ICER ($ per QALY gained) |
|
|---|---|---|---|
| Vaccination strategy | Median (80% UI) | Median (80% UI) | Median (80% UI) |
| Vaccinating persons aged 9-26 years (comparator) | - | - | - |
| Vaccinating all mid-adults | 250.6 (249.1 to 251.9) | 125.1 (19.4 to 204.5) | 2,005,000 (1,227,000 to >10M) |
| Vaccinating mid-adults with higher sexual activity† | 86.9 (80.8 to 92.8) | 98.5 (to 189.5) | 763,000 (447,000 to 2,946,000) |
| Vaccinating mid-adults who have just separated | 96.5 (94.0 to 127.1) | 88.7 ( to 193.8) | 1,164,000 (525,000 to 4,627,000) |
Model projections are presented as the median, and 10th and 90th percentiles (80% UI) of results from the 50 best-fitting parameter sets identified through calibration. Open stable population of 1 million persons.
Definitions: mid-adults=Women and men aged 27 through 45 years; QALY=Quality-Adjusted Life-Year; ICER=Incremental Cost-Effectiveness Ratio; UI=Uncertainty Interval; M=million.
Base case: Vaccinating persons aged 27-45 years; 9-valent cost per dose, $262 in persons aged 19-45 years; Discount rate, 3% per annum for future costs and benefits; time horizon, 100 years from start of HPV vaccination.
Higher sexual activity: persons who are expected to have ≥10 lifetime partners (Sexual activity levels L2 and L3 - see Technical Appendix).
Figure 1. Cost-effectiveness of vaccinating mid-adults in the context of an ongoing vaccination program for persons aged 9-26 years.
Results for A) base-case analysis, and sensitivity analysis for B) scenarios of vaccination of mid-adults with higher sexual activity (≥10 lifetime partners, Sexual activity levels L2 and L3 - see Technical Appendix), and C) scenarios of vaccination of mid-adults who have just separated. Box plots represent the median, and 10th, 25th, 75th and 90th percentiles of projections for the 50 best-fitting parameter sets. The dotted line represents the median for the comparator for the base case in sensitivity analyses.
Definitions: mid-adults=Women and men aged 27 through 45 years; QALY=Quality-adjusted life year.
*. Base case: Vaccinating persons aged 27-45 years; 9-valent cost per dose, $262 in persons aged 19 to 45 years; Discount rate, 3% per annum for future costs and benefits; time horizon, 100 years from start of HPV vaccination.
†. Infrequently screened women: Women screened less frequently than every 5 years, or never being screened (Screening behavior levels S3 and S4 in HPV-ADVISE - see Technical Appendix).
‡: We separated the 50 parameter sets into the parameter sets that have lower probability of natural immunity following clearance in females (≤40% vs >40%) and faster progression to CIN1/2/3 (e.g., average median time from infection to CIN1 of 10 vs 15 months, and average median time from infection to CIN3 of 32 vs 36 months) (22/50 parameter sets) and those that have a higher natural immunity and slower progression (28/50 parameter sets).
§. We assumed 50% increase in incidence of vulvar, anal and oropharyngeal cancers. This is illustrative and reflects the maximal scale of the observed increasing trends in HPV-related cancer incidence.(13, 14) For this scenario, we also assume that the incidence of vulvar, anal and oropharyngeal cancer in the absence of vaccination is 50% higher than in the base case.
‖. Higher sexual activity: persons who are expected to have ≥10 lifetime partners (Sexual activity levels L2 and L3 - see Technical Appendix).
¶. Moderate sexual activity: persons who are expected to have 2-9 lifetime partners (Sexual activity level L1 - see Technical Appendix).
Table 2.
