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

Forecasts 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.

The potential health and economic value of SARS-CoV-2 vaccination alongside physical distancing in the UK: transmission model-based future scenario analysis and economic evaluation (preprint)

Background In response to the coronavirus disease 2019 (COVID-19), the UK adopted mandatory physical distancing measures in March 2020. Vaccines against the newly emerged severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may become available as early as late 2020. We explored the health and economic value of introducing SARS-CoV-2 immunisation alongside physical distancing scenarios in the UK. Methods We used an age-structured dynamic-transmission and economic model to explore different scenarios of immunisation programmes over ten years. Assuming vaccines are effective in 5-64 year olds, we compared vaccinating 90% of individuals in this age group to no vaccination. We assumed either vaccine effectiveness of 25% and 1-year protection and 90% re-vaccinated annually, or 75% vaccine effectiveness and 10-year protection and 10% re-vaccinated annually. Natural immunity was assumed to last 45 weeks in the base case. We also explored the additional impact of physical distancing. We considered benefits from disease prevented in terms of quality-adjusted life-years (QALYs), and costs to the healthcare payer versus the national economy. We discounted at 3.5% annually and monetised health impact at 20,000 per QALY to obtain the net monetary value, which we explored in sensitivity analyses. Findings Without vaccination and physical distancing, we estimated 147.9 million COVID-19 cases (95% uncertainty interval: 48.5 million, 198.7 million) and 2.8 million (770,000, 4.2 million) deaths in the UK over ten years. Vaccination with 75% vaccine effectiveness and 10-year protection may stop community transmission entirely for several years, whereas SARS-CoV-2 becomes endemic without highly effective vaccines. Introducing vaccination compared to no vaccination leads to economic gains (positive net monetary value) of 0.37 billion to +1.33 billion across all physical distancing and vaccine effectiveness scenarios from the healthcare perspective, but net monetary values of physical distancing scenarios may be negative from societal perspective if the daily national economy losses are persistent and large. Interpretation Our model findings highlight the substantial health and economic value of introducing SARS-CoV-2 vaccination. Given uncertainty around both characteristics of the eventually licensed vaccines and long-term COVID-19 epidemiology, our study provides early insights about possible future scenarios in a post-vaccination era from an economic and epidemiological perspective.

Projecting contact matrices in 177 geographical regions: an update and comparison with empirical data for the COVID-19 era (preprint)

Mathematical models have played a key role in understanding the spread of directly-transmissible infectious diseases such as Coronavirus Disease 2019 (COVID-19), as well as the effectiveness of public health responses. As the risk of contracting directly-transmitted infections depends on who interacts with whom, mathematical models often use contact matrices to characterise the spread of infectious pathogens. These contact matrices are usually generated from diary-based contact surveys. However, the majority of places in the world do not have representative empirical contact studies, so synthetic contact matrices have been constructed using more widely available setting-specific survey data on household, school, classroom, and workplace composition combined with empirical data on contact patterns in Europe. In 2017, the largest set of synthetic contact matrices to date were published for 152 geographical locations. In this study, we update these matrices with the most recent data and extend our analysis to 177 geographical locations. Due to the observed geographic differences within countries, we also quantify contact patterns in rural and urban settings where data is available. Further, we compare both the 2017 and 2020 synthetic matrices to out-of-sample empirically-constructed contact matrices, and explore the effects of using both the empirical and synthetic contact matrices when modelling physical distancing interventions for the COVID-19 pandemic. We found that the synthetic contact matrices reproduce the main traits of the contact patterns in the empirically-constructed contact matrices. Models parameterised with the empirical and synthetic matrices generated similar findings with few differences observed in age groups where the empirical matrices have missing or aggregated age groups. This finding means that synthetic contact matrices may be used in modelling outbreaks in settings for which empirical studies have yet to be conducted.