Public health impact of UK COVID-19 booster vaccination programs during Omicron predominance

ABSTRACT

Background

We aimed to estimate the public health impact of booster vaccination against COVID-19 in the UK during an Omicron-predominant period.

Research design and methods

A dynamic transmission model was developed to compare public health outcomes for actual and alternative UK booster vaccination programs. Input sources were publicly available data and targeted literature reviews. Base case analyses estimated outcomes from the UK’s Autumn–Winter 2021–2022 booster program during January–March 2022, an Omicron-predominant period. Scenario analyses projected outcomes from Spring and in Autumn 2022 booster programs over an extended time horizon from April 2022–April 2023, assuming continued Omicron predominance, and explored hypothetical program alternatives with modified eligibility criteria and/or increased uptake.

Results

Estimates predicted that the Autumn–Winter 2021–2022 booster program averted approximately 12.8 million cases, 1.1 million hospitalizations, and 290,000 deaths. Scenario analyses suggested that Spring and Autumn 2022 programs would avert approximately 6.2 million cases, 716,000 hospitalizations, and 125,000 deaths; alternatives extending eligibility or targeting risk groups would improve these benefits, and increasing uptake would further strengthen impact.

Conclusions

Boosters were estimated to provide substantial benefit to UK public health during Omicron predominance. Benefits of booster vaccination could be maximized by extending eligibility and increasing uptake.

KEYWORDS:

• COVID-19
• COVID-19 booster vaccination
• dynamic transmission model
• omicron variant (B.1.1.529)
• public health impact
• SARS-CoV-2

1. Introduction

The coronavirus disease 2019 (COVID-19) pandemic has had a substantial public health impact in the UK [1], with approximately 19.9 million confirmed cases of COVID-19 [2], 854,000 hospitalizations [3], and 166,000 deaths [4] as of 22 September 2022. Vaccination against SARS-CoV-2, the highly transmissible virus responsible for the COVID-19 pandemic [5], is central to the UK Government approach to living with and managing the virus [6].

Following a rapid and successful mass vaccination campaign that began in December 2020 [7], evidence emerged suggesting the gradual waning of vaccine effectiveness (VE) [8–10]. This waning protection prompted the UK to offer a booster dose of BNT162b2 (manufactured by Pfizer-BioNTech) or mRNA-1273 (manufactured by Moderna) to sustain protection against clinical illness [11]. In September 2021, the UK authorized a single booster dose for vulnerable populations to protect against severe consequences of COVID-19 and reduce the impact of COVID-19 on the National Health Service (NHS) during the Winter 2021–2022 respiratory season. These vulnerable populations consisted of individuals aged ≥50 years, those living in a care home, those aged 16–49 years in a clinical risk group, frontline healthcare workers, and individuals aged ≥16 years who were household contacts of immunosuppressed individuals [11]. By 22 December 2021, the UK Joint Committee on Vaccination and Immunisation (JCVI) had expanded booster eligibility to all individuals aged ≥16 years and those aged 12–15 years who were severely immunosuppressed, were in a clinical risk group, or were a household contact of an immunosuppressed person [12]. Over 32 million individuals in the UK (50.8% of the population) had received a booster dose as part of the Autumn–Winter 2021–2022 (Autumn–Winter_2021–2022) booster program by 20 March 2022 [13].

The SARS-CoV-2 Omicron variant (B.1.1.529) [14] emerged in late November 2021 as the UK Autumn–Winter_2021–2022 booster program was being expanded, and became predominant in the UK by mid-December 2021 [15]. Although the Omicron variant has been associated with lower rates of severe disease outcomes [16–18], its higher transmissibility than earlier variants [14,18] created a risk of case surges that could place a heavy burden on NHS resources and disrupt society due to high levels of community illness and lost productivity [19].

More recently, the UK and several other countries, including Israel [20,21] and the US [22], introduced programs to administer a second booster dose based on early evidence that the protection provided by boosters was lower and shorter-lived against Omicron than against earlier variants of concern [23]. Following JCVI advisement [24], the UK began a Spring 2022 booster program on 21 March 2022 [25] in which a second booster dose was offered to the most vulnerable populations, consisting of individuals aged ≥75 years, those living in a care home, and individuals aged ≥12 years who were severely immunosuppressed [24]. The JCVI has also advised an expanded booster program for Autumn 2022, in which booster doses should be offered to all individuals aged ≥50 years, individuals aged 5–49 years who are in a clinical risk group or are a household contact of people with immunosuppression, individuals aged 16–49 years who are caretakers, individuals residing or working in care homes, and frontline healthcare workers [26].

Little is known about the public health impact of the UK’s Autumn–Winter_2021–2022 booster program or the predicted impact of the Spring 2022 and Autumn 2022 (Spring&Autumn₂₀₂₂) booster programs. A real-world effectiveness study showed that the Autumn–Winter_2021–2022 booster program provided strong protection against symptomatic disease, hospitalization, and death [27], but this study was conducted during the initial months of the program, when the Delta variant (B.1.617.2) was predominant.

The Omicron variant differs from prior variants in its greater transmissibility [14,18] and lower severity [16–18]. Further, COVID-19 vaccines show reduced effectiveness and shorter durability against Omicron infections relative to infections with prior variants [28–30]. Thus, estimating the impact of booster programs during Omicron predominance would provide important information on the public health value of continuing revaccination using currently available COVID-19 vaccines. In addition, evaluating hypothetical program alternatives by altering the eligibility criteria or increasing uptake could provide insight on the relative value of boosting different population subgroups. This study examined the public health impact of booster programs in the UK during Omicron predominance and the potential impact of varying booster eligibility and uptake.

2. Methods

2.1. Overview

A compartmental dynamic transmission model was used to estimate the impact of booster vaccination by capturing direct (individual-level) and indirect (herd-level) effects of SARS-CoV-2 transmission in the entire UK population. Model compartments were separated by age group (< ‍6‍ months; 6 months – 4 years; 5–11 years; 12–15 years; 16–17 years; 18–29 years; 30–49 years; 50–64 years; 65–74 years; ≥75 years) to align with the UK vaccine delivery plan.

In the base case, the model was used to retrospectively calculate the impact of the actual UK Autumn–Winter_2021–2022 booster program over a three-month time horizon from 3 January 2022 until 3 April 2022. Scenario analyses projected the potential outcomes of actual or hypothetical alternative Spring&Autumn₂₀₂₂ booster programs over a one-year time horizon from 4 April 2022 to 2 April 2023.