Incremental number of mid-adults needed to vaccinate (NNV) to prevent one additional HPV-related cancer in the context of an ongoing HPV vaccination program for persons aged 9-26 years
| Incremental number needed to vaccinate (NNV) |
|||
|---|---|---|---|
| Any HPV-related cancers* |
Female HPV-related cancers* |
Male HPV-related cancers* |
|
| Median (80% UI) | Median (80% UI) | Median (80% UI) | |
| BASE-CASE† ANALYSIS | |||
| Vaccinating all mid-adults | 7,670 (5,550 to 12,120) | 13,820 (8,520 to 32,040) | 18,550 (15,390 to 20,600) |
| Vaccinating mid-adults with higher sexual activity§ | 3,190 (2,230 to 4,750) | 5,180 (3,380 to 10,490) | 8,060 (7,050 to 9,390) |
| Vaccinating mid-adults who have just separated | 5,150 (3,800 to 8,250) | 8,440 (5,150 to 16,410) | 14,810 (12,270 to 16,440) |
|
| |||
| SENSITIVITY ANALYSES | |||
|
| |||
| Vaccinating mid-adults with higher sexual activity § | |||
| Base case† (Vaccinating mid-adults with higher sexual activity§) | 3,190 (2,230 to 4,750) | 5,180 (3,380 to 10,490) | 8,060 (7,050 to 9,390) |
| Age | |||
| Vaccinating persons aged 27-35 years | 2,700 (1,530 to 5,720) | 4,460 (2,100 to 29,790) | 6,600 (5,820 to 7,800) |
| Gender | |||
| Vaccinating mid-adult women only | 3,300 (1,850 to 24,960) | 4,290 (2,240 to >1M) | 14,210 (8,110 to 28,910) |
| Vaccinating mid-adult men only | 2,350 (1,650 to 6,210) | 4,190 (2,460 to 21,910) | 5,440 (4,370 to 7,720) |
| Screening | |||
| Vaccinating infrequently screened women‡ only | 890 (320 to >1M)‖ | 1,030 (320 to >1M)‖ | 15,510 (4,580 to >1M)‖ |
| Natural history | |||
| Faster progression and lower immunity | 2,770 (1,500 to 3,950) | 4,260 (2,300 to 7,500) | 7,860 (4,060 to 9,400) |
| Assuming 50% increase in incidence of vulvar, anal and oropharyngeal cancers† | 2,530 (1,810 to 3,430) | 4,590 (3,130 to 8,890) | 5,480 (4,790 to 6,400) |
|
| |||
| Vaccinating mid-adults who have just separated | |||
| Base case† (Vaccinating mid-adults who have just separated) | 5,150 (3,800 to 8,250) | 8,440 (5,150 to 16,410) | 14,810 (12,270 to 16,440) |
| Age | |||
| Vaccinating persons aged 27-35 years | 4,730 (3,210 to 11,380) | 7,010 (4,480 to 59,540) | 13,050 (10,390 to 15,470) |
| Gender | |||
| Vaccinating women only | 4,670 (2,570 to >1M) | 6,740 (2,990 to >1m) | 23,530 (12,300 to 81,530) |
| Vaccinating men only | 4,220 (3,020 to 9,770) | 6,990 (4,350 to 133,540) | 10,390 (9,170 to 11,870) |
| Screening | |||
| Vaccinating infrequently screened women‡ only | 2,310 (510 to >1M)‖ | 2,820 (530 to >1M)‖ | 48,400 (6,650 to >1M)‖ |
| Natural history | |||
| Faster progression and lower immunity¶ | 5,130 (3,720 to 7,670) | 8,010 (5,340 to 14,860) | 14,630 (11,390 to 16,390) |
| Assuming 50% increase in incidence of vulvar, anal and oropharyngeal cancers† | 4,150 (3,200 to 6,140) | 7,300 (4,830 to 13,470) | 10,060 (8,360 to 11,180) |
| Level of sexual activity | |||
| Women and men with moderate sexual activity‡ | 14,190 (6,130 to >1M) | 22,700 (7,020 to >1M)‖ | 52,560 (30,500 to 108,980) |
| Women and men with higher sexual activity§ | 1,730 (1,040 to 3,000) | 2,750 (1,450 to 8,160) | 4,130 (3,200 to 7,240) |
| Gender and Level of sexual activity | |||
| Moderate sexual activity††, women only | 6,470 (2,850 to >1M)‖ | 7,290 (3,150 to >1M)‖ | 63,240 (20,640 to >1M)‖ |
| Higher sexual activity§, women only | 1,800 (750 to >1M) | 2,020 (880 to >1M) | 9,390 (4,020 to 46,120) |
| Screening and Level of sexual activity | |||
| Moderate sexual activity††, infrequently screened women‡ only | 1,390 (360 to >1M)‖ | 1,470 (400 to >1M)‖ | 29,770 (3,450 to >1M)‖ |
| Higher sexual activity§, infrequently screened women‡ only | 470 (120 to >1M)‖ | 640 (130 to >1M)‖ | 8,850 (1,080 to >1M)‖ |
Model projections are presented as the median, and 10th and 90th percentiles (80% UI) of results from the best-fitting parameter sets identified through calibration. When mid-adult vaccination is restricted to women or men, benefits in opposite gender are due entirely to indirect effects. See Supplementary Table S3 for number of persons vaccinated and cancers averted.
