Abstract:Uncertainty can be classified as either aleatoric (intrinsic randomness) or epistemic (imperfect knowledge of parameters). The majority of frameworks assessing infectious disease risk consider only epistemic uncertainty. We only ever observe a single epidemic, and therefore cannot empirically determine aleatoric uncertainty. Here, we characterise both epistemic and aleatoric uncertainty using a time-varying general branching process. Our framework explicitly decomposes aleatoric variance into mechanistic components, quantifying the contribution to uncertainty produced by each factor in the epidemic process, and how these contributions vary over time. The aleatoric variance of an outbreak is itself a renewal equation where past variance affects future variance. We find that, superspreading is not necessary for substantial uncertainty, and profound variation in outbreak size can occur even without overdispersion in the offspring distribution (i.e. the distribution of the number of secondary infections an infected person produces). Aleatoric forecasting uncertainty grows dynamically and rapidly, and so forecasting using only epistemic uncertainty is a significant underestimate. Therefore, failure to account for aleatoric uncertainty will ensure that policymakers are misled about the substantially higher true extent of potential risk. We demonstrate our method, and the extent to which potential risk is underestimated, using two historical examples.

What problem does this paper attempt to address?

The problem this paper attempts to address is how to comprehensively consider uncertainty in infectious disease modeling, particularly the impact of aleatoric uncertainty on prediction and risk assessment. Most existing infectious disease risk assessment frameworks primarily focus on epistemic uncertainty, which is the uncertainty due to incomplete knowledge of parameters. However, the actual observed epidemic occurs only once, so aleatoric uncertainty cannot be empirically determined. This paper proposes a time-varying general branching process model that incorporates both aleatoric and epistemic uncertainties, explicitly decomposing the variance of aleatoric uncertainty into mechanistic components, quantifying the contribution of each factor to the uncertainty in the epidemic process and its variation over time. Specifically, the main objectives of the paper include: 1. **Quantifying Aleatoric Uncertainty**: Using the branching process model to quantify the impact of aleatoric uncertainty on epidemic dynamics, particularly in the absence of superspreading events, where the epidemic size can still exhibit significant variation. 2. **Assessing Predictive Uncertainty**: Demonstrating that predictions considering only epistemic uncertainty can severely underestimate actual uncertainty, potentially leading policymakers to misjudge the potential risks. 3. **Providing Methodological Tools**: Using two historical cases (the 2003 Hong Kong SARS epidemic and the early 2020 UK COVID-19 epidemic) to illustrate how to use the model to assess aleatoric uncertainty and conduct risk assessment with limited data. In summary, this paper aims to improve the accuracy of infectious disease prediction and risk assessment by introducing a new analytical framework that comprehensively considers both aleatoric and epistemic uncertainties, providing more reliable support for policy-making.

Intrinsic Randomness in Epidemic Modelling Beyond Statistical Uncertainty

Epidemic modelling: aspects where stochasticity matters

Epidemic modelling requires knowledge of the social network

Epidemic models with uncertainty in the reproduction number

Stochastic Epidemic Modelling

Superspreading and Heterogeneity in Epidemics

Deterministic epidemic models overestimate the basic reproduction number of observed outbreaks

Epidemic Forecast Follies

Epidemionics: from the host-host interactions to the systematic analysis of the emergent macroscopic dynamics of epidemic networks

A comprehensive probabilistic analysis of approximate SIR‐type epidemiological models via full randomized discrete‐time Markov chain formulation with applications

Inference in Stochastic Epidemic Models via Multinomial Approximations

Inferring Epidemics from Multiple Dependent Data via Pseudo-Marginal Methods

Avoidable errors in the modeling of outbreaks of emerging pathogens, with special reference to Ebola

Modelling preventive measures and their effect on generation times in emerging epidemics

Statistical Mechanics of Inference in Epidemic Spreading

Challenges in Estimating Time-Varying Epidemic Severity Rates from Aggregate Data

Modelling the Stochastic Importation Dynamics and Establishment of Novel Pathogenic Strains using a General Branching Processes Framework

Inference for ecological dynamical systems: a case study of two endemic diseases

Stochastic Modeling of an Infectious Disease, Part I: Understand the Negative Binomial Distribution and Predict an Epidemic More Reliably

Analysing Multiple Epidemic Data Sources

Epidemic Models with Random Infectious Period