Cost-effectiveness of using a 20-valent pneumococcal conjugate vaccine to directly protect adults in England at elevated risk of pneumococcal disease

ABSTRACT

Objectives

Despite the current pneumococcal vaccination program in England for older adults and adults with underlying conditions, disease burden remains high. We evaluated cost-effectiveness of 20-valent pneumococcal conjugate vaccine (PCV20) compared to current pneumococcal recommendations for adults in England.

Methods

Lifetime outcomes/costs of invasive pneumococcal disease (IPD) and community-acquired pneumonia (CAP) among adults aged 65–99 years and adults aged 18–64 years with underlying conditions in England were projected using a deterministic cohort model. Vaccination with PCV20 was compared with 23-valent pneumococcal polysaccharide vaccine (PPV23) from the National Health Service perspective.

Results

PCV20 was cost saving compared with PPV23 in base case and most sensitivity analyses. In the base case, replacing PPV23 with PCV20 prevented 7,789 and 140,046 cases of IPD and hospitalized CAP, respectively, and 22,199 associated deaths, resulting in incremental gain of 91,375 quality-adjusted life-years (QALYs) and incremental savings of £160M. In probabilistic sensitivity analyses, PCV20 (vs. PPV23) was cost saving in 85% of simulations; incremental cost per QALY was below £30,000 in 99% of simulations.

Conclusions

PCV20 vaccination in adults aged 65–99 years and those aged 18–64 years with underlying comorbidities in England is expected to prevent more hospitalizations, save more lives, and yield lower overall costs than current recommendations for PPV23.

KEYWORDS:

• Streptococcus pneumoniae
• vaccination
• immunization
• cost-effectiveness
• pneumococcal conjugate vaccine

1. Introduction

Streptococcus pneumoniae (pneumococcus) causes significant morbidity and mortality in both children and adults in the United Kingdom (UK) [1]. Invasive pneumococcal disease (IPD) is the most serious manifestation causing severe illness with a high risk of mortality [2–4]. Community-acquired pneumonia (CAP), however, is far more common and associated with significant morbidity and mortality [5–8]. Between 2010 and 2017, there were more than 4,000 cases of IPD in England and Wales annually [3]. Estimates of the burden of hospitalized CAP vary considerably, ranging from 80,000 to half a million cases, and a recent study of UK adults found 37% of persons hospitalized with CAP between 2013 and 2018 had pneumococcal pneumonia [5,6,9].

To reduce the risk of pneumococcal infection in the UK, routine vaccination is currently recommended by the Joint Committee on Vaccination and Immunisation (JCVI) for infants, at-risk children and adults (i.e. aged ≥2 years), and all older adults (i.e. aged ≥65 years) [10]. Two vaccines are currently available for use in the UK: a 23-valent pneumococcal polysaccharide vaccine (PPV23) manufactured by Merck Sharpe & Dohme (MSD), for which evidence suggests protection against vaccine-type IPD (VT-IPD) only [11–26], and Pfizer’s 13-valent pneumococcal conjugate vaccine (PCV13), which has been shown to protect against VT-IPD and vaccine-type CAP (VT-CAP) [27].

PPV23 has been recommended for all persons aged ≥2 years with underlying conditions (e.g. chronic comorbidities and immunosuppressive conditions) since 1992 and for all older adults since 2003 [28]. Although PPV23 coverage is high (65% among adults aged ≥65 years), surveillance data suggest that its use has provided only low to moderate short-term direct protection against IPD and has not had a population-level impact on IPD burden given that serotypes unique to PPV23 continue to account for a large proportion of IPD among older adults [2,29]. Immunization of children aged <2 years with PCVs – including 7-valent PCV (PCV7) from 2006–2009 and PCV13 from 2010 to present – has prevented a significant number of cases of pneumococcal disease and associated deaths through the direct protection of young children and indirect protection of older children and adults [3,30]. Nonetheless, the burden of both IPD and CAP remains substantial among adults [6,9,30].