Definitions: mid-adults=Women and men aged 27 through 45 years; NNV=Additional number of persons needed to vaccinate to prevent 1 additional HPV-related cancer; UI=Uncertainty Interval; M=million.
HPV-related cancers include cancers of the cervix, vulva, vagina, anus, and oropharynx for females and cancers of the penis, anus, and oropharynx for males. “Any HPV-related cancers” include cancers for both females and males. HPV-related cancers are cancers at these sites caused by HPV, also sometimes referred to as HPV-attributable cancers.
Base case: Vaccinating persons aged 27-45 years; 9-valent cost per dose, $262 in persons aged 19-45 years; Time horizon, 100 years from start of HPV vaccination.
Infrequently screened women: Women screened less frequently than every 5 years or never being screened (Screening behavior levels S3 and S4 in HPV-ADVISE - see Technical Appendix).
Higher sexual activity: persons who are expected to have ≥10 lifetime partners (Sexual activity levels L2 and L3 - see Technical Appendix).
Less than 75% of the parameter sets projected measurable additional HPV-related cancer cases averted for these scenarios because of stochasticity and small effect size.
We separated the 50 parameter sets into the parameter sets that have lower probability of natural immunity following clearance in females (≤40% vs >40%) and faster progression to CIN1/2/3 (e.g., average median time from infection to CIN1 of 10 vs 15 months, and average median time from infection to CIN3 of 32 vs 36 months) (22/50 parameter sets) and those that have a higher natural immunity and slower progression (28/50 parameter sets).
The value of 50% is illustrative and reflects the maximal scale of the observed increasing trends in HPV-related cancer incidence.(13, 14) For this scenario, we also assume that the incidence of vulvar, anal and oropharyngeal cancer in the absence of vaccination is 50% higher than in the base case.
Moderate sexual activity: persons who are expected to have 2-9 lifetime partners (Sexual activity level L1 - see Technical Appendix).
Results were most sensitive to level of sexual activity and screening behavior (Table 2–3, Figure 1B and C, Supplementary Table S2, and Supplementary Figures S2B and C). Vaccinating infrequently screened (less frequently than every 5 years or never) mid-adult women with higher sexual activity and who have just separated produced the lowest ICER ($86,000 (80%UI: 3,000 to >10 million) per QALY gained, Table 3) and the lowest NNV (470 (80% UI: 120 to >1 million) mid-adults vaccinated to prevent one additional cancer, Table 2, Supplementary Table S2). Vaccinating only mid-adult men with higher sexual activity produced lower ICERs and NNVs than only mid-adult women. Results were similar when vaccinating only men or women who have just separated (Table 2 and 3, Figure 1B and C, Supplementary Table S2). When varying the economic parameters, cost-effectiveness results were most sensitive to assumptions about the vaccine cost per dose, discount rate, and time horizon. Lower discount rates produced lower mid-adult vaccination ICERs, while shorter time horizons produced higher ICERs (Table 3). This is because the additional vaccination costs are mostly incurred during the first years of mid-adult vaccination while most of its benefits occur later (Supplementary Figure S3).
