A model-based evaluation of the cost-effectiveness of paediatric and elderly vaccination against pneumococcal infection in England

This study employs a novel low-dimensional model of pneumococcal serotype interactions to demonstrate that while the initial PCV7 introduction was not cost-effective due to rapid serotype replacement, switching to PCV13 and implementing PCV20 for both paediatric and elderly populations in England offers high cost-effectiveness and significant willingness to pay.

The Gist

Imagine the human body as a bustling city, and the bacteria Streptococcus pneumoniae (pneumococcus) as a group of rival gangs. These gangs aren't all the same; they wear different colored jackets, known as serotypes. Some gangs are mild nuisances, just hanging out in the nose and throat (carriage), while others are violent criminals that break into the bloodstream or brain, causing serious diseases like pneumonia and meningitis (Invasive Pneumococcal Disease, or IPD).

For years, the city's police force (the public health system) tried to stop these gangs using vaccines. But there's a catch: the gangs are clever. If you arrest the members wearing "Red Jackets" (Vaccine Type A), the "Blue Jacket" gang members might move in to take over the empty territory. This is called serotype replacement.

This paper is like a high-tech simulation game run by mathematicians and doctors in England. They built a digital twin of the country's population to answer a big question: "Is it worth spending more money to upgrade our police force to stop even more gangs?"

Here is the story of their findings, broken down simply:

1. The History of the "Gang Wars"

• The First Move (2006): The city introduced a vaccine (PCV7) that stopped 7 specific gangs. It worked great at first, especially for kids. The number of crimes dropped.
• The Backlash: But the other gangs (non-vaccine types) saw their chance. They moved in, filled the gaps, and started causing trouble. Because the new gangs were just as dangerous, the total crime rate didn't drop as much as hoped in the long run.
• The Upgrade (2010): The police upgraded to a better vaccine (PCV13) that covered 13 gangs. This helped again, but the "replacement" gangs kept evolving.
• The Current Situation: Today, the biggest problem isn't the old gangs anymore; it's the new, unvaccinated gangs. And the people getting hurt the most aren't the kids anymore—they are the elderly.

2. The New Proposal: The "Super-Vaccine" (PCV20)

The researchers looked at a new, powerful vaccine called PCV20, which covers 20 different gangs. They asked: Should we switch from the old 13-gang vaccine to this new 20-gang one for children? And should we switch the elderly from their current vaccine to this new one?

They ran their simulation forward for 50 years to see what would happen.

3. The Verdict: What's Worth the Price Tag?

The researchers calculated the "Willingness to Pay" (WTP). Think of this as the maximum price the city should pay for a vaccine dose to make sure it's a good deal. If the vaccine costs less than this number, it's a "buy." If it costs more, it's a "waste."

• For Children (The Switch to PCV20):

  • The Result: It's a good deal, but only if the price is right.
  • The Analogy: Imagine the 13-gang vaccine is a standard lock. The 20-gang vaccine is a high-tech smart lock. The study says the smart lock is worth the extra cost if the extra price is less than about £76 per dose (over a long 50-year period).
  • The Catch: If the vaccine is too expensive, the "replacement gangs" might still find a way to cause trouble, making the extra cost hard to justify.

• For the Elderly (Switching from PPV23 to PCV20):

  • The Result: It's a maybe, but less exciting than for kids.
  • The Analogy: The elderly are currently using an older, simpler vaccine (PPV23). Switching to the new 20-gang vaccine is beneficial, but the "price tag" for it to be worth it is much lower (around £11 to £32 extra per dose).
  • Why? The elderly population is huge and getting older, so stopping them from getting sick saves a lot of money. However, the new vaccine is already quite expensive, so the math is tight.

• The "Booster" Idea (A second shot at age 75):

  • The Result: This is actually a great idea.
  • The Analogy: Think of the vaccine protection like a fading memory. As people get older, they forget how to fight the gangs. Giving a second "reminder shot" at age 75 is highly cost-effective because it catches the people who are most vulnerable and most likely to get sick.

4. The "Hindsight" Surprise

The researchers also looked back at the 2006 decision to introduce the first vaccine (PCV7).

• The Surprise: If they had run this simulation back then, they would have said, "Don't do it! It's not worth the money!"
• Why? Because they couldn't predict that the "replacement gangs" would move in so quickly. The math shows that because the new gangs took over so fast, the first vaccine actually ended up costing more than the health benefits it provided in the long run. This is a rare case where looking back with better data changes the conclusion of the past.