Two new, higher valent PCVs – PCV15 and PCV20 – have recently been approved by the Medicines and Healthcare products Regulatory Agency for use among adults in the UK, however, at the time of writing JCVI has not published revised pneumococcal vaccination guidelines since their approval. PCV15 (manufactured by MSD) includes the serotypes in PCV13 (1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, 23F) plus two additional serotypes (22F, 33F) and PCV20 (manufactured by Pfizer) includes the serotypes in PCV15 plus five additional serotypes (8, 10A, 11A, 12F, 15B). Although PCV13 is only recommended among a very limited subgroup of high-risk UK adults, it is anticipated that these new PCVs may reduce disease burden in adults versus PPV23 – due to their broader serotype coverage than existing conjugate vaccines and expected greater, longer-lasting protection compared to PPV23 [2,31–36] – and thus may be a cost-effective use of healthcare resources. However, analyses of cost-effectiveness of use of higher-valent PCVs among adults in England have not been previously published. Accordingly, we evaluated the cost-effectiveness of PCV20 alone as well as various other strategies against a single dose of PPV23 in a population of adults currently eligible for pneumococcal vaccination in England.

2. Methods

2.1. Model description

A deterministic model with a Markov-type process was used to depict the lifetime risk of clinical outcomes and economic costs of pneumococcal disease in a hypothetical closed population comprising all adults aged 65–99 years and adults aged 18–64 years with underlying conditions in England (i.e. those adults currently eligible for vaccination with PPV23) (Figure 1). The model population is stratified at model entry by age (i.e. in one-year increments) and risk profile (i.e. low, moderate, or high risk). Persons may transition to a higher risk group, but not to a lower risk group, during the modeling horizon. Persons in the model population may be assumed to receive any of the following at model entry or subsequently: PCV20, PCV20→PPV23 (i.e. PCV20 at model entry [i.e. year 1] followed by PPV23 at start of year 2), PCV15→PPV23 (i.e. PCV15 at model entry followed by PPV23 in year 2), PPV23, PPV23+PPV23 (i.e. PPV23 at model entry and again at start of year 6), or no vaccine.

Figure 1. Model schematic.

Expected clinical outcomes and economic costs are projected annually for the model population, based on age, risk profile, disease/fatality rates, vaccination status, and unit costs of vaccinations and medical care. IPD includes bacteremia and meningitis, and all-cause CAP is stratified by care setting. Vaccinated persons may be at lower risk of future IPD and all-cause CAP; the magnitude of vaccine-associated risk reduction depends on clinical presentation, vaccine type(s), proportion of disease that is vaccine-preventable, age, time since vaccination, and risk profile. Risk of death from IPD, all-cause CAP, and other causes depends upon age and risk profile. Expected medical treatment costs for IPD and all-cause CAP are generated based on event rates and unit costs in relation to care setting. Vaccination costs are tallied in the year of vaccine administration. Clinical outcomes and economic costs projected for each vaccination strategy include IPD and all-cause CAP cases and attributable deaths, life-years (LYs), quality-adjusted LYs (QALYs), and costs of vaccination and medical treatment.

2.2. Model estimation

A summary of methods employed to estimate key model parameters is set forth below; corresponding parameter values are summarized in Table 1. A detailed description of methods employed to estimate all model parameters is provided in Supplement A.

Table 1. Base case model input values.