The health outcomes estimated were averted COVID-19 cases (asymptomatic and symptomatic), general ward and intensive care unit (ICU) hospitalizations, long COVID cases, and deaths. Healthcare resource use saved was calculated from the NHS perspective and included general practitioner (GP) visits and hospital bed days. Greater societal impacts were estimated using a human capital approach [31] considering averted productivity loss for persons both in and outside of the workforce.

The model structure and inputs are described below, with further details available in the Supplementary methods. Ethical approval was not needed for this study.

2.2. Dynamic transmission model

The model had a Susceptible-Exposed-Infectious-Recovered [32] structure (Figure 1). The ‘susceptible’ compartment included those without vaccine- or infection-induced protection; this cohort was vulnerable to infection and severe disease. Separate compartments contained those who were currently protected following a single primary dose (V₁), two primary doses (V₂), or booster vaccination(s) (V_B). An additional compartment, ‘booster waning’ (V_BW), captured a cohort whose booster-induced protection against hospitalization and death was still effective, but whose protection against infection had waned. ‘Exposed’ and ‘infectious’ compartments were each split by vaccination status. ‘Exposed’ compartments were split by vaccination status and age, with the age-stratified CoMix matrix [33,34] determining the force of infection and rate of transitioning into the ‘exposed’ compartments. ‘Infectious’ compartments were split by vaccination status and divided into ‘infectious symptomatic and infectious asymptomatic’ compartments.

Figure 1. Dynamic transmission model structure.

Compartments – all compartments were further stratified by age: S – susceptible; V_{1,2,B, BW –} vaccinated with dose 1, dose 2, boosted, and boosted with protection against infection having waned; E, EV_1,2,B – exposed by vaccination status; I_{S, V1S, V2S, VBS} – infectious and symptomatic by vaccination status; I_{A, V1A, V2A, VBA} – infectious and asymptomatic by vaccination status; R – recovered; Parameters: υ_{1, 2, B} – coverage by vaccination status; λ – force of infection; σ_{1, 2, B} – vaccine effectiveness against infection by vaccination status; ψ_{0,1, 2, B, BW} – waning rate (1/duration of immunity) by infection-induced immunity and vaccination status; ε_{1, 2, B} – probability of displaying symptoms by vaccination status; γ – recovery rate (1/duration of symptoms); Κ – rate of progression to infectious disease

The ‘recovered’ compartment contained those who had recently experienced and recovered from COVID-19 infection. Individuals in the ‘recovered’ compartment were assumed to be fully protected against infection (i.e. have infection-induced immunity) for a period after infection.

The base case model was fit to UK epidemiology data from 4 January 2022 until 28 March 2022 based on UK Government COVID-19 surveillance reports [35]. At the start of the model horizon, individuals began in the ‘susceptible,’ ‘V₁,’ ‘V₂,’ ‘V_B,’ ‘infectious symptomatic,’ ‘infectious asymptomatic,’ or ‘recovered’ compartments. In scenario analyses, individuals could also begin in the ‘exposed’ or the ‘V_BW’ compartment. The starting model cohorts and the sources from which they are derived are described in Supplementary Table 1 (base case) and Supplementary Table 2 (scenario analyses).

Cohorts could transition from the ‘susceptible’ compartment to become ‘exposed’ or receive a booster vaccination (entering V_B). For simplicity, we assumed that individuals who have not yet received primary series at the start of the time horizon stayed unvaccinated and could not receive the booster dose. Cohorts waned from V₁ or V₂ to the ‘susceptible’ compartment based on the rate of waning vaccine-induced protection against infection. The waning rate was calculated as the inverse of the duration of vaccine-induced protection, which was assumed to be six months for either dose of primary series vaccine (Table 1). Cohorts waned from the V_B to the V_BW compartment based on the rate of waning vaccine-induced protection against infection, calculated as the weighted average of the rate of loss of protection against symptomatic disease reported for both BNT162b2 and mRNA-1273 boosters and based on 100% use of BNT162b2 primary series vaccine from the 31 March 2022 UK COVID-19 weekly vaccine surveillance report [36]. Cohorts then waned from V_BW to ‘susceptible’ based on the rate of waning of vaccine-induced protection against hospitalization. This rate of waning was calculated as the weighted average of the rate of loss of protection against hospitalization reported for both BNT162b2 and mRNA-1273 boosters and based on 100% use of BNT162b2 primary series vaccine as in the UK COVID-19 weekly vaccine surveillance report from 31 March 2022 [36]. Infection could lead to recovery, upon which infection-induced immunity would begin, or to death.

Table 1. Vaccine- and infection-induced protection inputs.

	PV dose 1	PV dose 2	Booster	Prior Infection
Effectiveness (%)
Infection	50.5 [41] ^a	74.5 [41] ^a	55.0 [43] ^b	100 ^c
Reduction in probability of symptoms ^d	5.9 [41]	19.1 [41]	12.0 [42]	-
Hospitalization	86.6 [40] ^e	97.0 [37] ^f	90.0 [43] ^b	-
Mortality	88.5 [40] ^e	97.0 [37] ^f	90.0 [43] ^b	-
Duration of protection (months)
Infection	6.00 ^g	6.00 ^g	5.43 ^h	2.04 ^i
Reduction in probability of symptoms ^d	6.00 ^g	6.00 ^g	9.72 ^h	-
Hospitalization	6.00 ^g	6.00 ^g	9.72 ^h	-
Mortality	6.00 ^g	6.00 ^g	9.72 ^h	-
Rate of waning (per week)
Infection	0.042 ^j	0.042 ^j	0.043 ^k	0.114 ^j
Reduction in probability of symptoms ^d	0.042 ^j	0.042 ^j	0.024 ^l	-
Hospitalization	0.042 ^j	0.042 ^j	0.024 ^l	-
Mortality	0.042 ^j	0.042 ^j	0.024 ^l	-

Abbreviations: DoP, duration of protection; PV, primary series vaccination; VE, vaccine effectiveness

^aBrand-agnostic VE calculated from Pouwels et al. (pre-print) [41] UK-specific data for vaccine-specific effectiveness estimates during a Delta-dominant period, and weighted by vaccine uptake

^bUK Vaccine Effectiveness Expert Panel consensus estimates of booster VE within 0–3 months based on the use of BNT162b2 or mRNA-1273 booster and assuming primary series vaccination with BNT162b2, as reported in UK COVID-19 weekly vaccine surveillance report from 24 March 2022 [43]

