RT Journal Article
SR Electronic
T1 Transformation of forecasts for evaluating predictive performance in an epidemiological context
JF medRxiv
FD Cold Spring Harbor Laboratory Press
SP 2023.01.23.23284722
DO 10.1101/2023.01.23.23284722
A1 Bosse, Nikos I.
A1 Abbott, Sam
A1 Cori, Anne
A1 van Leeuwen, Edwin
A1 Bracher, Johannes
A1 Funk, Sebastian
YR 2023
UL http://medrxiv.org/content/early/2023/01/24/2023.01.23.23284722.abstract
AB Forecast evaluation plays an essential role in the development cycle of predictive epidemic models and can inform their use for public health decision-making. Common scores to evaluate epidemiological forecasts are the Continuous Ranked Probability Score (CRPS) and the Weighted Interval Score (WIS), which are both measures of the absolute distance between the forecast distribution and the observation. They are commonly applied directly to predicted and observed incidence counts, but it can be questioned whether this yields the most meaningful results given the exponential nature of epidemic processes and the several orders of magnitude that observed values can span over space and time. In this paper, we argue that log transforming counts before applying scores such as the CRPS or WIS can effectively mitigate these difficulties and yield epidemiologically meaningful and easily interpretable results. We motivate the procedure threefold using the CRPS on log-transformed counts as an example: Firstly, it can be interpreted as a probabilistic version of a relative error. Secondly, it reflects how well models predicted the time-varying epidemic growth rate. And lastly, using arguments on variance-stabilizing transformations, it can be shown that under the assumption of a quadratic mean-variance relationship, the logarithmic transformation leads to expected CRPS values which are independent of the order of magnitude of the predicted quantity. Applying the log transformation to data and forecasts from the European COVID-19 Forecast Hub, we find that it changes model rankings regardless of stratification by forecast date, location or target types. Situations in which models missed the beginning of upward swings are more strongly emphasized while failing to predict a downturn following a peak is less severely penalized. We conclude that appropriate transformations, of which the natural logarithm is only one particularly attractive option, should be considered when assessing the performance of different models in the context of infectious disease incidence.Competing Interest StatementThe authors have declared no competing interest.Funding StatementNIB received funding from the Health Protection Research Unit (grant code NIHR200908). SA's work was funded by the Wellcome Trust (grant: 210758/Z/18/Z). AC's acknowledges funding by the NIHR, the Sergei Brin foundation, USAID, and support by the Academy of Medical Sciences Springboard scheme, funded by the AMS, Wellcome Trust, BEIS, the British Heart Foundation and Diabetes UK (REF:SBF005\1044). EvL acknowledges funding by the National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) in Modelling and Health Economics (grant number NIHR200908) and the European Union's Horizon 2020 research and innovation programme - project EpiPose (101003688). The work of JB was supported by the Helmholtz Foundation (https://www.helmholtz.de/) via the SIMCARD Information and Data Science Pilot Project. SF's work was supported by the Wellcome Trust (grant: 210758/Z/18/Z). This work was supported by the National Institute for Health and Care Research (NIHR) Health Protection Research Unit (HPRU) in Modelling and Health Economics, which is a partnership between the UK Health Security Agency (UKHSA), Imperial College London, and the London School of Hygiene & Tropical Medicine (grant code NIHR200908). The views expressed are those of the authors and not necessarily those of the UK Department of Health and Social Care (DHSC), NIHR, or UKHSA.Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesI confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesAll data and code are available online at https://github.com/epiforecasts/transformation-forecast-evaluation