				Age and Risk Profile
				18–49 Years			50–64 Years			65–74 Years				75–84 Years				85-99 Years
				Moderate	High		Moderate	High		Low	Moderate	High		Low	Moderate	High		Low	Moderate	High
No. of adults (in thousands)				1,254.6	449.5		1,784.1	1,315.7		3,392.3	1,220.8	963.3		1,970.8	1,005.5	818.2		626.0	456.0	379.1
Annual disease incidence (per 100,000)
	Bacteremia			18.7	42.0		17.6	48.9		4.8	18.8	62.3		8.1	21.9	94.7		15.0	40.5	175.4
	Meningitis			2.1	4.7		1.7	4.8		0.3	1.1	3.6		0.3	0.7	3.0		0.1	0.3	1.5
	All-cause hospitalized CAP			247.7	452.4		700.6	1,279.3		769.7	2,301.4	4,202.5		939.7	2,809.7	5,130.8		1,498.7	4,481.2	8,183.1
Annual mortality/case-fatality (per 100)
	General population			0.1	0.2		0.6	0.8		1.1	1.7	2.3		3.0	4.5	6.0		9.6	14.4	19.2
	Bacteremia			6.0	7.2		12.8	15.4		14.0	17.5	21.0		22.0	27.5	33.0		32.3	40.3	48.4
	Meningitis			6.0	7.2		12.8	15.4		14.0	17.5	21.0		22.0	27.5	33.0		32.3	40.3	48.4
	All-cause hospitalized CAP			3.4	4.1		6.7	8.0		7.9	9.9	11.8		10.3	12.9	15.5		14.5	18.1	21.7
General population health-state utilities				0.7745	0.7346		0.7234	0.6852		0.9283	0.7308	0.7082		0.8921	0.6799	0.6607		0.8191	0.6038	0.5643
Vaccine coverage
	PCV20*			43.4%	37.5%		43.4%	37.5%		42.9%	79.0%	83.7%		73.3%	89.6%	95.5%		77.9%	85.9%	85.7%
	PPV23**			43.4%	37.5%		43.4%	37.5%		42.9%	79.0%	83.7%		73.3%	89.6%	95.5%		77.9%	85.9%	85.7%
VE PCV vs. VT-IPD
	Year 1			81.5%	65.2%		79.2%	63.3%		75.0%	75.0%	60.0%		75.0%	75.0%	60.0%		75.0%	75.0%	60.0%
	Year 5			81.5%	65.2%		79.2%	63.3%		75.0%	75.0%	60.0%		75.0%	75.0%	60.0%		75.0%	75.0%	60.0%
	Year 10			63.1%	50.5%		61.2%	49.0%		58.0%	58.0%	46.4%		58.0%	58.0%	46.4%		58.0%	58.0%	46.4%
	Year 15			37.2%	29.8%		36.2%	28.9%		34.3%	34.3%	27.4%		34.3%	34.3%	27.4%		34.3%	34.3%	27.4%
	Year 16+			0.0%	0.0%		0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%
VE PCV vs. VT-CAP
	Year 1			55.6%	44.5%		51.3%	41.1%		45.0%	45.0%	36.0%		45.0%	45.0%	36.0%		45.0%	45.0%	36.0%
	Year 5			55.6%	44.5%		51.3%	41.1%		45.0%	45.0%	36.0%		45.0%	45.0%	36.0%		45.0%	45.0%	36.0%
	Year 10			43.0%	34.4%		39.7%	31.8%		34.8%	34.8%	27.9%		34.8%	34.8%	27.9%		34.8%	34.8%	27.9%
	Year 15			25.4%	20.3%		23.5%	18.8%		20.6%	20.6%	16.4%		20.6%	20.6%	16.4%		20.6%	20.6%	16.4%
	Year 16+			0.0%	0.0%		0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%
VE PPV23 vs. VT-IPD
	Year 1			32.7%	17.0%		32.4%	16.9%		55.9%	31.1%	16.2%		50.8%	28.2%	14.7%		37.6%	20.9%	10.9%
	Year 5			24.9%	13.0%		24.7%	12.8%		42.6%	23.7%	12.3%		38.7%	21.5%	11.2%		28.7%	15.9%	8.3%
	Year 10+			0.0%	0.0%		0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%		0.0%	0.0%	0.0%

*Vaccine uptake of PCVs was the same in sequential strategies involving PCV20→PPV23 and PCV15→PPV23; in such strategies, 100% of persons who received PCV and were alive in year 2 were assumed to receive PPV23

**Vaccine uptake was the same in the PPV23 strategy with revaccination; 100% of persons who received the initial dose of PPV23 were assumed to receive the second dose assuming they were alive in year 6

2.2.1. Population

The model population size (N = 15.6 M) and age distribution were based on the estimated population in England in 2022, the year when PCV20 became available for adults [37]. Persons in each age group were allocated into low-, moderate- (immunocompetent with underlying medical conditions), and high-risk (immunocompromised) subgroups based on risk factors defined in The Green Book of Immunisation against infectious disease, Chapter 25 [28] and published risk distributions [38] (see Supplement A for a comprehensive description of conditions considered).

2.2.2. Rates of IPD

Annual incidence of IPD was estimated by age (in one-year increments) and risk profile using the most recent age-specific disease rates available [2,3], age-specific distributions by risk profile, and age-specific odds ratios (ORs) for IPD by risk profile [38]. IPD rates were apportioned between bacteremia and meningitis based on a recent prospective study [4].

2.2.3. Rates of all-cause CAP

Annual incidence of all-cause hospitalized CAP was estimated by age and risk profile using age-specific rates of radiologically confirmed pneumonia from a recent prospective study conducted in Bristol, UK [5], age-specific distributions by risk profile [38], and age-specific fully adjusted ORs for all-cause hospitalized CAP by risk profile [39]. All-cause outpatient CAP was not considered in base case analyses.