^cAssumed protection against re-infection after Omicron infection

^dBased on proportion of patients reporting any symptoms 1–20 days after PV dose 1 or ≥14 days after PV dose 2 compared to those not vaccinated. Assumed to be the same for all vaccines

^eBrand-agnostic VE for PV dose 1 calculated from Andrews et al. (pre-print) [40] UK-specific data for vaccine-specific effectiveness estimates for all age groups during a Delta-dominant period, at least 4 weeks after PV dose 1, and weighted by vaccine uptake of 50% BNT162b2 and 50% ChAdOx1

^fCalculated as the midpoint of UK Vaccine Effectiveness Expert Panel consensus estimates of VE against the Delta variant for PV dose 2, within 0–3 months of vaccination, as reported in the UK COVID-19 weekly vaccine surveillance report from 20 January 2022. Estimates are brand-agnostic and are characterized as ‘high-confidence’ estimates (Hospitalization: 95–99%; Mortality: 95–99%)

^gAssumed

^hDuration of protection calculated as 1/waning rate

ⁱThe DoP against re-infection after Omicron infection was calculated by multiplying the assumed DoP of infection-induced immunity after Delta infection (assumed to be 9 months) by the relative strength of protection against infection from Delta (84% based on [65]) versus Omicron (19% based on [66]). This DoP begins after the patient leaves the ‘infectious’ compartment

^jRate of waning protection calculated as 1/duration of protection

^kWaning rate calculated as the rate of loss of protection against symptomatic disease based on the use of BNT162b2 or mRNA-1273 boosters from the UK COVID-19 weekly vaccine surveillance report from 31 March 2022 [36]

^lWaning rate calculated as the rate of loss of protection against hospitalization based on the use of BNT162b2 or mRNA-1273 boosters, from the UK COVID-19 weekly vaccine surveillance report from 31 March 2022 [36]

The model was calibrated to weekly UK age-specific case rates and hospitalization rates reported by the UK Health Security Agency. Model calibration is described further in the Supplementary methods.

2.3. Base case model inputs

2.3.1. Vaccine inputs

Rates of booster uptake in the model were based on UK age-specific rates of booster uptake through 14 June 2022 (Supplementary Table 3) [38]. Coverage rates for high-risk individuals and their caregivers were based on UK data for vaccinations received by these groups from 8 December 2020 until 31 May 2022 [39]. As only a single booster dose was offered during the base case time horizon, the base case did not consider the possibility of multiple booster doses.

Effectiveness inputs for primary series [40,41] and booster [42–44] vaccination-induced protection against infection and hospitalization were derived from published UK estimates (Table 1). People were assumed to have received primary series vaccination during a period of Delta predominance and booster vaccination during a period of Omicron predominance. As such, VE inputs for primary series vaccination were derived from sources focusing on a Delta-predominant period and inputs for booster vaccination were derived from sources focusing on an Omicron-predominant period. Effectiveness values for primary series vaccinations were brand-agnostic. When relevant data were available, primary series VE was weighted according to vaccine-specific uptake. Primary series uptake was assumed to comprise 50% BNT162b2 vaccines and 50% ChAdOx1 vaccines. Booster effectiveness was based on expert consensus estimates from the UK weekly COVID-19 report from 24 March 2022 [43] that assumed the use of both BNT162b2 and mRNA-1273 booster vaccines and, for simplicity, 100% BNT162b2 primary series use. Primary series vaccination was assumed to protect against both infection and hospitalization for six months, and waning was assumed to occur at a constant rate based on the rate for waning protection against infection (Table 1). Boosters protected against infection for 5.43 months and against hospitalization for 9.72 months based on calculations from UK surveillance data from 31 March 2022 [36]. Waning occurred at a constant rate based on these durations of protection.

2.3.2. Exposure inputs

The ‘exposed’ cohorts were calculated from age-stratified and week-specific UK CoMix data [45,46] based on mixing within the model population, in conjunction with vaccination status. Mixing data for adults (aged ≥18 years) were stratified by age group (18–29 years, 30–39 years, 40–49 years, 50–‍59 years, 60–69 years, and ≥70 years), based on data through 28 December 2021 [46]. Mixing data for children (aged 0–17 years) were stratified by age group (0–4 years, 5–11 years, 12–‍17 years), based on data through 9 December 2021 [45].

2.3.3. Health inputs

Health inputs and sources are detailed in Table 2. Infections were based on UK infection epidemiology using UK Health Security Agency (UKHSA) Pillar 1 (swab testing in UKHSA labs and NHS hospitals) and Pillar 2 (swab testing through commercial partnerships, either processed in a lab or using rapid lateral flow tests) [47] case data from 4 January 2022 to 28 March 2022 [48]. The Omicron variant was dominant during this period, comprising between 93% [49,50] and >99.5% of infections [51]; as such, all infections were assumed to be of the Omicron variant.

Table 2. Health and NHS resource use inputs.

	Age group (years)
	<0.5	0.5–4	5–11	12–15	16–17	18–29	30–49	50–64	65–74	≥ 75	All ages	Source
Health
Probability of symptoms (%)	40	40	25	37	42	51	59	72	76	76	-	Derived from [56]
Proportion hospitalized ^a (%)	1.7	1.7	0.4	0.4	0.9	1.0	1.9	5.7	17.3	45.7	-	^b
Proportion of hospitalized in ICU (%)	0.00	0.00	0.00	0.00	0.00	10.20	26.36	30.41	20.49	6.15	-	^c
Infection fatality ratio ^a,d (%)	0.00	0.00	0.00	0.00	0.00	0.00	0.01	0.39	2.99	20.51	-	Calculated from [52]
High risk
Proportion high risk (%)	5.2	5.2	8.9	9.5	11.2	16.9	26.3	40.5	58.4	76.2	^−	^e
Risk ratio of hospitalization ^f	-	-	-	-	-	-	-	-	-	-	1.70	[53]
Risk ratio of ICU admission ^f	-	-	-	-	-	-	-	-	-	-	1.12	Cited in [54]
Risk ratio of mortality ^f	-	-	-	-	-	-	-	-	-	-	2.15	Cited HR in [54]
Proportion with long COVID (%)
Patients receiving outpatient care ^f	-	-	-	-	-	-	-	-	-	-	32	[67]
Hospitalized patients ^f	-	-	-	-	-	-	-	-	-	-	51	[67]
NHS resource use
Proportion visiting GP ^f,g (%)	-	-	-	-	-	-	-	-	-	-	100	Assumed
Bed days general ward ^f (days/case)	-	-	-	-	-	-	-	-	-	-	8.5	Calculated from [68]
Bed days ICU ^f (days/case)	-	-	-	-	-	-	-	-	-	-	10.06	[63]

