Comparing human and model-based forecasts of COVID-19 in Germany and Poland

1AbstractForecasts based on epidemiological modelling have played an important role in shaping public policy throughout the COVID-19 pandemic. This modelling combines knowledge about infectious disease dynamics with the subjective opinion of the researcher who develops and refines the model and often also adjusts model outputs. Developing a forecast model is difficult, resource- and time-consuming. It is therefore worth asking what modelling is able to add beyond the subjective opinion of the researcher alone. To investigate this, we analysed different real-time forecasts of cases of and deaths from COVID-19 in Germany and Poland over a 1-4 week horizon submitted to the German and Polish Forecast Hub. We compared crowd forecasts elicited from researchers and volunteers, against a) forecasts from two semi-mechanistic models based on common epidemiological assumptions and b) the ensemble of all other models submitted to the Forecast Hub. We found crowd forecasts, despite being overconfident, to outperform all other methods across all forecast horizons when forecasting cases (weighted interval score relative to the Hub ensemble 2 weeks ahead: 0.89). Forecasts based on computational models performed comparably better when predicting deaths (rel. WIS 1.26), suggesting that epidemiological modelling and human judgement can complement each other in important ways.

Numbers of close contacts of individuals infected with SARS-CoV-2 and their association with government intervention strategies

Background Contact tracing is conducted with the primary purpose of interrupting transmission from individuals who are likely to be infectious to others. Secondary analyses of data on the numbers of close contacts of confirmed cases could also: provide an early signal of increases in contact patterns that might precede larger than expected case numbers; evaluate the impact of government interventions on the number of contacts of confirmed cases; or provide data information on contact rates between age cohorts for the purpose of epidemiological modelling. We analysed data from 140,204 close contacts of 39,861 cases in Ireland from 1st May to 1st December 2020. Results Negative binomial regression models highlighted greater numbers of contacts within specific population demographics, after correcting for temporal associations. Separate segmented regression models of the number of cases over time and the average number of contacts per case indicated that a breakpoint indicating a rapid decrease in the number of contacts per case in October 2020 preceded a breakpoint indicating a reduction in the number of cases by 11 days. Conclusions We found that the number of contacts per infected case was overdispersed, the mean varied considerable over time and was temporally associated with government interventions. Analysis of the reported number of contacts per individual in contact tracing data may be a useful early indicator of changes in behaviour in response to, or indeed despite, government restrictions. This study provides useful information for triangulating assumptions regarding the contact mixing rates between different age cohorts for epidemiological modelling.

Caprine Arthritis Encephalitis Virus Disease Modelling Review

Mathematical modelling is used in disease studies to assess the economical impacts of diseases, as well as to better understand the epidemiological dynamics of the biological and environmental factors that are associated with disease spreading. For an incurable disease such as Caprine Arthritis Encephalitis (CAE), this knowledge is extremely valuable. However, the application of modelling techniques to CAE disease studies has not been significantly explored in the literature. The purpose of the present work was to review the published studies, highlighting their scope, strengths and limitations, as well to provide ideas for future modelling approaches for studying CAE disease. The reviewed studies were divided into the following two major themes: Mathematical epidemiological modelling and statistical modelling. Regarding the epidemiological modelling studies, two groups of models have been addressed in the literature: With and without the sexual transmission component. Regarding the statistical modelling studies, the reviewed articles varied on modelling assumptions and goals. These studies modelled the dairy production, the CAE risk factors and the hypothesis of CAE being a risk factor for other diseases. Finally, the present work concludes with further suggestions for modelling studies on CAE.

Literature Review on Epidemiological Modelling, Spatial Modelling and Artificial Intelligence for COVID-19

Introduction: Following the outbreak of Coronavirus (COVID-19) in Wuhan, China in December 2019, the World Health Organisation (WHO) has declared this infectious disease as a pandemic. Unlike previous infectious outbreaks such as Severe Acute Respiratory Syndrome (SARS) and Middle Eastern Respiratory syndrome (MERS), the high transmission rate of COVID-19 has resulted in worldwide spread. The countries with the highest recorded incidence and mortality rates are the US and UK. Rationale/Objective: This review will compare studies that have used epidemiological models for disease forecasting and other models that have identified sociodemographic factors associated with COVID-19. We will evaluate several models, from basic equation-based mathematical models to more advanced machine-learning ones. Our expectation is that by identifying high impact models used by policy makers and discussing their limitations, we can identify possible areas for future research. Evidence Review: The bibliographic database google scholar was used to search keywords such as ‘COVID-19’, ‘epidemiological modelling’ and ‘machine learning’. We examined data review articles, research studies and government-released articles. Results: We identified that the current SEIR model used by the UK government lacked the spatial modelling to enable an accurate prediction of disease spread. We discussed that machine-learning systems which can identify high-risk groups can be used to establish the disparities in COVID-19 death in BAME groups. We found that most of the data hungry AI models used were limited by the lack of datasets available. Conclusion: In conclusion, advances in AI methods for infectious disease have overcome challenges presented in mathematical models. Whilst limitations do exist, when optimised, these highly advanced models have a great potential in public health surveillance, particularly infectious disease transmission.