Table 3.
Univariate sensitivity analysis: Cost-effectiveness of vaccinating mid-adults in the context of an ongoing vaccination program for persons aged 9-26 years
| Sensitivity analysis | ICER ($ per QALY gained) |
||
|---|---|---|---|
| Vaccinating mid-adults with higher sexual activity* | Vaccinating mid-adults who have just separated | ||
|
| |||
| Median (80% UI) | Median (80% UI) | ||
| Base case ‡ | |||
| Vaccinating all mid-adults | 763,000 (447,000 to 2,947,000) | 1,164,000 (525,000 to 4,627,000) | |
| Age | |||
| Vaccinating persons aged 27-35 years | 505,000 (292,000 to >10M) | 1,113,000 (366,000 to >10M)† | |
| Gender | |||
| Vaccinating women only | 751,000 (249,000 to >10M) | 970,000 (326,000 to >10M)† | |
| Vaccinating men only | 662,000 (316,000 to >10M) | 841,000 (432,000 to >10M) | |
| Screening | |||
| Vaccinating infrequently screened women§ only | 161,000 (29,000 to >10M)† | 151,000 (40,000 to >10M)† | |
| Natural history | |||
| Faster progression & lower immunity¶ | 632,000 (355,000 to 2,895,000) | 1,164,000 (707,000 to 3,566,000) | |
| Assuming 50% increase in incidence of vulvar, anal and oropharyngeal cancers** | 659,000 (404,000 to 1,877,000) | 982,000 (481,000 to 2,797,000) | |
| Level of sexual activity | |||
| Women and men with moderate sexual activity‖ | NA | 2,110,000 (535,000 to >10M)† | |
| Women and men with higher sexual activity† | NA | 338,000 (132,000 to >10M) | |
| Gender and Level of sexual activity | |||
| Moderate sexual activity, women only‖ | NA | 945,000 (231,000 to >10M)† | |
| Higher sexual activity, women only* | NA | 289,000 (81,000 to >10M) | |
| Screening and Level of sexual activity | |||
| Moderate sexual activity‖, infrequently screened women§ only | NA | 160,000 (24,000 to >10M)† | |
| Higher sexual activity*, infrequently screened women§ only | NA | 86,000 (3,000 to >10M)† | |
| Economic analysis parameters | |||
| Vaccination cost per dose (base=$261.81)††: | |||
| $168.98 | 484,000 (282,000 to 1,870,000) | 739,000 (334,000 to 2,958,000) | |
| $272.12 | 794,000 (466,000 to 3,065,000) | 1,211,000 (547,000 to 4,812,000) | |
| Maximum health care costs‡‡ | 752,000 (431,000 to 2,892,000) | 1,145,000 (520,000 to 4,585,000) | |
| Maximum disease burden‡‡ | 631,000 (362,000 to 2,024,000) | 924,000 (453,000 to 3,211,000) | |
| Discount rate: | |||
| 0% | 282,000 (186,000 to 563,000) | 494,000 (342,000 to 883,000) | |
| 5% | 1,363,000 (661,000 to >10M) | 1,791,000 (702,000 to >10M) | |
| 50-year time horizon | 1,256,000 (537,000 to >10M) | 1,744,000 (604,000 to >10M) | |
Model projections are presented as the median, and 10th and 90th percentiles (80% UI) of results from the best-fitting parameter sets identified through calibration.
Definitions: mid-adults=Women and men aged 27 through 45 years; ICER=Incremental Cost-Effectiveness Ratio; QALY=Quality-Adjusted Life-Year; UI=Uncertainty Interval; M=million; NA=Not applicable.