2.2.4. Effectiveness of PCVs

For PCVs, vaccine effectiveness (VE) was based principally on randomized control trial data from the Community-Acquired Pneumonia Immunization Trial in Adults (CAPiTA) [27], which provides efficacy estimates for PCV13 and has been used in previous PCV20 cost-effectiveness analyses [40]. Initial effectiveness of PCVs against VT-IPD and VT-CAP for low- and moderate-risk persons aged ≥18 years was based on data from CAPiTA and post hoc analyses [27,41]. Initial VE against VT-IPD and VT-CAP for high-risk persons aged 18–99 years was assumed to be equal to 80% of corresponding values for low-/moderate-risk persons [42,43]. Initial VE-PCV was assumed to persist for 5 years, consistent with CAPiTA [27,44], and to wane thereafter, as follows: 5% annual decline during years 6–10, 10% annual decline during years 11–15, and no efficacy from year 16 through the end of the modeling horizon [41].

2.2.5. Effectiveness of PPV23

For VE-PPV23, we used the most recent real-world effectiveness studies conducted among UK adults, which are in line with assumptions for VE-PPV23 employed in previously published cost-effectiveness analyses. Risk-specific VE-PPV23 against VT-IPD for persons aged ≥18 years was derived from the study by Djennad et al. [2]. VE-PPV23 against VT-CAP was assumed to be zero based on findings by Lawrence et al. which do not provide robust evidence that PPV23 confers any meaningful protection against hospitalized CAP to older UK adults and is consistent with findings of a recent meta-analysis [26,45]. Beyond year 1 of the modeling horizon, VE-PPV23 against VT-IPD was assumed to decline [2] linearly, as follows: decline to 76.2% of initial VE by year 5, and decline to no efficacy by year 10.

2.2.6. Serotype coverage

Age-specific vaccine serotype coverage for IPD was based on prospectively collected surveillance data for 2016/2017 [2,3]. Vaccine serotype coverage for all-cause CAP was based on serotype coverage for pneumococcal CAP and the proportion of CAP due to pneumococcus in 2017/2018 [6]. Serotype coverage for pneumococcal CAP was assumed to be the same irrespective of age.

Herd effects for serotypes unique to PCV15 and PCV20 (vs. PCV13) were assumed to begin one year after the corresponding pediatric immunization program is expected to be implemented (i.e. model years 3 and 4, respectively). Serotype coverage for PPV23 was also reduced to account for the impact of herd effects on serotypes common to PPV23 and PCV20. Reductions in serotype coverage due to herd effects from childhood vaccination were based on the observed impact of PCV13 on the five additional serotypes in PCV13 not in PCV7 (excluding serotype 3) [3]. Reduction in serotype coverage was assumed to peak by year 10 and to be sustained at that level thereafter.

2.2.7. Costs

Costs of IPD (bacteremia, £4,542; meningitis, £5,600) and all-cause hospitalized CAP (£4,192) were derived based on average costs per finished consultant episode (FCE) from the 2019/2020 National Health Service (NHS) reference costs [46] and the average number of FCEs per admission from the Hospital Episode Statistics for England [9]. Prices of PPV23 (£16.80), PCV13 (£49.10), PCV15 (£50.30), and PCV20 (£56.50) were based on list values [47,48]. Vaccine administration cost (£10.06) was based on the General Practice Contract (2019) [49].

2.2.8. Utilities

Age- and risk-specific general population health-state utility values (HSUVs) were based on a published study of HSUVs for English adults [50]. For persons who experienced IPD or all-cause CAP, an annual utility decrement (0.13) was applied during the year in which the event occurred [51]. Both utilities and disutilities were estimated based on data sources which employed the EQ-5D instrument.

2.3. Analyses

2.3.1. Base Case

Clinical outcomes and economic costs were projected for the model population under PCV20 alone, PCV20→PPV23, PCV15→PPV23, and PPV23+PPV23 (hypothetical recommendations) and PPV23 (single dose, currently recommended for eligible NHS patients). Vaccine coverage for each strategy was assumed on par with current PPV23 coverage levels [29]. Future costs, LYs, and QALYs were discounted at 3.5% annually; analyses were conducted from the NHS (i.e. healthcare system) perspective. Cost-effectiveness was assessed considering the £20,000–30,000/QALY gained threshold range employed by the National Institute for Health and Care Excellence (NICE) and JCVI [52].

2.3.2. Subgroup and Sensitivity

Subgroup analyses were conducted comparing PCV20 versus PPV23 among all adults aged 65–99 years and adults aged 18–64 years with underlying conditions, respectively. Selected sensitivity analyses comparing PCV20 versus PPV23 were conducted in which alternative input values were employed for key model parameters. Probabilistic sensitivity analyses (PSA) were conducted (1,000 replications) for all strategies considered to account for uncertainty surrounding estimates of key model parameters. Inputs employed in sensitivity analyses and corresponding methods of estimation are described in Supplement A.