5. The Bottom Line

The study concludes that England is facing a shifting battlefield.

1. Kids: Switching to the new 20-gang vaccine is likely a smart move, provided the price isn't too high.
2. Elderly: Switching them to the new vaccine is good, but the price needs to be very low to make sense.
3. Booster: Giving an extra shot to 75-year-olds is a very smart investment.

The Big Warning: The researchers admit their crystal ball isn't perfect. They don't have data on every single "gang" (serotype), and they don't know exactly how long the new vaccine will last in the real world. If a brand-new, super-dangerous gang appears that isn't in their model, all their predictions could be wrong.

In short: The police force needs to upgrade its weapons to stop the new gangs, but they need to shop around for a good price, and they definitely need to give the elderly a second reminder shot before they forget how to fight.

Technical Summary

1. Problem Statement

Pneumococcal infection (Streptococcus pneumoniae) causes significant morbidity and mortality, particularly in young children and the elderly, leading to invasive pneumococcal disease (IPD) and community-acquired pneumonia (CAP). While vaccination programs using Pneumococcal Conjugate Vaccines (PCVs) in children and Pneumococcal Polysaccharide Vaccines (PPVs) in the elderly have reduced specific serotype cases, they have triggered serotype replacement. Non-vaccine serotypes (NVTs) have emerged to fill the ecological niche left by vaccine-targeted strains, shifting the disease burden increasingly toward the elderly population.

The core problem addressed is the uncertainty regarding the long-term cost-effectiveness of transitioning from current vaccination strategies (PCV13 for children, PPV23 for the elderly) to higher-valency vaccines (PCV20). Policymakers lack robust, dynamic models that account for the complex interactions between 26+ serotypes, age-structured transmission, and the waning of immunity over time to determine the optimal "Willingness to Pay" (WTP) thresholds for these transitions.

2. Methodology

The authors developed a novel, low-dimensional deterministic transmission model combined with a health economic framework.

A. Epidemiological Model

• Structure: Unlike traditional compartmental models that require tracking every possible combination of serotype infections (which would result in $>5 \times 10^{15}$ equations), the authors adapted the Andreasen et al. framework. They model the number of hosts susceptible, vaccinated, infected (carriage), or resistant to each specific serotype independently, assuming independence between serotypes for susceptibility but allowing for interaction via competition (replacement) and cross-immunity.
• Scope: The model tracks 26 serotypes (grouped into PCV7, PCV13-additional, PCV20-additional, and "Others"), 9 age groups (ranging from 0–3 months to 75+ years), and 3 vaccine types (PCV7, PCV13, PCV20). This reduces the system to 3,042 equations.
• Data Fitting: The model was fitted to English data from 2000–2023, including:
  • IPD Data: Over 114,000 serotyped invasive cases from the UK Health Security Agency (UKHSA).
  • Carriage Data: Five carriage studies (2001/02, 2008, 2012, 2015, 2018) covering ~2,400 individuals.
  • Parameters Inferred: Serotype-specific transmission/recovery rates, age-dependent case-carriage ratios, vaccine efficacy against carriage (inferred for PCV13), and the impact of COVID-19 NPIs.
• Inference: A Bayesian, likelihood-informed approach was used to maximize the fit to historical data, with IPD data providing the primary constraint due to its volume compared to carriage data.

B. Health Economic Model

• Costs & QALYs: The model calculates direct healthcare costs (hospitalization, treatment) and Quality-Adjusted Life Year (QALY) losses associated with IPD, pCAP, and sequelae (e.g., meningitis, hearing loss).
• Discounting: Costs and health effects were discounted at 3.5% annually.
• WTP Calculation: The authors calculated the Willingness to Pay (WTP) threshold for vaccine doses. This represents the maximum price per dose (or price difference between vaccines) where the intervention remains cost-effective.
  • Two metrics were reported: a central estimate (valuing 1 QALY at £20,000) and a conservative estimate accounting for parameter uncertainty (valuing 1 QALY at £30,000, ensuring 90% of scenarios are cost-effective).
• Time Horizons: Simulations were run over 5, 10, 25, and 50 years to capture both short-term dynamics and long-term serotype replacement effects.