Higher sexual activity: persons who are expected to have ≥10 lifetime partners (Sexual activity levels L2 and L3 - see Technical Appendix).
Less than 75% of the parameter sets projected incremental QALY gains for these scenarios because of stochasticity and small effect size.
Base case: Vaccinating persons aged 27-45 years; 9-valent cost per dose, $262 in persons aged 19-45 years; Discount rate, 3% per annum for future costs and benefits; time horizon, 100 years from start of vaccination.
Infrequently screened women: Women screened less frequently than every 5 years or never being screened (Screening behavior levels S3 and S4 in HPV-ADVISE - see Technical Appendix).
Moderate sexual activity: persons who are expected to have 2-9 lifetime partners (Sexual activity level L1 - see Technical Appendix).
We separated the 50 parameter sets into the parameter sets that have lower probability of natural immunity following clearance in females (≤40% vs >40%) and faster progression to CIN1/2/3 (e.g., average median time from infection to CIN1 of 10 vs 15 months, and average median time from infection to CIN3 of 32 vs 36 months) (22/50 parameter sets) and those that have a higher natural immunity and slower progression (28/50 parameter sets).
The value of 50% is illustrative and reflects the maximal scale of the observed increasing trends in HPV-related cancer incidence.(13, 14) For this scenario, we also assume that the incidence of vulvar, anal and oropharyngeal cancer in the absence of vaccination is 50% higher than in the base case.
Vaccination costs are based on the CDC vaccine price list as of 1 August 2018 (www.cdc.gov/vaccines/programs/vfc/awardees/vaccine-management/price-list/index.html) and include the cost of vaccine administration.
Maximum estimates from the U.S. literature (see Supplement Table S1 for health care costs used and the HPV-ADVISE Technical appendix [available at www.marc-brisson.net/HPVadvise-US.pdf] for detailed QALY weights and case fatality.
DISCUSSION
In this analysis, we examined the cost-effectiveness of expanding HPV vaccination to subgroups of the mid-adult United States population that may be at higher risk of acquiring HPV. Under all scenarios investigated, HPV vaccination of mid-adults was much less cost-effective than vaccinating individuals aged ≤26 years. The estimated cost per QALY gained by vaccinating mid-adult was about $2 million for all mid-adults, $1.2 million for mid-adults who have just separated from a stable relationship, $0.7 million for mid-adults with higher sexual activity (adults who are expected to have ≥10 lifetime partners), and $86,000 for infrequently screened mid-adult women with higher sexual activity and who have just separated. In contrast, vaccinating individuals aged 9-26 years is projected to be cost-saving (cost per QALY < $0) compared to no vaccination.(9) Similarly, the number of individuals needed to be vaccinated to prevent one HPV-related cancer (NNV) was notably higher for mid-adults than for those aged ≤26 years.
The current analysis projected fewer benefits per dose and higher ICERs from vaccinating mid-adults than our previous analysis(9) because of greater herd effects of HPV vaccination through age 26 years that reduce the additional benefits achievable by mid-adult vaccination. Three factors contribute to the greater projected herd effects compared to our previous analysis. First, our new historical vaccination coverage estimates are based on data up to 2020 (Supplementary Figure S1), which indicates higher coverage than assumed in the previous analysis based on 2016 data.(9) Second, mid-adult vaccination starts in 2021 in the current analysis (vs 2018 in the previous publication), which provides 3 years of additional herd effects from the historical vaccination program. The majority of the QALYs gained by extending vaccination to age 45 years come from vaccinating those up to age 35 years. An additional 3 years of herd effects can have a sizeable effect on HPV incidence in the 27-35 years age group, which spans only 9 years. Third, in the current analysis, men aged 22-26 years are now included in the reference group to reflect the modified recommendation in 2019. Thus, the QALYs gained by vaccinating men aged 22-26 years are now included in the comparison strategy instead of the mid-adult strategy, and the additional herd effects from vaccinating men aged 22-26 years reduce the projected benefits of vaccinating mid-adults.
