Intrinsic randomness in epidemic modelling beyond statistical uncertainty
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 determ...
| Glavni autori: | , , , , , , , , , |
|---|---|
| Format: | Journal article |
| Jezik: | English |
| Izdano: |
Springer Nature
2023
|
| _version_ | 1826310419399049216 |
|---|---|
| author | Penn, MJ Laydon, DJ Penn, J Whittaker, C Morgenstern, C Ratmann, O Mishra, S Pakkanen, MS Donnelly, CA Bhatt, S |
| author_facet | Penn, MJ Laydon, DJ Penn, J Whittaker, C Morgenstern, C Ratmann, O Mishra, S Pakkanen, MS Donnelly, CA Bhatt, S |
| author_sort | Penn, MJ |
| collection | OXFORD |
| description | 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 | 2024-03-07T07:51:56Z |
| format | Journal article |
| id | oxford-uuid:985eccc0-3f84-4803-81d5-78fa51f61114 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:51:56Z |
| publishDate | 2023 |
| publisher | Springer Nature |
| record_format | dspace |
| spelling | oxford-uuid:985eccc0-3f84-4803-81d5-78fa51f611142023-07-18T11:23:31ZIntrinsic randomness in epidemic modelling beyond statistical uncertaintyJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:985eccc0-3f84-4803-81d5-78fa51f61114EnglishSymplectic ElementsSpringer Nature2023Penn, MJLaydon, DJPenn, JWhittaker, CMorgenstern, CRatmann, OMishra, SPakkanen, MSDonnelly, CABhatt, SUncertainty 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. |
| spellingShingle | Penn, MJ Laydon, DJ Penn, J Whittaker, C Morgenstern, C Ratmann, O Mishra, S Pakkanen, MS Donnelly, CA Bhatt, S Intrinsic randomness in epidemic modelling beyond statistical uncertainty |
| 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 |
| work_keys_str_mv | AT pennmj intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT laydondj intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT pennj intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT whittakerc intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT morgensternc intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT ratmanno intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT mishras intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT pakkanenms intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT donnellyca intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty AT bhatts intrinsicrandomnessinepidemicmodellingbeyondstatisticaluncertainty |