Table 1 Summary of data sources, assumptions, and prior distributions

From: A Bayesian network analysis of the Pfizer COVID-19 vaccine in the paediatric population

Model inputs

Data sources, assumptions, rationale (references)

Vaccine effectiveness against symptomatic infection

Data from the Increase Community Access to Testing (ICATT) platform was used33, on page 2215 of the attached reference. Here data were collected from 49 states in the USA, of children and adolescents presenting to drive-through sites for testing between 26/12/2021 and 21/02/2022 (predominantly Omicron transmission period). A test negative case control analysis was conducted to calculate vaccine effectiveness against symptomatic infection. Immunocompromised cases were excluded from the analysis.

A child was considered vaccinated if tests were taken at least 2 weeks after the second vaccine dose. We used data for vaccine effectiveness at 2 months post second dose.

Vaccine effectiveness (reported on page 2215) is 16.6% in the 12–15 year-old age group in the period 30–90 days post second dose. We assumed the same vaccine effectiveness for our age cohort of 12–17.

Vaccine effectiveness was reported as 28.9% in the 5–11 year-old age group the period 30–90 days post second dose33.

Please see supplementary Table 1 for final assumptions.

Risk of symptomatic infection under current transmission status

Cases of COVID-19 from NSW between 1/1/2022 and 30/7/22 were used to estimate risk of symptomatic infection under current transmission status for our age cohorts of 5–11 and 12–17 year-olds34. Found at: https://www.health.nsw.gov.au/Infectious/covid-19/Documents/weekly-covid-overview-20220730.pdf.

Census data from Australian Bureau of Statistics was used to estimate the probability of each specific age and gender cohort being infected. This reports Australian demographic data from the 31/12/202135

Please see supplementary Tables 2-2.5 for calculations and assumptions.

Community transmission levels

We selected community transmission scenarios of 1%, 2%, 5% and 10% of the population being infected with SARS-CoV-2 over a 2-month period for the purposes of the model. This refers to percentage of the whole population infected over a 2 month period.

Risk of developing (background) myocarditis

A population based retrospective observational study of linked administrative databases from Ontario, Canada was used to populate the risk of developing background myocarditis. The study identifies hospitalisations and emergency department visits for myocarditis between 2015 and 2020. The study would miss mildly symptomatic and asymptomatic myocarditis, however would capture all cases of clinical concern.

As the incidence rates of myocarditis are significantly higher in the 16–17 cohort compared for 12–15 cohort, we have stratified this outcome to 5–11, 12–15 and 16–17. Furthermore the paper reports the incidence of myocarditis over a 12 month period, we have converted this to a 3 month period.

Supplemental Table S6 of the paper provides the data stratified to sex and age. We assumed the same myocarditis rates in the 16–19 age cohort presented in the paper for our 16–17 age cohort.

Please see supplementary table 3 for final assumptions36.

Risk of developing Pfizer vaccine-associated myocarditis

For ages 5–11 and 12–17 years, second dose data were reported by the Therapeutic Goods Administration’s weekly vaccine update released on 25/08/22. The TGA reported myocarditis cases included all those diagnosed after a second Pfizer vaccine; however it did not define if cases were attributable to vaccination. Therefore, we assume this includes the background myocarditis rates. Furthermore, the TGA reports that the majority of cases of vaccine-associated myocarditis occur in the first 3–5 days post vaccination, with data collected up to the 10 day mark.

Data will be updated and booster information included once more Australian data are available. We assume that the data includes both the background myocarditis rate and risk of developing Pfizer induced myocarditis.

See Supplementary Table 4 for final assumptions37.

Risk of developing

SARS-CoV-2 infection-induced myocarditis

Data sourced from a retrospective cohort of 40 health care systems using electronic health records between January 1 2021 and January 31 2022. The study population included persons with documented SARS-CoV-2 testing, viral illness diagnostic codes or COVID-19 vaccination during the study period. The authors then compared incidence of cardiac complications, in particular myocarditis and pericarditis post infection and vaccination.

The incidence rate of myocarditis was calculated in the first 21 days post SARS-CoV-2. Cases were excluded if they had received a vaccination in the 30 days prior to or post SARS-CoV-2 infection38.

See supplementary Table 5 for final assumptions.

Vaccine effectiveness against hospitalisation from COVID-19

The data were sourced from a case control, test-negative design of vaccine effectiveness against laboratory confirmed COVID-19 leading to hospitalisation in 31 hospitals across 23 states in the USA. The study period was from 1/7/2021- 17/2/2022, with the Omicron period defined from 19/12/2021- 17/2/2022.

Adolescents were defined as 12–18 years of age, and we assumed the same vaccine effectiveness for 12–17-year-olds. We used vaccine effectiveness data for 2–22 2 doses of Pfizer vaccination (43%).

For 5–11-year-olds, the vaccine effectiveness reported was 68% during the Omicron period. The median time since vaccination was 34 days, the paper does not stratify vaccine effectiveness data for time from vaccination The paper reported vaccine effectiveness at 2–22 weeks post vaccination for 12–18 year olds, but median time since vaccination for 5–11 year olds39.

See Supplementary Table 6 for final assumptions.