To our knowledge, our study is the first to examine the cost-effectiveness and NNV of vaccinating subgroups of mid-adults at higher risk of acquiring HPV infection. The study has four main strengths. First, our model (HPV-ADVISE) was calibrated to represent highly stratified United States-specific sexual behavior, HPV epidemiology, and health care resource use. Second, HPV-ADVISE projections were shown to be consistent with age-specific post-vaccination HPV infection and anogenital warts diagnosis prevalence data from the United States(9) and Australia.(15) Third, projections were made using 50 parameter sets to capture uncertainty in the natural history of HPV infection and related diseases, and variability in sexual behavior data. Had we not used 50 parameter sets to account for uncertainty in the natural history of HPV, our uncertainty intervals would not have been as wide and likely would not reflect the degree of uncertainty in the ICERs and NNVs of HPV vaccination of mid-adults. Finally, we performed extensive sensitivity analyses on key parameters and assumptions.
Our study also has limitations. First, though our model reproduces short-term observed post-vaccination herd effects, long-term herd effects of vaccinating younger age cohorts on mid-adult women and men remain uncertain. If HPV-ADVISE overestimates the herd effects of vaccination through age 26 years in the United States, the ICERs and NNVs of vaccinating mid-adult men and women might be more favorable than we estimated. Second, natural history of HPV infections in mid-adults is not well understood. HPV-ADVISE assumes that progression from HPV infection to disease is independent of age. However, if as has been suggested, a proportion of re-detection is due to reactivations of latent recurring infections(16) or deposition,(17) and new infections later in life have a smaller risk of progressing to cervical cancer,(18–20) mid-adult vaccination would produce lower benefits and higher ICERs and NNVs than projected. These benefits will also depend on the proportion of newly detected infections that are truly recent acquisitions and the potential for reactivated infections to produce HPV-related diseases. Third, historical vaccination coverage in the model could be underestimated, as we assumed that maximal vaccine efficacy occurs at course completion. Recent immunogenicity and efficacy studies suggest that one dose of HPV vaccine might offer similar protection to 2 doses or 3 doses.(21, 22) If so, this would result in a higher percentage of protection among persons aged 9 to 17 years than assumed, as observed coverage for 1 dose is higher than for 2 doses.(23) However, higher historical coverages would result in higher ICERs and NNVs due to larger herd effects in mid-adults by vaccinating younger age cohorts. On the other hand, if less than 3 doses were recommended for adults, vaccinating mid-adults with 1 or 2 doses instead of 3, as we assumed, would result in much lower ICERs (if a 1-dose or 2-dose schedule is non-inferior to a 3-dose schedule). However, NNVs would not change. Fourth, surveillance data show an increasing trend in vulvar, anal and oropharyngeal cancers.(13, 14) In the HPV-ADVISE model, cancer incidence is stable over time in the absence of vaccination. However, even when assuming a scenario of 50% increase in incidence of vulvar, anal, and oropharyngeal cancers in sensitivity analysis, mid-adult vaccination ICERs and NNVs remain relatively high (Table 3 and Figures 1B–C). Fifth, cervical cancer screening guidelines and practice have changed in the United States.(24) More effective cervical cancer screening could result in higher ICERs and NNVs for mid-adult vaccination. Sixth, our conclusions should not be generalized to low- to middle-income countries (LMICs) or other countries with different sexual behaviour, HIV prevalence, screening participation, and vaccination coverages as the US. Finally, issues related to the disruption of routine vaccination and cervical cancer screening due to the COVID-19 pandemic were not taken into account.
In conclusion, our results suggest that the cost-effectiveness and NNV of mid-adult vaccination improve when limited to mid-adult subgroups at higher risk of acquiring HPV infection. However, HPV vaccination of mid-adults remains much less cost-effective and produces much higher NNVs than vaccinating females and males aged 26 years or less under all scenarios investigated.