Abbreviations: GP, general practitioner; HR, hazard ratio; ICNARC, Intensive Care National Audit & Resource Center; ICU, intensive care unit; NHS, National Health Service; UKHSA, UK Health Security Agency

^aOf symptomatic cases only

^bThe case to hospitalizations ratio was calculated using observed age-stratified UKHSA case and hospitalization data from weeks 45–53, 2020 [55], a period before vaccines against SARS-CoV-2 were widely available. Further details are provided in the Supplementary methods for model calibration

^cDerived from 2021 ICNARC report data on the overall probability of being admitted to an ICU and the age distribution of cases in ICU [63], considering the age distribution of hospitalized cases from UKHSA data

^dIFR inputs were adjusted from the IFRs reported by Nyberg et al. [52] to remove the effects of vaccination and high-risk mortality that were applied in the model simulation itself, to exclude asymptomatic cases, and to match the overall observed IFR in the model calibration period (3 January–3 April 2022). Further details are provided in the Supplementary methods for model inputs

^ePopulation-size weighted average of point prevalence for each age group based on data in Walker et al., 2021 [76]

^fAssumed to be the same across all age groups

^gOf non-hospitalized symptomatic cases only

The proportions of infections that were ‘infectious symptomatic’ (i.e. symptomatic infection) and ‘infectious asymptomatic’ (i.e. asymptomatic infection) were based on UK infection data stratified by vaccination status [41] and age based on age-specific clinical fraction estimates from a multi-national modeling study [56]. Symptomatic cohorts were considered 3.85 times more infectious than asymptomatic cohorts, based on data from a real-world study of community transmission conducted in Singapore [57]. For both asymptomatic and symptomatic cases, a 5-day latent (‘exposed’) period followed by a 5-day ‘infectious’ period were assumed based on average values from published COVID-19 models [58–62]. Asymptomatic cases were assumed to never lead to GP visits, hospitalization, long COVID, or death.

In cases in which hospitalization occurred, patients were assumed to be hospitalized after a 1-week delay from becoming ‘infectious.’ The age-stratified probabilities of ICU admission for hospitalized patients were derived from 2021 Intensive Care National Audit & Research Centre (ICNARC) data on patients admitted to the ICU with confirmed COVID-19 [63]. A proportion of the population within each age group was considered ‘high risk’ according to the clinical risk groups defined in the UK Green Book on Immunisation Against Infectious Disease, chapter 14a on COVID-19 [64], and were assigned a higher probability of hospitalization, ICU admission, and death (Table 2).

Infected patients gained infection-induced protection lasting 2.04 months after they leave the ‘infectious’ compartment. This duration of protection was calculated by multiplying the duration of infection-induced immunity after Delta infection (assumed to be 9 months) by the relative strength of infection-induced protection against Delta (84% [65]) compared with Omicron (19% [66]) (Table 1).

The proportion of patients with COVID-19 developing long COVID was assumed to be 51% for hospitalized patients and 32% for non-hospitalized symptomatic patients, based on data from the Belgian Health Care Knowledge Center for the prevalence of long COVID reported 1–3 months after initial infection [67].

2.3.4. NHS resource use inputs

NHS healthcare resource use was calculated for all cases receiving outpatient or inpatient medical care. All non-hospitalized symptomatic cases were assumed to have one outpatient (GP) visit (Table 2). Length of stay (bed days) in a general ward [68] or ICU [63] was calculated from UK COVID-19 patient data (Table 2 and Supplementary methods).

2.3.5. Productivity loss inputs

Productivity loss was estimated using a human capital approach [31] and included the number of productive days for persons both in and outside of the paid workforce. Length of COVID-19-related absence was based on the duration of isolation for diagnosed cases (10 days) [69] plus the length of stay for hospitalized cases (Supplementary methods). For those in the paid workforce, productive work days were calculated using UK Office of National Statistics (ONS) workforce participation rates [70]. For the entire model population, productive unpaid work days (e.g. time spent volunteering, caring for others, cleaning) were calculated using UK ONS data collected in 2015 [71]. Unpaid work performed was estimated for each week based on age and weighted by gender as in Mendes et al. 2022 [72].

2.4. Scenario analyses

Scenario analyses explored the public health impact of Spring&Autumn₂₀₂₂ booster programs over an extended time horizon of 4 April 2022 to 2 April 2023 and varied eligible populations and/or uptake. In all scenarios, Omicron was assumed to comprise 100% of infections throughout the time horizon.

2.4.1. Scenario 1 – UK Spring&Autumn₂₀₂₂ booster programs

Scenario 1 estimated the impact of Spring&Autumn₂₀₂₂ booster programs recommended by the JCVI for defined eligible populations. Spring 2022 booster eligibility was limited to individuals aged ≥75 years, and uptake was based on UK booster uptake data as of 12 June 2022 [38] and multiplied by 79.2% to account for the lower uptake of further boosters [73] (Supplementary Table 3).

Autumn 2022 booster eligibility was limited to individuals aged ≥50 years and those aged 5–‍49 years who were in UK Green Book clinical risk groups [26,64]. For individuals aged ≥50 years, Autumn 2022 booster uptake was assumed to be the same as in the Autumn–Winter 2021–2022 booster program up to January 2022 [74]. For individuals aged 5–49 years in clinical risk groups, uptake of 70.0% was assumed across all relevant age groups based on NHS data reporting observed booster uptake among individuals aged 16–64 years in a high-risk or caretaker group through 31 May 2022 [39]. Projected booster program outcomes were compared to a counterfactual scenario in which there were no Spring&Autumn₂₀₂₂ programs.