3. Results

3.1. Base case

Clinical and economic outcomes and full incremental cost-effectiveness results for all strategies evaluated are presented in Table 2. The cost-efficiency frontier – which comprises the PPV23 single dose strategy, PCV20 alone strategy, and PCV20→PPV23 strategy – indicates that the PCV15→PPV23 and PPV23+PPV23 strategies are dominated by strategies involving PCV20 (Figure 2). Although PCV20→PPV23 is on the cost-efficiency frontier, it was not cost-effective because its incremental cost-effectiveness ratio (ICER) compared with PCV20 alone (£686,948/QALY) was above the upper bound of the cost-effectiveness range. PCV20 alone was therefore deemed the most relevant hypothetical vaccination strategy.

Figure 2. Cost-effectiveness plane for alternative vaccination strategies among moderate- and high-risk adults aged 18–64 years and all adults aged 65–99 years in England (N = 15,635,909).

Table 2. Results of base case analyses of alternative vaccination strategies among moderate- and high-risk adults aged 18–64 years and all adults aged 65–99 years in England (N = 15,635,909).

A. Clinical and Economic Outcomes
		PPV23 (1 dose)	PCV20	PCV20→PPV23	PPV23+PPV23	PCV15→PPV23
Clinical Outcomes
	No. of cases
	IPD	117,807	110,018	109,709	117,031	113,644
	All-cause hospitalized CAP	7,483,128	7,343,082	7,343,116	7,483,223	7,384,442
	No. of deaths
	IPD	37,612	35,191	35,106	37,377	36,247
	All-cause hospitalized CAP	1,122,633	1,102,854	1,102,860	1,122,651	1,108,621
	Life-years (discounted)	179,787,906	179,906,103	179,906,580	179,788,893	179,867,258
	Quality-adjusted life-years (discounted)	132,546,963	132,638,338	132,638,681	132,547,656	132,608,379
Economic Outcomes (in millions)
	Medical care	£19,971.1	£19,432.8	£19,431.6	£19,968.7	£19,604.2
	Vaccination	£256.2	£634.9	£871.4	£424.0	£812.2
	Medical + vaccination	£20,227.3	£20,067.7	£20,303.0	£20,392.7	£20,416.4
B. Incremental Cost-Effectiveness
Strategy		Cost (in millions)	QALYs	Incremental Cost (in millions)	Incremental QALYs	ICER
PCV20		£20,067.7	132,638,338	–	–	–
PPV23 (1 dose)		£20,227.3	132,546,963	£159.6	−91,375	Dominated
PCV20→PPV23		£20,303.0	132,638,681	£235.3	343	£686,948
PPV23+PPV23		£20,392.7	132,547,656	£89.7	−91,024	Dominated
PCV15→PPV23		£20,416.4	132,608,379	£113.4	−30,302	Dominated

Compared with PPV23 (single dose), PCV20 alone would prevent an additional 7,789 IPD cases, 140,046 hospitalized CAP cases, and 22,199 pneumococcal-related deaths and save £538.31 million in medical care costs (Table 3). Total vaccination costs with PCV20 would increase by £378.70 million, and thus use of PCV20 would save the NHS £159.61 million. With total net costs lower by £159.61 million and discounted QALYs higher by 91,375, PCV20 was dominant versus PPV23; incremental net monetary benefit was £1.99 billion at £20,000/QALY gained and £2.90 billion at £30,000/QALY gained.

Table 3. Incremental results of base case, subgroup, and sensitivity analyses comparing PCV20 alone vs. PPV23 (single dose) among moderate- and high-risk adults aged 18–64 and all adults aged 65–99 years in England (N = 15,635,909).