Risk of hospitalisation from COVID-19

5–11 age group: We used data from a Singaporean study conducted between 21/1/22 and 8/4/22 of all reported cases of COVID-19 to the Ministry of Health. The study collected data on all hospital admissions in the unvaccinated population with 146 hospitalised cases amongst a 16,909 unvaccinated children. Therefore 0.86% of the unvaccinated population between 5 and 11 year-olds were admitted to hospital31. We assume all hospitalisations were due to SARS-CoV-2 infection.

12–17 age group: We used data from four linked New York databases for COVID-19 cases and vaccinations between 30/11/2021 and 30/1/2022. From the 3/1/2022 over 90% of sequenced cases were Omicron. Therefore, from the tables provided in the study we calculated the total number of cases of COVID-19 in the unvaccinated population and the total number of hospitalised cases in the unvaccinated population from the 3/1/2022 to 30/1/202232.

See supplementary Table 7 for final assumptions.

Vaccine effectiveness against a severe outcome from COVID-19

The data were sourced from a case control, test-negative design of vaccine effectiveness against Critical COVID-19 in 31 hospitals across 23 states in the USA. In the paper, critical COVID-19 was defined as cases requiring life support (i.e., non invasive ventilation, invasive ventilation, vasoactive infusions, extracorporeal membrane oxygenation or death). In our model, we defined severe outcome as Intensive Care Unit admission or death.

The study period was from 1/7/2021 to 17/2/2022, with the Omicron period from 19/12/2021 to 17/2/2022.

Adolescents were defined as 12–18 years of age, and we assumed the same vaccine effectiveness for 12–17 year-olds. Vaccine effectiveness against critical COVID-19 is reported as 79%.

The study had insufficient data to report vaccine effectiveness against Critical COVID-19 in the 5–11 age group, we assume the same vaccine effectiveness in the 5–11 age group as 12–1839.

See Supplementary Table 8 for final assumptions.

Risk of suffering a severe outcome from COVID-19

“ Severe Outcome” is defined as Intensive Care Admission or death for SARS-CoV-2 infection.

The data were sourced by approaching the Paediatric Active Enhanced Surveillance (PAEDS) authors29. This is a hospital-based active surveillance system for multiple childhood diseases including COVID-19. Pre-print data was made available by the group for suffering a severe outcome due to COVID-19 in Australia between the 1/12/2021 and 31/08/2022, defined as the Omicron period. “Severe Outcome” where COVID-19 was not the primary cause for admission, but an associated diagnosis were excluded from the data. Furthermore vaccination status for cases is not available, with the “Risk of Severe Outcome from COVID-19” in the unvaccinated population, likely being an underestimation, as the data was collected when vaccinations were available for the paediatric population.

Please see supplementary Table 9.

Vaccine effectiveness against SARS-CoV-2 infection-induced MSI

The evaluation of COVID-19 vaccine and MIS-C was conducted across 29 hospitals in 22 US states between the 1st July 2021 and 7th April 2022. A test negative case control design was used to estimate vaccine effectiveness. Fully vaccinated was defined as 2 doses of Pfizer Vaccine at least 28 days prior to hospital admission40. The Omicron period was defined as January 1st onwards, the study reports the Omicron variant exceeded 50% of all SARS-CoV-2 infections from the 18th December onwards, and with a likely lag of 2–4 weeks from infection to development of MSI, January 1st was deemed the Omicron period.

For the 12–18 age group, Fig. 4 of the paper shows and aOR of 0.08 of developing MSI in vaccinated population vs unvaccinated for the Omicron variant and therefore a Vaccine Effectiveness of 92%.

The paper does not stratify the 5–11 age group by variant and reports an overall aOR of 0.22, of developing MSI in vaccinated vs unvaccinated 5–11 years olds and therefore vaccine efficacy of 78% (1-aOR*100). We have assumed this vaccine effectiveness to represent the Omicron variant

See Supplementary Table 10 for final assumptions.

Risk of developing

SARS-CoV-2 infection-induced MIS-C

The data were sourced by approaching the Paediatric Active Enhanced Surveillance (PAEDS) authors29. This is a hospital-based active surveillance system for multiple childhood diseases including COVID-19. Pre-print data was made available by the group for “Risk of developing SARS-CoV-2 infection induced MIS-C” in Australia between the 1/12/2021 and 31/08/2022, defined as the Omicron period. Furthermore, vaccination status for cases is not available, with the “Risk of developing SARS-CoV-2 infection induced MIS-C” in the unvaccinated population, likely being an underestimation, as the data was collected when vaccinations were available for the paediatric population in Australia.

Risk of suffering a severe outcome from SARS-CoV-2 infection-induced MIS-C

Data was sourced from a study of 29 hospitals in 22 US states between 1st July 2021 and 7th April 2022, therefore spanning both Omicron and Delta variants. This study was test negative case control, primarily designed to estimate vaccine effectiveness vs MIS-C. In supplementary Table 4 however the paper does describe the percentage of cases admitted to intensive care during the Omicron predominant period. It reports 58.5% of MSI cases were admitted to Intensive care in the Omicron predominant period across both age cohorts40.

Please see supplementary Table 12 for final assumptions