Supplementary Material
ACKNOWLEDGMENTS
MB designed the study, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. JFL drafted the article. JFL, LEM, HWC, and MD commented on the study design and model structure. MB and JFL performed the analysis. HWC and LEM provided United States specific data necessary for the analysis. All authors contributed to the interpretation of results, critically revised the manuscript for important intellectual content and approved the final version submitted for publication.
FUNDING
This work was supported by a contract from the Centers for Disease Control and Prevention (CDC; contract 00HCVGEB-2018-25176), a Fonds de recherche du Québec - Santé (FRQS) Research Scholars award (to MB), and a foundation scheme grant from the Canadian Institutes of Health Research (grant number FDN-143283). This research was enabled in part by support provided by WestGrid (www.westgrid.ca), Compute Ontario (www.computeontario.ca), Calcul Québec (www.calculquebec.ca) and the Digital Research Alliance of Canada (www.alliancecan.ca).
DISCLAIMER
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC).
Primary Funding Source:
Centers for Disease Control and Prevention
Footnotes
DISCLOSURES
JFL, HWC, LEM, MD, and MB have no conflicts of interest to declare.
REPRODUCIBLE RESEARCH STATEMENT
Study protocol: The article, Supplement, and technical appendix publicly available on Dr. Brisson’s Web site (www.marc-brisson.net/HPVadvise-US.pdf) are designed to provide sufficient information for an interested reader to replicate the analysis. Statistical code: Not applicable. Data set: Not applicable; this study did not use an analytic data set but used simulated data generated via our model, HPV-ADVISE. Model outputs are available upon reasonable request.
REFERENCES
- 1.Markowitz L, Dunne EF, Saraiya M, Lawson H, Chesson H, Unger E. Quadrivalent Human Papillomavirus Vaccine: Recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbidity and mortality weekly report. Recommendations and reports / Centers for Disease Control 2007;56(RR-2):1–24. [PubMed] [Google Scholar]
- 2.Centers for Disease Control and Prevention. Recommendations on the use of quadrivalent human papillomavirus vaccine in males-Advisory Committee on Immunization Practices (ACIP), 2011. MMWR Morb Mortal Wkly Rep. 2011;60(50):1705–8. [PubMed] [Google Scholar]
- 3.U.S. Food & Drug Administration (FDA). FDA approves expanded use of Gardasil 9 to include individuals 27 through 45 years old, 2018. FDA News Release. Accessed at https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm622715.htm on October 24, 2018. [Google Scholar]
- 4.Meites E, Szilagyi PG, Chesson HW, Unger ER, Romero JR, Markowitz LE. Human Papillomavirus Vaccination for Adults: Updated Recommendations of the Advisory Committee on Immunization Practices. MMWR Morb Mortal Wkly Rep. 2019;68(32):698–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chesson HW 2019;Pages https://stacks.cdc.gov/view/cdc/82631/cdc_82631_DS1.pdf on September 28 2021.
- 6.de Sanjose S, Brotons M, Pavon MA. The natural history of human papillomavirus infection. Best Pract Res Clin Obstet Gynaecol. 2018;47:2–13. [DOI] [PubMed] [Google Scholar]
- 7.Brisson M, Laprise JF, Chesson HW, Drolet M, Malagon T, Boily MC, et al. Health and Economic Impact of Switching from a 4-Valent to a 9-Valent HPV Vaccination Program in the United States. J Natl Cancer Inst. 2016;108(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brisson M, Laprise JF, Drolet M, Van de Velde N, Boily MC. Technical Appendix HPV-ADVISE CDC, 2014. Available at http://marc-brisson.net/HPVadvise-US.pdf. [Google Scholar]
- 9.Laprise JF, Chesson HW, Markowitz LE, Drolet M, Martin D, Benard E, et al. Effectiveness and Cost-Effectiveness of Human Papillomavirus Vaccination Through Age 45 Years in the United States. Ann Intern Med. 2020;172(1):22–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bureau of Labor Statistics, United States Department of Labor, 2021. Consumer Price Index. Item Medical Care. Accessed at https://data.bls.gov/timeseries/CUUR0000SAM?include_graphs=false&output_type=column&years_option=all_years on January 12, 2022. [Google Scholar]
- 11.Centers for Disease Control and Prevention (CDC);Pages https://www.cdc.gov/vaccines/imz-managers/coverage/nis/teen/index.html on August 2021.
