COVID-19 inequalities in England: a mathematical modelling study of transmission risk and clinical vulnerability by socioeconomic status (preprint)

BackgroundThe COVID-19 pandemic resulted in major inequalities in infection burden between areas of varying socioeconomic deprivation in many countries, including England. Areas of higher deprivation tend to have a different population structure - generally younger - which can increase viral transmission due to higher contact rates in school-going children and working-age adults. Higher deprivation is also associated with higher presence of chronic comorbidities, which were convincingly demonstrated to be risk factors for severe COVID-19 disease. These two major factors need to be combined to better understand and quantify their relative importance in the observed COVID-19 inequalities. MethodsWe used UK Census data on health status and demography stratified by decile of the Index of Multiple Deprivation (IMD), which is a measure of socioeconomic deprivation. We calculated epidemiological impact using an age-stratified COVID-19 transmission model, which incorporated different contact patterns and clinical health profile by decile. To separate the contribution of each factor, we also considered a scenario where the clinical health profile of all deciles was at the level of the least deprived. We also considered the effectiveness of school closures in each decile. ResultsIn the modelled epidemics in urban areas, the most deprived decile experienced 9% more infections, 13% more clinical cases, and a 97% larger peak clinical size than the least deprived; we found similar inequalities in rural areas. 21% of clinical cases and 16% of deaths in England observed under the model assumptions would not occur if all deciles experienced the clinical health profile of the least deprived decile. We found that more deaths were prevented in more affluent areas during school closures. ConclusionsThis study demonstrates that both clinical and demographic factors synergise to generate health inequalities in COVID-19, that improving the clinical health profile of populations would increase health equity, and that some interventions can increase health inequalities.

The interplay between pre-season population susceptibility and the effectiveness of the rolled out vaccination control the timing and size of an emerging seasonal influenza wave in England (preprint)

Relaxing social distancing measures and reduced level of influenza over the last two seasons may lead to a winter 2022 influenza wave in England. We used an established model for influenza transmission and vaccination to evaluate the rolled out influenza immunisation programme over October to December 2022. Specifically, we explored how the interplay between pre-season population susceptibility and influenza vaccine efficacy control the timing and the size of a possible winter influenza wave. Our findings suggest that susceptibility affects the timing and the height of a potential influenza wave, with higher susceptibility leading to an earlier and larger influenza wave while vaccine efficacy controls the size of the peak of the influenza wave. With pre-season susceptibility higher than pre-COVID-19 levels, under the planned vaccine programme an early influenza epidemic wave is possible, its size dependent on vaccine effectiveness against the circulating strain. If pre-season susceptibility is low and similar to pre-COVID levels, the planned influenza vaccine programme with an effective vaccine could largely suppress a winter 2022 influenza outbreak in England.

Evaluating Forecasts with scoringutils in R (preprint)

Evaluating forecasts is essential in order to understand and improve forecasting and make forecasts useful to decision-makers. Much theoretical work has been done on the development of proper scoring rules and other scoring metrics that can help evaluate forecasts. In practice, however, conducting a forecast evaluation and comparison of different forecasters remains challenging. In this paper we introduce scoringutils, an R package that aims to greatly facilitate this process. It is especially geared towards comparing multiple forecasters, regardless of how forecasts were created, and visualising results. The package is able to handle missing forecasts and is the first R package to offer extensive support for forecasts represented through predictive quantiles, a format used by several collaborative ensemble forecasting efforts. The paper gives a short introduction to forecast evaluation, discusses the metrics implemented in scoringutils and gives guidance on when they are appropriate to use, and illustrates the application of the package using example data of forecasts for COVID-19 cases and deaths submitted to the European Forecast Hub between May and September 2021

Determinants of RSV epidemiology following suppression through pandemic contact restrictions (preprint)