2.4.2. Scenario 2 – boosters for individuals aged ≥5 years

Scenario 2 assessed the impact of hypothetical alternative Spring&Autumn₂₀₂₂ programs extending booster eligibility to all individuals aged ≥5 years. Booster uptake for individuals aged 5–11 years was assumed based on UK observed coverage for the first dose of COVID-19 primary series vaccination in 5- to 11-year-olds as of 12 June 2022 [38] (Supplementary Table 3). Booster uptake for individuals aged 12–17 years was assumed based on UK observed coverage for 3 or more doses in 12- to 17-year-olds as of 12 June 2022 [38]. Booster uptake for individuals aged ≥18 years was based on UK observed coverage for 3 or more doses as of 3 April 2022 [48] and stratified by age group. Projected booster program outcomes were compared to a counterfactual scenario in which there were no Spring&Autumn₂₀₂₂ programs.

2.4.3. Scenario 3 – increased uptake for individuals aged ≥5 years

Scenario 3 assessed the impact of hypothetical alternative Spring&Autumn₂₀₂₂ programs extending the offer of Spring&Autumn₂₀₂₂ boosters to all individuals aged ≥5 years, while also assuming increased uptake. Booster uptake in individuals aged 5–15 years was assumed to align with seasonal flu coverage of 57.2% (Supplementary Table 3) for school-age children, based on flu coverage rates for 2021–2022 [75]. Booster uptake for individuals aged ≥16 years was assumed to reach the same levels observed for the second primary dose, based on age-stratified uptake data through 12 June 2022: aged 16–17 years: 35.0%; 18–29: 49.9%; 30–49: 64.3%; 50–64: 72.1%; 65–74: 87.2%; ≥75: 92.4% [38]. Projected booster program outcomes were compared to a counterfactual scenario in which there were no Spring&Autumn₂₀₂₂ programs.

2.4.4. Scenario 4 – boosters for high-risk individuals only

Scenario 4 assessed the impact of limited Spring&Autumn₂₀₂₂ booster programs that only offered boosters to high-risk individuals aged ≥5 years. Uptake of 70.0% was assumed across all relevant age groups based on NHS data reporting observed booster uptake among individuals aged 16–64 years in a high-risk or caregiver group through 31 May 2022 [39] (Supplementary Table 3). High-risk individuals were defined according to clinical risk groups identified in the UK Green Book [64] and Walker et al. [76] and stratified by age. Projected booster program outcomes were compared to a counterfactual scenario in which there were no Spring&Autumn₂₀₂₂ programs.

2.4.5. Scenario 5 – increased uptake in high-risk-only programs

Scenario 5 assessed the impact of limited Spring&Autumn₂₀₂₂ booster programs that only offered boosters to high-risk individuals aged ≥ 5 years and assumed increased uptake of 84.9% in this population; the uptake rate was based on the proportion of this population completing primary series vaccination [39] (Supplementary Table 3). High-risk individuals were defined according to clinical risk groups identified in the UK Green Book [64] and Walker et al. [76] and stratified by age. Projected booster program outcomes were compared to a counterfactual scenario in which there were no Spring&Autumn₂₀₂₂ programs.

3. Results

3.1. Base case results

The Autumn–Winter_2021–2022 booster program was estimated to have averted approximately 12.8 million infections (64% reduction), 1.1 million hospitalizations (91% reduction), 3.7 million long COVID cases (68% reduction), and 290,000 deaths (93% reduction) in the UK between January and March 2022 relative to a counterfactual scenario in which booster vaccination did not occur (Figure 2 and Table 3).

Figure 2. Base case health outcomes: new cases, hospitalizations, and deaths by week.

Base case new health outcomes (i.e. incidence) by week per 100,000 population with (green) and without (black) the Autumn–Winter_2021–2022 booster vaccination program. A. New cases per week per 100,000 population. B. New hospitalizations per week per 100,000 population. C. New deaths per week per 100,000 population.

Table 3. Base case results: Modeled population-level outcomes between January and April 2022 with the Autumn–Winter 2021–2022 booster program.

Outcomes ^a	No booster	Booster	Outcomes averted with booster ^b
			Number of outcomes averted	% reduction
Health outcomes (n)
COVID-19 cases	19,893,662	7,093,296	12,800,366	64%
Asymptomatic	10,049,496	3,990,662	6,058,834	60%
Symptomatic	9,844,166	3,102,634	6,741,532	68%
Hospitalizations	1,200,143	111,349	1,088,795	91%
General ward cases	1,009,413	91,668	917,745	91%
ICU cases	190,731	19,680	171,050	90%
Long COVID cases	5,491,621	1,759,386	3,732,235	68%
Deaths	310,854	21,258	289,596	93%
NHS resource use outcomes
GP visits (n)	8,644,023	2,991,286	5,652,737	65%
Bed days in general ward	8,580,008	779,179	7,800,829	91%
Bed days in ICU	1,621,211	167,283	1,453,928	90%
Patient productivity loss
Days of work lost	554,648,660	193,289,694	361,358,966	65%

Abbreviations: COVID-19, coronavirus disease 2019; GP, general practitioner; ICU, intensive care unit

^aNew outcomes (i.e. incidence) estimated for the period between 3 January 2022 and 3 April 2022

^bDifference and percent reduction with booster relative to no booster

From the NHS perspective, the Autumn-Winter_2021–2022 booster program was estimated to have averted 5.7 million GP visits (65% reduction) and freed up approximately 7.8 million general ward bed days (91% reduction) and 1.5 million ICU bed days (90% reduction) to treat patients with other conditions.

From a societal perspective, the Autumn-Winter_2021–2022 booster program was estimated to have averted approximately 361 million days (65% reduction) of productivity loss.

3.2. Scenario results

Projected new outcomes (i.e. incidence) were estimated over an extended time horizon from 4 April 2022–2 April 2023 as detailed below. Figure 3 depicts predicted week-by-week new hospitalizations averted in each scenario.

Figure 3. Scenario projections: Weekly hospitalizations averted with Spring&Autumn_2021–2022 boosters.

Modeled projections of new hospitalizations averted with Spring&Autumn₂₀₂₂ booster vaccination in scenario analyses. New hospitalizations averted are depicted weekly over the extended time horizon of 4 April 2022 to 2 April 2023 for each of the modeled scenarios. Scenario 1: Actual (Spring 2022) and anticipated (Autumn 2022) booster programs. Scenario 2: Alternative Spring&Autumn₂₀₂₂ booster programs extending eligibility to individuals aged ≥5 years. Scenario 3: Alternative Spring&Autumn₂₀₂₂ booster programs extending eligibility to individuals aged ≥5 years and assuming increased uptake among all age groups. Scenario 4: Alternative Spring&Autumn₂₀₂₂ booster programs in which only individuals in Green Book [64] clinical risk groups would be eligible. Scenario 5: Alternative Spring&Autumn₂₀₂₂ booster programs in which only individuals in Green Book [64] clinical risk groups would be eligible and assuming increased uptake in these groups.