- 12.Chesson HW, Meites E, Ekwueme DU, Saraiya M, Markowitz LE. Cost-effectiveness of nonavalent HPV vaccination among males aged 22 through 26years in the United States. Vaccine. 2018;36(29):4362–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Van Dyne EA, Henley SJ, Saraiya M, Thomas CC, Markowitz LE, Benard VB. Trends in Human Papillomavirus-Associated Cancers - United States, 1999-2015. MMWR Morb Mortal Wkly Rep. 2018;67(33):918–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Liao CI, Francoeur AA, Kapp DS, Caesar MAP, Huh WK, Chan JK. Trends in Human Papillomavirus-Associated Cancers, Demographic Characteristics, and Vaccinations in the US, 2001-2017. JAMA Netw Open. 2022;5(3):e222530. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Drolet M, Laprise JF, Brotherton JML, Donovan B, Fairley CK, Ali H, et al. The Impact of Human Papillomavirus Catch-Up Vaccination in Australia: Implications for Introduction of Multiple Age Cohort Vaccination and Postvaccination Data Interpretation. J Infect Dis. 2017;216(10):1205–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Malagon T, Trottier H, El-Zein M, Villa LL, Franco EL, Ludwig-McGill Cohort S. Human papillomavirus intermittence and risk factors associated with first detections and redetections in the Ludwig-McGill cohort study of adult women. J Infect Dis. 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Malagon T, Burchell AN, El-Zein M, Guenoun J, Tellier PP, Coutlee F, et al. Estimating HPV DNA Deposition Between Sexual Partners Using HPV Concordance, Y Chromosome DNA Detection, and Self-reported Sexual Behaviors. J Infect Dis. 2017;216(10):1210–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Plummer M, Peto J, Franceschi S, International Collaboration of Epidemiological Studies of Cervical C. Time since first sexual intercourse and the risk of cervical cancer. Int J Cancer. 2012;130(11):2638–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rodriguez AC, Schiffman M, Herrero R, Hildesheim A, Bratti C, Sherman ME, et al. Longitudinal study of human papillomavirus persistence and cervical intraepithelial neoplasia grade 2/3: critical role of duration of infection. J Natl Cancer Inst. 2010;102(5):315–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gage JC, Katki HA, Schiffman M, Fetterman B, Poitras NE, Lorey T, et al. Age-stratified 5-year risks of cervical precancer among women with enrollment and newly detected HPV infection. Int J Cancer. 2015;136(7):1665–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Barnabas RV, Brown ER, Onono MA, Bukusi EA, Njoroge B, Winer RL, et al. Efficacy of single-dose HPV vaccination among young African women. NEJM Evid. 2022;1(5):EVIDoa2100056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Basu P, Malvi SG, Joshi S, Bhatla N, Muwonge R, Lucas E, et al. Vaccine efficacy against persistent human papillomavirus (HPV) 16/18 infection at 10 years after one, two, and three doses of quadrivalent HPV vaccine in girls in India: a multicentre, prospective, cohort study. Lancet Oncol. 2021;22(11):1518–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pingali C, Yankey D, Elam-Evans LD, Markowitz LE, Valier MR, Fredua B, et al. National Vaccination Coverage Among Adolescents Aged 13-17 Years - National Immunization Survey-Teen, United States, 2021. MMWR Morb Mortal Wkly Rep. 2022;71(35):1101–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.United States Preventive Services Task Force (USPSTF), Curry SJ, Krist AH, Owens DK, Barry MJ, Caughey AB, et al. Screening for Cervical Cancer: US Preventive Services Task Force Recommendation Statement. JAMA. 2018;320(7):674–86. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.