Introduction COVID-19 related non-pharmaceutical interventions (NPIs) led to a suppression of RSV circulation in winter 2020/21 throughout Europe and an off-season resurgence in Summer 2021 in several European countries. We explore how such temporary interruption may shape future RSV epidemiology and what factors drive the associated uncertainty. Methods We developed an age-structured dynamic transmission model to simulate pre-pandemic RSV infections and hospitalisations. We sampled parameters governing RSV seasonality, immunity acquisition and duration of post-infection immunity and retained those simulations that qualitatively fit the UK’s pre-pandemic epidemiology. From Spring 2020 to Summer 2021 we assumed a 50% reduced contact frequency, returning to pre-pandemic levels from mid-May 2021. We simulated transmission forwards until 2023 and evaluated the impact of the sampled parameters on the projected trajectories of RSV hospitalisations. Results Following a lifting of contact restrictions in summer 2021 the model replicated an out-of-season resurgence of RSV. If unmitigated, paediatric RSV hospitalisation incidence in the 2021/22 season was projected to increase by 32% to 67% compared to pre-pandemic levels. The size of the increase depended most on whether infection risk was primarily determined by immunity acquired from previous exposure or general immune maturation. While infants were less affected, the increase in seasonal hospitalisation incidence exceeded 100% in 1-2 year old children and 275% in 2-5 year old children, respectively, in some simulations where immunity from previous exposure dominated. Consequently, the average age of a case increased by 1 to 5 months, most markedly if there was strong immunity acquisition from previous exposure. If immunity to infection was largely determined by age rather than previous exposure, the 2021/22 season started earlier and lasted longer but with a peak incidence lower or similar to pre-pandemic levels. For subsequent seasons, simulations suggested a quick return to pre-pandemic epidemiology, with some slight oscillating behaviour possible depending on the strength of post-exposure immunity. Conclusion COVID-19 mitigation measures stopped RSV circulation in the 2020/21 season and generated immunity debt that will likely lead to a temporary increase in RSV burden in the season following the lifting of restrictions, particularly in 1 to 5 year old children. A more accurate understanding of immunity drivers for RSV is needed to better predict the size of such an increase and plan a potential expansion of pharmaceutical and non-pharmaceutical mitigation measures.

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.

Long-term health-related quality of life in non-hospitalised COVID-19 cases with confirmed SARS-CoV-2 infection in England: Longitudinal analysis and cross-sectional comparison with controls (preprint)

Background This study measured the long-term health-related quality of life of non-hospitalised COVID-19 cases with PCR-confirmed SARS-CoV-2(+) infection using the recommended instrument in England (the EQ-5D). Methods Prospective cohort study of SARS-CoV-2(+) cases aged 12-85 years and followed up for six months from 01 December 2020, with cross-sectional comparison to SARS-CoV-2(-) controls. Main outcomes were loss of quality-adjusted life days (QALDs); physical symptoms; and COVID-19-related private expenditures. We analysed results using multivariable regressions with post-hoc weighting by age and sex, and conditional logistic regressions for the association of each symptom and EQ-5D limitation on cases and controls. Results Of 548 cases (mean age 41.1 years; 61.5% female), 16.8% reported physical symptoms at month 6 (most frequently extreme tiredness, headache, loss of taste and/or smell, and shortness of breath). Cases reported more limitations with doing usual activities than controls. Almost half of cases spent a mean of £18.1 on non-prescription drugs (median: £10.0), and 52.7% missed work or school for a mean of 12 days (median: 10). On average, all cases lost 15.9 (95%-CI: 12.1, 19.7) QALDs, while those reporting symptoms at month 6 lost 34.1 (29.0, 39.2) QALDs. Losses also increased with older age. Cumulatively, the health loss from morbidity contributes at least 21% of the total COVID-19-related disease burden in England. Conclusions One in 6 cases report ongoing symptoms at 6 months, and 10% report prolonged loss of function compared to pre-COVID-19 baselines. A marked health burden was observed among older COVID-19 cases and those with persistent physical symptoms. summary Losses of health-related quality of life in non-hospitalised COVID-19 cases increase by age and for cases with symptoms after 6 months. At a population level, at least 21% of the total COVID-19-related disease burden in England is attributable to morbidity.

How immunity from and interaction with seasonal coronaviruses can shape SARS-CoV-2 epidemiology (preprint)

We hypothesised that cross-protection from seasonal epidemics of human coronaviruses (HCoVs) could have affected SARS-CoV-2 transmission, including generating reduced susceptibility in children. To determine what the pre-pandemic distribution of immunity to HCoVs was, we fitted a mathematical model to 6 years of seasonal coronavirus surveillance data from England and Wales. We estimated a duration of immunity to seasonal HCoVs of 7.3 years (95%CI 6.8 - 7.9) and show that, while cross-protection between HCoV and SARS-CoV-2 may contribute to the age distribution, it is insufficient to explain the age pattern of SARS-CoV-2 infections in the first wave of the pandemic in England and Wales. Projections from our model illustrate how different strengths of cross-protection between circulating coronaviruses could determine the frequency and magnitude of SARS-CoV-2 epidemics over the coming decade, as well as the potential impact of cross-protection on future seasonal coronavirus transmission.