3. Key Contributions

1. High-Dimensional Serotype Modeling: The paper presents a computationally efficient method to model the dynamics of 26 interacting serotypes, overcoming the "curse of dimensionality" that typically forces models to aggregate serotypes into broad categories.
2. Retrospective Validation: The model successfully retrospectively analyzed the introduction of PCV7 (2006) and PCV13 (2010), revealing that the initial cost-effectiveness of PCV7 was likely overestimated due to unanticipated rapid serotype replacement.
3. Dynamic WTP Thresholds: The study provides specific WTP thresholds for transitioning to PCV20 in both paediatric and elderly populations, explicitly accounting for the interplay between paediatric herd immunity and elderly direct protection.
4. Booster Strategy Analysis: It evaluates the cost-effectiveness of adding a PCV20 booster dose at age 75, a strategy not previously modeled in this context.

4. Key Results

A. Retrospective Analysis (2006 & 2010)

• PCV7 (2006): The model suggests that introducing PCV7 in 2006 was not cost-effective at any price point when viewed over long time horizons. Rapid serotype replacement negated the initial benefits, leading to a negative WTP (e.g., -£17 for a 10-year horizon). This contrasts with earlier pre-introduction models that predicted cost-effectiveness at £10–15/dose.
• PCV13 (2010): Switching from PCV7 to PCV13 in 2010 was highly cost-effective. The WTP for the switch peaked at £60 per dose over a 50-year horizon, driven by the reduction in IPD cases caused by the additional 6 serotypes before replacement occurred.

B. Prospective Analysis (2026 Scenarios)

• Paediatric Switch (PCV13 $\to$ PCV20):
  • Over a 50-year horizon, the WTP for switching infants to PCV20 is £76–£84 (above the cost of PCV13), depending on the vaccine used in the elderly.
  • This switch is deemed cost-effective at current list prices (PCV20 list price is £56.80 vs. PCV13 at £49.10) for time horizons of 10 years or more.
• Elderly Switch (PPV23 $\to$ PCV20):
  • Switching the elderly (65+) from PPV23 to PCV20 is less cost-effective due to the high list price of PCV20 (£56.80) compared to PPV23 (£16.80).
  • The WTP threshold is £11–£32 (above PPV23 cost) over 50 years. At current list prices, this switch is not cost-effective.
• Combined Strategy: A simultaneous switch to PCV20 for both infants and the elderly yields a combined WTP of £63.
• Booster at Age 75: Introducing an additional PCV20 dose at age 75 (assuming 72% uptake) shows a WTP of £56 over 50 years. This is close to the list price, suggesting a booster strategy could be cost-effective if procurement costs are managed.

C. Sensitivity Analysis

• Time Horizon: WTP values are highly sensitive to the time horizon. Short horizons (5 years) show low WTP because the full benefits of herd immunity and serotype suppression take time to materialize.
• Vaccine Efficacy: The WTP is sensitive to the efficacy of PCV20 against carriage. If PCV20 efficacy against carriage drops significantly below PCV13 levels, the long-term cost-effectiveness diminishes due to faster serotype replacement.

5. Significance and Limitations

Significance:

• The study provides critical evidence for the UK Joint Committee on Vaccination and Immunisation (JCVI) regarding the transition to PCV20.
• It highlights that serotype replacement is a dominant factor that can render low-valency vaccines (like PCV7) cost-ineffective in the long run, justifying the move to higher-valency vaccines (PCV20).
• It identifies that while paediatric vaccination drives herd immunity, the elderly population now bears the majority of the disease burden, necessitating specific strategies like boosters or direct vaccination switches for this demographic.

Limitations:

• Data Scarcity: The model is limited by a lack of carriage data in older adults and limited longitudinal data on pCAP.
• Serotype Aggregation: The "Others" group aggregates many serotypes, potentially masking the rise of specific emerging strains.
• Efficacy Assumptions: PCV20 efficacy is inferred from PCV13 and immunological data rather than long-term real-world epidemiological data.
• Model Simplification: The assumption of multiplicative interactions between serotype and age (rather than a full matrix) simplifies the model but may miss specific age-serotype anomalies (e.g., Serotype 1 dominance in 15–44 year olds).

In conclusion, the paper argues that while the introduction of PCV7 was retrospectively not cost-effective due to replacement, the transition to PCV20 for children is highly cost-effective over the long term. However, switching the elderly to PCV20 or adding a booster at age 75 requires careful price negotiation to remain cost-effective.