			Subgroup			Sensitivity Analyses*
		Base Case	Moderate- and High-Risk Aged 18–64 Years	All Aged 65–99 Years		Hosp. CAP Rates from Pick et al.	2x Hosp. CAP Rates from Pick et al.	Plus Outpatient CAP	LT Consequences of IPD & CAP	VE-PPV23 vs. VT-CAP from Lawrence et al.	VE-PCV20 vs. ST3 CAP from Gessner et al.	Differential Discounting	Alt. Vaccine Waning	RWE of Vaccines
Clinical Outcomes
	No. of cases
	IPD	−7,789	−1,892	−5,897		−8,021	−7,927	−7,789	−7,043	−7,815	−7,651	−7,789	−6,617	−5,874
	All-cause hospitalized CAP	−140,046	−22,355	−117,691		−41,037	−80,087	−140,046	−119,221	−105,222	−165,231	−140,046	−118,323	−194,477
	All-cause outpatient CAP	–	–	–		–	–	−61,501	–	–	–	–	–	–
	No. of deaths
	IPD	−2,420	−287	−2,134		−2,516	−2,475	−2,420	−2,217	−2,431	−2,384	−2,420	−2,003	−1,810
	All-cause hospitalized CAP	−19,779	−1,936	−17,843		−6,209	−12,021	−19,779	−49,847	−15,201	−23,599	−19,779	−16,511	−27,754
	Life-years (discounted)	118,197	24,262	93,935		41,291	69,316	118,197	284,236	91,657	136,570	145,090	103,952	155,186
	Quality-adjusted life-years (discounted)	91,375	19,081	72,293		30,981	52,837	91,588	192,331	69,875	105,485	109,906	80,521	120,158
Economic Outcomes (in millions)
	Medical care	−£538.3	−£85.8	−£452.5		−£178.2	−£318.7	−£540.4	−£469.4	−£411.8	−£627.1	−£538.3	−£470.5	−£723.0
	Vaccination	£378.7	£78.7	£300.0		£378.7	£378.7	£378.7	£378.7	£378.7	£378.7	£378.7	£378.7	£378.7
	Medical + vaccination	−£159.6	−£7.2	−£152.4		£200.5	£60.0	−£161.7	−£90.7	−£33.1	−£248.4	−£159.6	−£91.8	−£344.3
Cost-Effectiveness
	Cost per LY	Dominant	Dominant	Dominant		£4,857	£865	Dominant	Dominant	Dominant	Dominant	Dominant	Dominant	Dominant
	Cost per QALY	Dominant	Dominant	Dominant		£6,473	£1,135	Dominant	Dominant	Dominant	Dominant	Dominant	Dominant	Dominant

LT: long-term; RWE: real world effectiveness; ST3: serotype 3

*Detailed information on parameter values employed in sensitivity analyses can be found in Supplement A

3.2. Subgroup and Sensitivity

PCV20 was found to be cost-effective – if not cost saving – versus PPV23 in all subgroup and sensitivity analyses (Table 3). Sensitivity analyses also suggest that PCV20 alone would remain cost saving (vs. PPV23) at prices up to 4.4 times (£73.23) the PPV23 list price, and would be cost-effective at prices up to 15.8 to 21.5 times (£264.82 and £360.61) the PPV23 list price at £20,000/QALY and £30,000/QALY, respectively. In PSA, PCV20 alone was cost saving versus PPV23 in 85% of simulations and cost-effective in 99% of simulations (WTP = £30,000/QALY) (Figure 3). The cost-effectiveness acceptability frontier indicates that PCV20 is always the optimal strategy (Figure 4).

Figure 3. Scatterplot for cost-effectiveness of PCV20 alone versus PPV23 (single dose) among moderate- and high-risk adults aged 18–64 years and all adults aged 65–99 years in England (N = 15,635,909).WTP: willingness-to-pay

Figure 4. Cost-effectiveness acceptability curve for alternative pneumococcal vaccination strategies among moderate- and high-risk adults aged 18–64 years and all adults aged 65–99 years in England (N = 15,635,909).CEAF: cost-effectiveness acceptability frontier*The CEAF illustrates the probability of greatest net monetary benefit

effective

4. Discussion

With recent approval of next generation PCVs with expanded serotype coverage, we undertook a new evaluation to examine the potential cost-effectiveness of pneumococcal vaccination strategies – including PCV20 alone and others – relative to PPV23 alone (single dose). Our findings suggest replacing PPV23 with PCV20 alone for vaccination of adults aged 65–99 years as well as adults aged 18 to 64 years with underlying conditions in England would save lives and reduce health care resources and costs to the NHS. Furthermore, PCV20 alone was the most efficient vaccination strategy among those explored, dominating the strategy involving revaccination with PPV23 (2 doses total), as well as PCV15 in sequence with PPV23. Our finding that PCV20 alone is a cost saving replacement for PPV23 reflects the latest pre-COVID serotype coverage for pneumococcal disease among adults [2,3,6] and the expected greater and longer-lasting protection provided by PCV20 compared to PPV23, which largely offsets the 7% difference in IPD serotype coverage between vaccines (PPV23 non-PCV20 serotypes: 2, 9N, 17F, and 20; PCV20 non-PPV23 serotypes: 6A/6C) [2,27,44,53,54].