3.2.1. Scenario 1 – UK Spring and Autumn 2022 booster programs

Spring&Autumn₂₀₂₂ booster programs offering boosters to older adults and high-risk groups were estimated to avert approximately 6.2 million infections (13% reduction), 716,000 hospitalizations (36% reduction), 1.9 million long COVID cases (15% reduction), and 125,000 deaths (45% reduction) over the time horizon of April 2022 – April 2023 compared to offering no Spring&Autumn₂₀₂₂ boosters (Table 4 and Supplementary Table 5). These Spring&Autumn₂₀₂₂ programs were estimated to free up approximately 2.7 million GP appointments (13% reduction), 5.2 million general ward bed days (37% reduction), and 839,000 ICU bed days (30% reduction) to treat patients with other conditions. An estimated 165 million days of productivity loss (12% reduction) would be averted.

Table 4. Scenario results: Modeled percent reduction in population-level outcomes averted between April 2022 and April 2023 with alternative Spring & Autumn 2022 booster program scenarios.

Outcomes ^a	Scenario 1 ^b	Scenario 2 ^c	Scenario 3 ^d	Scenario 4 ^e	Scenario 5 ^f
Health outcomes
COVID-19 cases	13%	64%	75%	27%	33%
Asymptomatic	12%	63%	74%	25%	31%
Symptomatic	15%	66%	76%	29%	34%
Hospitalizations	36%	78%	85%	48%	55%
General ward cases	37%	78%	85%	48%	55%
ICU cases	30%	78%	84%	47%	54%
Long COVID cases	15%	65%	75%	28%	33%
Deaths	45%	80%	86%	54%	62%
NHS resource use outcomes
GP visits	13%	65%	75%	27%	32%
Bed days in general ward	37%	78%	85%	48%	55%
Bed days in ICU	30%	78%	84%	47%	54%
Patient productivity loss
Total days lost	12%	64%	75%	26%	32%

Abbreviations: COVID-19, coronavirus disease 2019; GP, general practitioner; ICU, intensive care unit

^aNew outcomes (i.e. incidence) predicted for the period between 4 April 2022 and 2 April 2023

^bActual (Spring 2022) and anticipated (Autumn 2022) booster programs

^cAlternative Spring&Autumn₂₀₂₂ booster programs extending eligibility to individuals aged ≥ 5 years

^dAlternative Spring&Autumn₂₀₂₂ booster programs extending eligibility to individuals aged ≥ 5 years and assuming increased uptake among all age groups

^eAlternative Spring&Autumn₂₀₂₂ booster programs in which only individuals in Green Book [64] clinical risk groups would be eligible

^fAlternative Spring&Autumn₂₀₂₂ booster programs in which only individuals in Green Book [64] clinical risk groups would be eligible and assuming increased uptake in these groups

3.2.2. Scenario 2 – boosters for individuals aged ≥5 years

Spring&Autumn₂₀₂₂ booster programs offering boosters to individuals aged ≥5 years were estimated to avert approximately 30.0 million infections (64% reduction), 1.6 million hospitalizations (78% reduction), 8.3 million long COVID cases (65% reduction), and 222,000 deaths (80% reduction) over the time horizon of April ‍2022 – April 2023 compared to offering no Spring&Autumn₂₀₂₂ boosters (Table 4 and Supplementary Table 6). The programs were estimated to free up approximately 13.2 million GP appointments (65% reduction), 11.1 million general ward bed days (78% reduction), and 2.2 million ICU bed days (78% reduction). An estimated 859 million days of productivity loss (64% reduction) would be averted.

3.2.3. Scenario 3 – increased uptake for individuals aged ≥5 years

Spring&Autumn₂₀₂₂ booster programs including individuals aged ≥5 years and assuming increased uptake across all age groups were estimated to avert approximately 32.9 million infections (75% reduction) (Table 4 and Supplementary Table 7), 1.6 million hospitalizations (85% reduction), 9.2 million long COVID cases (75% reduction), and 228,000 deaths (86% reduction) over the time horizon of April 2022 – April 2023 compared to offering no Spring&Autumn₂₀₂₂ boosters. The programs were estimated to free up approximately 14.5 million GP appointments (75% reduction), 11.5 million general ward bed days (85% reduction), and 2.3 million ICU bed days (84% reduction). An estimated 953 million days of productivity loss (75% reduction) would be averted.

3.2.4. Scenario 4 – boosters for high-risk individuals only

Offering Spring&Autumn₂₀₂₂ boosters to only high-risk individuals (aged ≥5 years) was estimated to avert approximately 12.6 million infections (27% reduction) (Table 4 and Supplementary Table 8), 892,000 hospitalizations (48% reduction), 3.6 million long COVID cases (28% reduction), and 136,000 deaths (54% reduction) over the time horizon of April 2022 – April 2023 compared to offering no Spring&Autumn₂₀₂₂ boosters. High-risk-only Spring&Autumn₂₀₂₂ booster programs were estimated to free up approximately 5.5 million GP appointments (27% reduction), 6.4 million general ward bed days (48% reduction), and 1.2 million ICU bed days (47% reduction) to treat patients with other conditions. An estimated 355 million days of productivity loss (26% reduction) would be averted.

3.2.5. Scenario 5 – improved uptake in high-risk-only programs

Offering Spring&Autumn₂₀₂₂ boosters to only high-risk individuals aged ≥ 5 years and assuming improved uptake among this population was estimated to avert approximately 15.2 million infections (33% reduction) (Table 4 and Supplementary Table 9), 1.0 million hospitalizations (55% reduction), 4.3 million long COVID cases (33% reduction), and 154,000 deaths (62% reduction) over the time horizon of April 2022 – April 2023 compared to offering no Spring&Autumn₂₀₂₂ boosters. High-risk Spring&Autumn₂₀₂₂ programs with increased uptake were estimated to free up approximately 6.7 million GP appointments (32% reduction), 7.3 million general ward bed days (55% reduction), and 1.4 million ICU bed days (54% reduction) to treat patients with other conditions. An estimated 428 million days of productivity loss (32% reduction) would be averted.