Time use and social mixing during and around festive periods: Potential changes in the age distribution of COVID-19 cases from increased intergenerational interactions (preprint)

RationaleAmid the ongoing coronavirus disease 2019 (COVID-19) pandemic in which many countries have adopted physical distancing measures, tiered restrictions, and episodic "lockdowns," the impact of potentially increased social mixing during festive holidays on the age distribution of new COVID-19 cases remains unclear. ObjectiveWe aimed to gain insights into possible changes in the age distribution of COVID-19 cases in the UK after temporarily increased intergenerational interactions in late December 2020. MethodWe modelled changes in time use and social mixing based on age-stratified contact rates using historical nationally-representative surveys and up-to-date Google mobility data from four weeks before and after the festive period. We explored changes in the effective reproduction number and the age distribution of cases, in four scenarios: (1) "normal": time use and contact patterns as observed historically, (2) "pre-lockdown": patterns as seen before the lockdown in November 2020, (3) "lockdown": patterns restricted as in November 2020, and (4) "festive break": similar to 3 but with social visits over the holiday period as in 1. ResultsAcross ages, the estimated Reff decreases during the festive break in scenarios 1-3 and returns to pre-holiday levels in scenarios 2-3, while remaining relatively stable in scenario 4. Relative incidence is likely to decrease in children aged 0-15 but increase in other ages. Changes in age distribution were large during the holidays, and are likely to start before the holidays for individuals aged 16-24 years in scenarios 1-3. ConclusionsOur modelling findings suggest that increased contacts during the festive period may shift the age distribution of COVID-19 cases from children towards adults. Given that COVID-19-related hospitalisations and deaths rise by age, more intergenerational mixing risks an increased burden in the period following the holidays. HighlightsO_LIHome visits are associated with increased intergenerational mixing. C_LIO_LIThe effective reproduction number is likely to remain stable or even reduce slightly due to a reduction in contacts at work and school. C_LIO_LIRelative incidence is likely to become lower in children, but higher in the C_LIO_LIolder (more vulnerable) age groups around the holiday period, which could lead to increased health care burden. C_LI

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

Background: Short-term forecasts of infectious disease can create situational awareness and inform planning for outbreak response. Here, we report on multi-model forecasts of Covid-19 in the UK that were generated at regular intervals starting at the end of March 2020, in order to monitor expected healthcare utilisation and population impacts in real time. Methods: We evaluated the performance of individual model forecasts generated between 24 March and 14 July 2020, using a variety of metrics including the weighted interval score as well as metrics that assess the calibration, sharpness, bias and absolute error of forecasts separately. We further combined the predictions from individual models to ensemble forecasts using a simple mean as well as a quantile regression average that aimed to maximise performance. We further compared model performance to a null model of no change. Results: In most cases, individual models performed better than the null model, and ensembles models were well calibrated and performed comparatively to the best individual models. The quantile regression average did not noticeably outperform the mean ensemble. Conclusions: 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.

Mental Health Status of Adolescents During the COVID-19 Pandemic: A Cross-sectional Survey among the Bangladeshi Graduate Students at Dhaka City (preprint)

ObjectivesTo identify the level of Mental Health Status of Adolescents During the COVID-19 Pandemic among the Bangladeshi Graduate Student at Dhaka MethodA cross-sectional survey was conducted with 330 students from different public and Private Universities in Dhaka, Bangladesh between April 01, 2020 and July 31, 2020 amid the COVID-19 lockdown period in Bangladesh. A standard, self-administered online questionnaire consisting of questions on socio-demographic variables, mental health status, as well as stress management sent to the respondents through social networking platforms. Data were analyzed using descriptive statistics, t-test, one-way ANOVA and correlation tests. ResultsThe mean score of mental health status was 2.08 based on four points scale. They felt problem in decision making (3.04), in doing the things well (2.92), in enjoying normal day to day life (2.88), in playing a useful part in life (2.85), in doing their task (2.75), living in perfectly well and in good health (2.70). The respondents also developed a suicidal tendency (2.55), felt nervous in strung-up (2.24), took longer time to do things (2.14), felt tightness and pressure in head (2.12), and found themselves pressurized by various stuff (2.05). This study also found a significant positive relationship between mental health status and age, living with parents, and parents attitude. Finally, this study revealed that the respondents managed their stress by chatting with their friends, parents and siblings, and by sleeping. ConclusionMental health status of adolescents was found moderate in this study. This study suggests further large-scale study including different socio-economic settings in order to figure out the real scenario of adolescents mental health status of the country during the pandemic.