To the best of our knowledge, this is the first study to report on cost-effectiveness of PCV20 use among adults in England, however, several analyses have been conducted focused on the United States (US). Findings presented by the US Advisory Committee on Immunization Practices (ACIP) pneumococcal work group suggest that use of PCV20 alone among US adults aged ≥65 years as well as those aged <65 years with risk factors would be cost saving compared to the 2019 US adult pneumococcal vaccines recommendation [55–57]. Based on consideration of all domains included in the evidence-to-recommendation framework, including cost-effectiveness analyses, the ACIP recently recommended the use of PCV20 alone among moderate- and high-risk adults aged 19–64 years and all adults aged ≥65 years [55].

Base case analyses described herein were conducted taking a conservative approach to several aspects of model parameterization. Although higher-valent PCVs have been neither licensed nor recommended for infants and children in England, herd effects from future pediatric use of PCV20 and PCV15 were considered. In addition, base case analyses did not account for the considerable burden of outpatient CAP [58]. Long-term clinical consequences and costs of pneumococcal disease also were not included though common following acute illness [7,59,60]. Finally, despite evidence suggesting that risk of death may remain elevated for several years following an acute hospitalization for IPD or CAP [7,8,60] – especially among the elderly and those with comorbidities (i.e. 65% of the English population) [8,37–39] – we considered only in-hospital pneumonia deaths.

Robust estimates of the incidence of hospitalized CAP in the UK are limited. In base case analyses, we employed recently published estimates from the first prospective study of CAP in the UK by Hyams et al. [5] which suggest that CAP incidence is three to five times higher than previous estimates based on data from Nottingham, which were employed in the 2016 JCVI analyses [5,6,61,62]. To evaluate the impact of CAP incidence rates on model outcomes, we conducted sensitivity analyses in which we conservatively assumed CAP incidence rates based on Pick et al. (Nottingham data) and twice the rates from Pick et al., respectively. The latter analysis was similar to one conducted by JCVI in 2016 to address concerns of likely substantial CAP under ascertainment [6]. Despite the low incidence rates, PCV20 was found to be cost-effective versus PPV23, with ICERs of approximately £6,500/QALY and £1,100/QALY, respectively. The difference in rates reported by Pick et al. and Hyams et al. is believed to be largely attributable to differences in study design. The Nottingham study was designed primarily to measure serotype distribution and trends for CAP [1,6,61]. The study by Hyams et al., however, was designed to estimate CAP incidence and also employed a novel, potentially more robust approach to estimate the population denominator used to calculate incidence rates [5]. Estimates by Hyams et al. are comparable to rates reported in other industrialized countries and are better aligned with the number of hospital admissions for pneumonia in the UK (based on ICD-10 code J18) [9,63].

Because the observed decline in serotype 3 disease among UK adults after the introduction of PCV13 among infants was lower than the decline in disease due to the other five serotypes that are in PCV13 but not in PCV7, and because incidence of IPD due to serotype 3 began increasing again in 2013/2014, there is some concern that PCV13 – and therefore also PCV20 – may be less effective in preventing serotype 3 disease [3]. However, as summarized in a recently published review article, evidence suggests that PCV13 is indeed effective against IPD and CAP due to serotype 3 [64–67]. We therefore conducted a scenario analysis using estimates of VE-PCV20 against serotype 3 CAP based on the study by Gessner et al. and found PCV20 alone to be dominant versus PPV23 [67].

Assumed effectiveness of PPV23 against CAP is an area of parameter uncertainty. Consistent with findings of multiple meta-analyses, assumptions employed in recently published economic evaluations, and the 2015 JCVI position, we assumed that PPV23 conferred no benefit against CAP in the base case [11–25,54,68,69]. However, given recent literature suggesting some limited effectiveness of PPV23 against VT-CAP, we conducted a scenario analysis based on estimates from the study by Lawrence et al. which employed real-world data from the UK [45,70]. Although the differences in cases and deaths due to hospitalized CAP and medical care costs were reduced, PCV20 remained dominant (vs. PPV23). Due to the high degree of uncertainty surrounding recent literature – both within and across studies – findings from analyses in which PPV23 is assumed to provide some protection against CAP should be considered with caution until higher-quality data in support of this assumption become available. We also note that in line with NICE and JCVI methodological guidance [52,71], analyses considering no vaccine (i.e. as a standalone strategy) vs. PPV23 were not conducted because PPV23 is the current standard of care in the UK.

