Fig. 5: Example Bayesian network (BN) for modelling the risk of developing background myocarditis over 2 months based on age and sex.

The output node, ‘Background myocarditis over 2 months’ is the child of two linked (arrow) parent nodes, ‘Age group’ and ‘Sex’. As these parent nodes do not have parent themselves, the probabilities of each of their possible states are determined by a prior distribution; the model adopts the age distribution of the Australian population and an even distribution of males and females. The conditional probability table for the outcome node ‘Background myocarditis over 2 months’, gives the probability for each state of this node dependent on the parent node states. a In the default state, the BN shows that the chance of developing background myocarditis (not from COVID-19 or the Pfizer vaccine) over 2 months is 0.003% (e.g., in a population of 100,000 people, we expect three to get myocarditis in a 2-month period). b An example of scenario analysis showing the chance of a 40–49-year-old male (underlined) developing background myocarditis over 2 months, the model calculates a 0.004% chance of myocarditis.