A retrospective cluster analysis of COVID-19 cases by county (preprint)

The COVID-19 pandemic in the U.S. has exhibited distinct waves, the first beginning in March 2020, the second beginning in early June, and additional waves currently emerging. Paradoxically, almost no county has exhibited this multi-wave pattern. We aim to answer three research questions: (1) How many distinct clusters of counties exhibit similar COVID-19 patterns in the time-series of daily confirmed cases?; (2) What is the geographic distribution of the counties within each cluster? and (3) Are county-level demographic, socioeconomic and political variables associated with the COVID-19 case patterns? We analyzed data from counties in the U.S. from March 1 to October 24, 2020. Time series clustering identified clusters in the daily confirmed cases of COVID-19. An explanatory model was used to identify demographic, socioeconomic and political variables associated the cluster patterns. Four patterns were identified from the timing of the outbreaks including counties experiencing a spring, an early summer, a late summer, and a fall outbreak. Several county-level demographic, socioeconomic, and political variables showed significant associations with the identified clusters. The timing of the outbreak is related both to the geographic location within the U.S. and several variables including age, poverty distribution, and political association. These results show that the reported pattern of cases in the U.S. is observed through aggregation of the COVID-19 cases, suggesting that local trends may be more informative. The timing of the outbreak varies by county, and is associated with important demographic, socioeconomic and geographic factors.

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

England has been heavily affected by the SARS-CoV-2 pandemic, with severe 'lock-down' mitigation measures now gradually being lifted. The real-time pandemic monitoring presented here has contributed to the evidence informing this pandemic management. Estimates on the 10th May showed lock-down had reduced transmission by 75%, the reproduction number falling from 2.6 to 0.61. This regionally-varying impact was largest in London of 81% (95% CrI: 77%-84%). Reproduction numbers have since slowly increased, and on 19th June the probability that the epidemic is growing was greater than 50% in two regions, South West and London. An estimated 8% of the population had been infected, with a higher proportion in London (17%). The infection-to-fatality ratio is 1.1% (0.9%-1.4%) overall but 17% (14%-22%) among the over-75s. This ongoing work will be key to quantifying any widespread resurgence should accrued immunity and effective contact tracing be insufficient to preclude a second wave.

Augmenting contact matrices with time-use data for fine-grained intervention modelling of disease dynamics: A modelling analysis (preprint)

Background: Social distancing is an important public health intervention to reduce or interrupt the sustained community transmission of emerging infectious pathogens, such as SARS-CoV-2 during the coronavirus disease 2019 (COVID-19) pandemic. We aimed to explore the impact on the epidemic curve of fewer contacts when individuals reduce the time they spend on selected daily activities. Methods: We combined the large-scale empirical data of a social contact survey and a time-use survey to estimate contact matrices by age group (0-15, 16-24, 25-44, 45-64, 65+) and daily activity (work, schooling, transportation, and four leisure activities: social visits, bar/cafe/restaurant visits, park visits, and non-essential shopping). We assumed that reductions in time are proportional to reductions in contacts. The derived matrices were then applied in an age-structured dynamic-transmission model of COVID-19 to explore the effects. Findings: The relative reductions in the derived contact matrices were highest when closing schools (in ages 0-14 years), workplaces (15-64 years), and stopping social visits (65+ years). For COVID-19, the closure of workplaces, schools, and stopping social visits had the largest impact on reducing the epidemic curve and delaying its peak, while the predicted impact of fewer contacts in parks, bars/cafes/restaurants, and non-essential shopping were minimal. Interpretation: We successfully augmented contact matrices with time-use data to predict the highest impact of social distancing measures from reduced contacts when spending less time at work, school, and on social visits. Although the predicted impact from other leisure activities with potential for close physical contact were minimal, changes in mixing patterns and time-use immediately after re-allowing social activities may pose increased short-term transmission risks, especially in potentially crowded environments indoors.