Intrinsic randomness in epidemic modelling beyond statistical uncertainty
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 empirical...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-06-01
|
| Series: | Communications Physics |
| Online Access: | https://doi.org/10.1038/s42005-023-01265-2 |
| _version_ | 1827916916027555840 |
|---|---|
| author | Matthew J. Penn Daniel J. Laydon Joseph Penn Charles Whittaker Christian Morgenstern Oliver Ratmann Swapnil Mishra Mikko S. Pakkanen Christl A. Donnelly Samir Bhatt |
| author_facet | Matthew J. Penn Daniel J. Laydon Joseph Penn Charles Whittaker Christian Morgenstern Oliver Ratmann Swapnil Mishra Mikko S. Pakkanen Christl A. Donnelly Samir Bhatt |
| author_sort | Matthew J. Penn |
| collection | DOAJ |
| description | 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. |
| first_indexed | 2024-03-13T03:21:29Z |
| format | Article |
| id | doaj.art-b4e068aec9fa4456ba8fa48ec68bccac |
| institution | Directory Open Access Journal |
| issn | 2399-3650 |
| language | English |
| last_indexed | 2024-03-13T03:21:29Z |
| publishDate | 2023-06-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Communications Physics |
| spelling | doaj.art-b4e068aec9fa4456ba8fa48ec68bccac2023-06-25T11:19:29ZengNature PortfolioCommunications Physics2399-36502023-06-01611910.1038/s42005-023-01265-2Intrinsic randomness in epidemic modelling beyond statistical uncertaintyMatthew J. Penn0Daniel J. Laydon1Joseph Penn2Charles Whittaker3Christian Morgenstern4Oliver Ratmann5Swapnil Mishra6Mikko S. Pakkanen7Christl A. Donnelly8Samir Bhatt9University of OxfordImperial College LondonUniversity of OxfordImperial College LondonImperial College LondonImperial College LondonUniversity of CopenhagenImperial College LondonUniversity of OxfordImperial College LondonAbstract 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.https://doi.org/10.1038/s42005-023-01265-2 |
| spellingShingle | Matthew J. Penn Daniel J. Laydon Joseph Penn Charles Whittaker Christian Morgenstern Oliver Ratmann Swapnil Mishra Mikko S. Pakkanen Christl A. Donnelly Samir Bhatt Intrinsic randomness in epidemic modelling beyond statistical uncertainty Communications Physics |
| title | Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| title_full | Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| title_fullStr | Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| title_full_unstemmed | Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| title_short | Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| title_sort | intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| url | https://doi.org/10.1038/s42005-023-01265-2 |
| work_keys_str_mv | AT matthewjpenn intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT danieljlaydon intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT josephpenn intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT charleswhittaker intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT christianmorgenstern intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT oliverratmann intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT swapnilmishra intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT mikkospakkanen intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT christladonnelly intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT samirbhatt intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty |