Refining epidemiological forecasts with simple scoring rules

@article{Moore2021RefiningEF,
  title={Refining epidemiological forecasts with simple scoring rules},
  author={Robert E. Moore and Conor Rosato and Simon Maskell},
  journal={Philosophical transactions. Series A, Mathematical, physical, and engineering sciences},
  year={2021},
  volume={380}
}
Estimates from infectious disease models have constituted a significant part of the scientific evidence used to inform the response to the COVID-19 pandemic in the UK. These estimates can vary strikingly in their bias and variability. Epidemiological forecasts should be consistent with the observations that eventually materialize. We use simple scoring rules to refine the forecasts of a novel statistical model for multisource COVID-19 surveillance data by tuning its smoothness hyperparameter… 

Figures and Tables from this paper

Inference of Stochastic Disease Transmission Models Using Particle-MCMC and a Gradient Based Proposal

Full Bayesian parameter estimations are obtained using gradient information and the No U-Turn Sampler (NUTS) when proposing new parameters of stochastic non-linear Susceptible-Exposed-Infected-Recovered (SEIR) and SIR models.

Technical challenges of modelling real-life epidemics and examples of overcoming these

The coronavirus disease 2019 (COVID-19) pandemic has highlighted the importance of mathematical modelling in informing and advising policy decision-making and this special issue aims to address some of these challenges alongside showcasing the usefulness of modelling as applied in this pandemic.

References

SHOWING 1-10 OF 37 REFERENCES

Fitting to the UK COVID-19 outbreak, short-term forecasts and estimating the reproductive number

This work presents in some detail the inference scheme employed for calibrating the Warwick COVID-19 model to the available public health data streams, and performs computational simulations to assess how the accuracy of short-term predictions varied over the time course of the outbreak.

Short-term forecasts to inform the response to the COVID-19 epidemic in the UK

Ensembles of multi-model forecasts can inform the policy response to the Covid-19 pandemic by assessing future resource needs and expected population impact of morbidity and mortality.

Predictive Model Assessment for Count Data

Proposals include a nonrandomized version of the probability integral transform, marginal calibration diagrams, and proper scoring rules, such as the predictive deviance, for the evaluation of probabilistic forecasts and the critique of statistical models for count data.

Statistical methods used to combine the effective reproduction number, R(t), and other related measures of COVID-19 in the UK

Combining estimates from multiple candidate models to better understand the variation of these outcome measures allows an overall assessment of the current response measures as well as an assessment of their uncertainty, allowing timely and informed decision making based on all of the available information.

Estimating the time-varying reproduction number of SARS-CoV-2 using national and subnational case counts

This decision-support tool can be used to assess changes in virus transmission both globally, regionally, nationally, and subnationally and allows public health officials and policymakers to track the progress of the outbreak in near real-time using an epidemiologically valid measure.

Simulation-Driven COVID-19 Epidemiological Modeling with Social Media

A simulation algorithm is presented that implements a simple agent-based model for disease transmission that works with a standard compartment epidemiological model for COVID-19 and how social media information in the form of daily symptom mentions can be incorporate into CO VID-19 epidemiological models.

Assessing the Accuracy of Time Series Model Forecasts of Count Observations: Comment

In this article, I suggest two approaches for assessing the accuracy of the Harvey-Fernandes model forecasts. First, I suggest using measures of forecast accuracy that were originally developed to

Analysis of temporal trends in potential COVID-19 cases reported through NHS Pathways England

A fully reproducible analysis of temporal trends in NHS Pathways reports until 14th May 2020 is presented, and it is demonstrated that rates of growth/decline and effective reproduction number estimated from these data may be useful in monitoring transmission.

Real-time nowcasting and forecasting of COVID-19 dynamics in England: the first wave

Real-time pandemic monitoring presented here has contributed to the evidence informing this pandemic management throughout the first wave, and continues to be key to quantifying any widespread resurgence, should accrued immunity and effective contact tracing be insufficient to preclude a second wave.

Real-time Nowcasting and Forecasting of COVID-19 Dynamics in England: the first wave?

Real-time pandemic monitoring presented here has contributed to the evidence informing this pandemic management, and will be key to quantifying any widespread resurgence should accrued immunity and effective contact tracing be insufficient to preclude a second wave.