A participatory modelling approach for investigating the spread of COVID-19 in countries of the Eastern Mediterranean Region to support public health decision-making

Early on in the COVID-19 pandemic, the WHO Eastern Mediterranean Regional Office recognised the importance of epidemiological modelling to forecast the progression of the COVID-19 pandemic to support decisions guiding the implementation of response measures. We established a modelling support team to facilitate the application of epidemiological modelling analyses in the Eastern Mediterranean Region (EMR) countries. Here, we present an innovative, stepwise approach to participatory modelling of the COVID-19 pandemic that engaged decision-makers and public health professionals from countries throughout all stages of the modelling process. Our approach consisted of first identifying the relevant policy questions, collecting country-specific data and interpreting model findings from a decision-maker’s perspective, as well as communicating model uncertainty. We used a simple modelling methodology that was adaptable to the shortage of epidemiological data, and the limited modelling capacity, in our region. We discuss the benefits of using models to produce rapid decision-making guidance for COVID-19 control in the WHO EMR, as well as challenges that we have experienced regarding conveying uncertainty associated with model results, synthesising and comparing results across multiple modelling approaches, and modelling fragile and conflict-affected states.

A participatory modelling approach for investigating the spread of COVID-19 in countries of the Eastern Mediterranean Region to support public health decision-making

Early on in the COVID-19 pandemic, the WHO Eastern Mediterranean Regional Office (WHO EMRO) recognised the importance of epidemiological modelling to forecast the progression of the COVID-19 pandemic to support decisions guiding the implementation of response measures. We established a modelling support team to facilitate the application of epidemiological modelling analyses in the Eastern Mediterranean Region (EMR) countries. Here we present an innovative, stepwise approach to participatory modelling of the COVID-19 pandemic that engaged decision-makers and public health professionals from countries throughout all stages of the modelling process. Our approach consisted of first identifying the relevant policy questions, collecting country-specific data, and interpreting model findings from a decision-maker’s perspective, as well as communicating model uncertainty. We used a simple modelling methodology that was adaptable to the shortage of epidemiological data, and the limited modelling capacity, in our region. We discuss the benefits of using models to produce rapid decision-making guidance for COVID-19 control in the WHO Eastern Mediterranean Region (EMR), as well as challenges that we have experienced regarding conveying uncertainty associated with model results, synthesizing and comparing results across multiple modelling approaches, and modelling fragile and conflict-affected states.

Numbers of close contacts of individuals infected with SARS-CoV-2 and their association with government intervention strategies

ABSTRACTBackgroundContact tracing is conducted with the primary purpose of interrupting transmission from individuals who are likely to be infectious to others. Secondary analyses of data on the numbers of close contacts of confirmed cases could also: provide an early signal of increases in contact patterns that might precede larger than expected case numbers; evaluate the impact of government interventions on the number of contacts of confirmed cases; or provide data information on contact rates between age cohorts for the purpose of epidemiological modelling.MethodsWe analysed data from 140,204 contacts of 39861 cases in Ireland from 1st May to 1st December 2020. Only ‘close’ contacts were included in the analysis. A close contact was defined as any individual who had had > 15 minutes face-to-face (<2 m) contact with a case; any household contact; or any individual sharing a closed space for longer than 2 hours, in any setting.ResultsThe number of contacts per case was overdispersed, the mean varied considerably over time, and was temporally associated with government interventions. Negative binomial regression models highlighted greater numbers of contacts within specific population demographics, after correcting for temporal associations. Separate segmented regression models of the number of cases over time and the average number of contacts per case indicated that a breakpoint indicating a rapid decrease in the number of contacts per case in October 2020 preceded a breakpoint indicating a reduction in the number of cases by 11 days.DiscussionThese data were collected for a specific purpose and therefore any inferences must be made with caution. The data are representative of contact rates of cases, and not of the overall population. However, the data may be a more accurate indicator of the likely degree of onward transmission than might be the case if a random sample of the population were taken. Furthermore, since we analysed only the number of close contacts, the total number of contacts per case would have been higher. Nevertheless, this analysis provides useful information for monitoring the impact of government interventions on the number of contacts; for helping pre-empt increases or decreases in case numbers, and for triangulating assumptions regarding the contact mixing rates between different age cohorts for epidemiological modelling.