4. Discussion

This study estimated that the Autumn–Winter_2021–2022 booster program provided substantial benefit to UK public health during January – March 2022, an Omicron-predominant period. Relative to a counterfactual scenario in which no booster program had taken place, the Autumn–Winter_2021–2022 booster program was estimated to have freed up approximately 9.3 million hospital bed days during this three-month period, alleviating significant burden to the NHS. An estimated ~3.7 million long COVID cases were averted, which may provide NHS resource savings in the longer term. Additionally, boosters were estimated to avert approximately 361 million days of productivity loss, suggesting considerable benefit to society.

UK Spring&Autumn₂₀₂₂ booster programs covering older adults and high-risk individuals, as per JCVI recommendation, were also estimated to generate substantial public health benefit relative to offering no boosters. Over a one-year period (from 4 April 2022–2 April 2023), the booster programs were estimated to reduce COVID-19 cases by 13%, hospitalizations by 36%, and deaths by 45%. Extending eligibility for Spring&Autumn₂₀₂₂ booster programs to individuals aged ≥5 years was estimated to improve benefit compared to actual or planned Spring&Autumn₂₀₂₂ booster programs, averting approximately twice as many hospitalizations (2.2x) and deaths (1.8x) and releasing nearly 7.2 million more bed days.

Extending eligibility to individuals aged ≥5 years combined with increasing booster uptake in all age groups in Spring&Autumn₂₀₂₂ programs was estimated to improve benefit further. This scenario reduced hospitalizations and deaths by an additional 6 percentage points compared to Spring&Autumn₂₀₂₂ programs that had extended eligibility but more realistic uptake assumptions. Such programs with extended eligibility and increased uptake among younger age groups, in line with JCVI program recommendations, would reduce hospitalizations by 85% relative to having no Spring&Autumn₂₀₂₂ programs.

Prior literature has not evaluated the public health impact of extending booster vaccination to children. However, a UK modeling study included a scenario analysis that considered the impact of offering children aged ≥5 years primary series vaccines and projected potential infection and severe disease outcomes between May and December 2022 [77]. In contrast with the present study, Barnard et al. found that vaccinating children would provide only a small improvement in public health impact and would not significantly alter the projected virus transmission dynamics. Differences between study findings are likely attributable to substantial differences in several all-age model inputs, for instance, the slower waning of infection- and vaccine-induced immunity and greater vaccination protection against transmission assumed in the prior model. These parameter choices in the prior model contribute to a much smaller susceptible population and reduced overall transmission, which would likely limit the potential impact of vaccinating children on reducing transmission.

Another UK model, based on data during Delta predominance, estimated that extending primary series vaccination to children aged 12–17 years would reduce hospitalizations and deaths by 8–‍10% and extending primary vaccination to children aged 5–11 years would have a far lesser impact [78]. Estimates in the present model showed a much greater impact of (booster-) vaccinating children; this may be due to the increased contagiousness of the Omicron variant and differences in model parameters. First, the prior model assumed behavioral modifications that reduced mixing within the model population. Variations in these mixing parameters influenced the estimated impact of vaccinating children. In addition, many children were not yet eligible for vaccination or could only receive a first primary series dose during much of the time horizon in the prior study. As periods before vaccination could confer no protection, the overall estimated impact of pediatric vaccination would inherently be lower in the prior study than in the present study where children were eligible for booster vaccination throughout the time horizon.

Similarly, limited work has estimated the effectiveness of extending booster vaccination to young and middle-aged adults [77,79]. Scenarios in this study extended eligibility for Spring&Autumn₂₀₂₂ booster programs to individuals aged ≥5 years, which includes individuals aged 5–74 years not eligible for Spring 2022 programs and individuals aged 5–49 years not eligible for Autumn 2022 programs. An Israeli modeling study on booster vaccination of young and middle-aged adults (aged 16–59 years) during Delta predominance estimated substantial public health benefit from including these populations in booster programs. Not offering an initial booster dose to individuals aged 16–59 years was estimated to increase cases, severe cases, and deaths by 272–‍397% [79]. The extent of benefit estimated from boosters was greater in the Israeli study than in the present study; this may be due to differences between the Delta and Omicron variants, a potential reduced benefit of subsequent versus initial boosters, societal differences in health or behavior, or other differences in model dynamics.

The public health value of vaccinating lower-risk groups remains widely debated, particularly for children and young people [80,81], who have relatively low rates of hospitalization [82] and mortality from COVID-19 [82,83]. Although children and lower-risk adults are unlikely to become severely ill themselves, these populations may nonetheless spread SARS-CoV-2, especially in a setting of a highly contagious variant such as Omicron. For instance, UK CoMix contact dynamics data used in this study suggest that children have substantial contacts with parents and grandparents. Limited real-world evidence from Denmark suggests that individuals who had received a booster were less likely than those who had received only primary vaccination to transmit COVID-19 Omicron infections within their households [84]. As such, it is reasonable that boosters could help prevent close-contact transmission to high-risk individuals, in addition to reducing overall spread.

The lack of information regarding which populations should be offered booster vaccination has led to different national policies, even among high-income countries with the resources to extend vaccine eligibility. For instance, the age at which the general population was eligible for a second booster in early–mid 2022 varied substantially among countries [20,21,24,85], from 50 years in the US [22] to 75 years in the UK [24].

As the UK has prioritized protecting the most vulnerable throughout its primary [86] and booster [11,24,26] vaccination programs, the present study also assessed the impact of boosting high-risk individuals. It is widely accepted that individuals with certain clinical conditions are more likely to experience severe illness from COVID-19 [64], but the overall public health impact attributable to boosting high-risk groups is unknown. Offering Spring&Autumn₂₀₂₂ boosters to only individuals in clinical risk groups was predicted to substantially reduce the severe outcomes of COVID-19 relative to not offering Spring&Autumn₂₀₂₂ booster vaccinations, reducing hospitalizations by 48% and deaths by 54%. Estimates suggested that overall cases would be averted by a more modest 27%. All outcomes were improved further in a scenario analysis assuming increased uptake in these alternative high-risk-only Spring&Autumn₂₀₂₂ programs, despite this population already having a high rate of booster coverage (70.0% observed uptake) [39]. High-risk-only programs assuming increased uptake to reach 84.9% coverage were estimated to provide a 55% reduction in hospitalizations, 62% reduction in deaths, and 33% reduction in COVID-19 cases relative to no Spring&Autumn₂₀₂₂ vaccination. However, these increased uptake projections may represent a ‘ceiling’ of potential public health benefit from high-risk-only programs. In contrast, alternative programs extending age-based eligibility and assuming increased uptake were projected to have the greatest overall public health impact, reducing hospitalizations by 85%, deaths by 86%, and cases by 75% relative to having no Spring&Autumn₂₀₂₂ programs.