To account for uncertainty in the future price of PCV20, we considered alternative prices and found that PCV20 alone remained cost saving (vs. PPV23) at prices up to 4.4 times the PPV23 list price and was cost-effective at prices up to 21.5 times the PPV23 list price (assuming WTP = £30,000/QALY). As PPV23 recently changed to being centrally procured [72], it may now be procured at a discounted price, in which case the price of PPV23 employed in our analyses may somewhat overestimate the current cost of adult pneumococcal vaccination to the NHS.

The true reduction in hospitalization costs with PCV20 (vs. PPV23) is underestimated in our base case analysis as the cost of IPD and CAP cases treated in intensive care – estimated at 5.2% of all hospitalized CAP cases in 2018/19 [73] – was not accounted for. Published literature suggests intensive care hospitalizations for pneumonia span 11 days on average at a cost of £1,500 per bed-day [74,75]. Given that patients with IPD generally experience more severe symptoms, it is likely that the proportion of patients admitted to intensive care for IPD is even greater, however, we have not found published estimates on IPD intensive care admissions. In addition, new research suggests NHS reference costs fail to account for the opportunity cost of hospitalization [76,77]. This may be especially relevant in the context of pneumonia, which tends to peak during winter when many hospitals are at capacity. Further research is also needed to estimate the impact of pneumonia on social care.

Societal costs of pneumococcal disease were also omitted from these analyses in line with JCVI guidance [52], thus excluding the benefits of vaccination in avoiding productivity loss to patient and carers. The adoption of a societal perspective in health technology assessments of vaccines has been discussed in the literature [78,79]. A recent cost-effectiveness analysis of COVID-19 vaccination by Public Health England and London School of Hygiene & Tropical Medicine researchers demonstrates the relevance of the broader impact of vaccinations [75].

The burden of pneumococcal disease in England has been greatly reduced in the past two years due to measures implemented to reduce transmission of COVID-19 (i.e. lockdowns, social distancing measures, mask-wearing) [80]. However, in light of recent reports of out-of-season respiratory syncytial virus outbreaks across the globe [81–85] as well as data showing PCV13-type IPD cases increased in 2021 (vs. prior years) in England and elsewhere [86–88], there is concern that major outbreaks of common respiratory disease may occur in England as pandemic restrictions are further relaxed [89]. It is therefore critical that delivery of adult vaccinations be prioritized, especially for older adults and others at elevated risk [90].

5. Conclusions

Results of our analyses suggest that the use of PCV20 in lieu of a single dose of PPV23 among all adults aged 65–99 years and adults aged 18–64 years with underlying conditions in England would be cost saving (i.e. more effective and less costly than the current program). Other strategies considered were either too costly or less efficient than PCV20. The prolonged duration of protection and greater effectiveness against vaccine-type IPD and CAP with use of PCV20 results in greater protection of vulnerable adults, and therefore has the potential to substantially reduce the clinical, resource, and financial burden of pneumococcal disease on the NHS.

Declaration of interest

The research described herein and manuscript development was supported by Pfizer Inc. D Mendes, R Sato, A Vyse, J Campling, G Ellsbury, and T Mugwagwa are employees and shareholders of Pfizer; no other relationships or activities that could appear to have influenced the submitted work. M Slack has received personal fees from GSK, Pfizer, Merck, AstraZeneca and Sanofi Pasteur as a speaker at international meetings and as a member of advisory boards (outside the scope of the submitted work). M Slack has also worked as a contractor for Pfizer and received remuneration from Pfizer. M Atwood, A Averin, and D Weycker are employees of Policy Analysis Inc. (PAI), which received financial support from Pfizer Inc. for this study (including manuscript preparation). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

Authorship was designated based on guidelines promulgated by the International Committee of Medical Journal Editors (2004). All persons who met criteria for authorship were included in the author list. The contribution of each of these persons to this study is as follows: (1) conception and design (all authors), acquisition of data (A Averin, D Mendes, A Vyse, T Mugwagwa), analysis or interpretation of data (all authors); and (2) preparation of manuscript (all authors), critical review of manuscript (all authors). All authors have read and approved the final version of the manuscript. The study sponsor, Pfizer Inc., reviewed the study research plan and study manuscript; data management, processing, and analyses were conducted by PAI. All final analytic decisions and the decision to submit for publication were made solely by study investigators.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14737167.2022.2134120

Additional information

Funding

This study was funded by Pfizer Inc.