Long COVID cases were also included in the model. According to model estimates, booster programs averted long COVID cases at similar rates to overall COVID-19 cases. The number of long COVID cases was estimated to be reduced by 68% in the Autumn–Winter 2021–2022 program relative to having no booster program. Spring&Autumn₂₀₂₂ programs were projected to reduce long COVID cases by 15% relative to having no booster programs. Hypothetical alternative Spring&Autumn₂₀₂₂ programs could have a greater impact; programs offering boosters only to high-risk individuals could reduce long COVID cases by 28%, while programs extending eligibility to all individuals aged ≥5 years could reduce long COVID cases by 65% relative to having no booster programs. Estimating the prevalence of long COVID is important, as long COVID may have far-reaching impacts on productivity loss and NHS resource use at present and for years to come.

The dynamic transmission model used in this study allowed herd effects of transmission and booster vaccination to be estimated among all persons in the UK, including those not eligible for booster vaccination. A further strength of this study was its use of clinical, vaccine, and NHS resource use inputs derived from published UK data. For instance, booster VE inputs were based on VE values for both BNT162b2 and mRNA-1273 boosters. As UK boosters have comprised a mix of these vaccine brands (76.4% BNT162b2 and 23.4% mRNA-1273 as of 26 July 2022 [87]), model estimates should better reflect effectiveness than analyses assuming use of a single vaccine brand. The base case and all scenarios employed booster programs timed to align with the schedule of actual UK programs as conducted (Autumn–Winter_2021–2022 and Spring 2022) or planned (Autumn 2022) [24]. In addition, the inclusion of Green Book [64] clinical risk groups in the model enabled a closer approximation of the impacts of severe COVID-19 disease, and the inclusion of long COVID outcomes may be useful when considering the longer-term benefits of vaccination. This study analyzed the public health impact of boosters considering productivity loss from both those in and not in the paid workforce, thereby providing a more comprehensive picture of booster impact across the UK population.

Calibrated case rate values provided a good replication of the observed case rates in the base case, overestimating total cases by 0.3% in the whole modeled population across the calibration period (Supplementary methods). Although hospitalization rates were also a calibration target, the model underestimated observed hospitalizations by 30%. The underestimation of hospitalizations is a limitation of the model and likely led to underestimating the number of hospitalizations averted by booster programs. As such, the model is thought to provide a conservative estimate of booster program impact on hospitalizations. Deaths were overestimated by 13% due to limitations in IFR input calculations; it is uncertain how this overestimation in deaths affected the estimation of booster program impact on mortality. This model used a deterministic approach in model calibration given the limited data availability for model parameters. This meant that it did not account for parameter uncertainty or sample from an assumed distribution of the uncertain parameter to generate a range of estimates.

Further limitations of this study include the use of several assumptions and simplifications in model inputs due to a lack of appropriate data or to limit model complexity. First, scenario analyses projected outcomes over an extended time horizon from April 2022–April 2023, assuming that Omicron predominance would continue throughout this period. If other variants were to emerge that varied in severity, transmissibility, or escape of vaccine- or infection-induced protection, model estimates would differ. As it is still unclear whether COVID-19 surges seasonally [88], the model did not apply seasonal forcing. Next, this study did not include the impact of non-pharmaceutical interventions (e.g. mask wearing, social distancing) that may have reduced the number of SARS-CoV-2 infections. The study also did not include the impact of pharmaceutical treatments (e.g. anti-virals, monoclonal antibodies) that may reduce the number of severe outcomes A further limitation was that the model did not explicitly account for cross-immunity from prior infections with the Delta variant. However, the infection rate parameter was calibrated to match observed infection rates, which means that the effect of cross-immunity, as expressed in observed infection targets, was also captured in calibrated infection rate parameters. In addition, consideration of socioeconomic status, which may affect both vaccine uptake rates and disease outcomes, was beyond the scope of this study. The results of this study cannot necessarily be generalized to populations, timeframes, or disease variants not included in the model.

Continued research is needed to characterize the public health impact of booster vaccination in the UK across different population subgroups and timeframes whilst considering other variables that may impact disease transmission or severity. Future work should investigate the optimal frequency of revaccination and how this may vary across different age and clinical risk groups.

5. Conclusions

This modeling study estimated that UK booster vaccination against SARS-CoV-2 provided substantial benefit to public health and would continue to provide benefit over an extended time horizon amid Omicron predominance.

The Autumn–Winter_2021–2022 booster program was estimated to have averted approximately 290,000 deaths and 1.1 million hospitalizations between January and March 2022, and actual Spring&Autumn₂₀₂₂ booster programs were estimated to avert a total of approximately 125,000 deaths and 716,000 hospitalizations between April 2022 and April 2023.

All public health outcomes estimated, including COVID-19 cases, long COVID cases, patient productivity loss, hospitalizations, and deaths, were improved considerably in counterfactual Spring&Autumn₂₀₂₂ program scenarios compared to a scenario based on the actual Spring&Autumn₂₀₂₂ programs. The greatest public health benefit was observed in a counterfactual scenario that extended booster eligibility to individuals ≥5 years old and assumed increased uptake across all eligible age groups.

These results suggest that public health benefits could be maximized by extending eligibility for booster vaccination programs and increasing uptake.

Declaration of interest

D Mendes, JL Nguyen, L Hamson, M Di Fusco, C Czudek, and J Yang are employees of Pfizer and may own Pfizer stock(s). R Chapman, E Aruffo, and P Gal are employees of Evidera, which received financial support from Pfizer, Inc. in connection with the study and the development of this manuscript. 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 material discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Reviewers on this manuscript have received an honorarium for their review work. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

Input into early model development was provided by Julie Roiz (Evidera) and was funded by Pfizer, Inc. Medical writing was provided by Jacqueline Janowich Wasserott, PhD and Jonathan Pitt, PhD (Evidera) and was funded by Pfizer, Inc.

Data availability statement

Data generated or analyzed during this study are available upon request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14760584.2023.2158816

Additional information

Funding

This study was sponsored by Pfizer, Inc.