Epidemiological Predictive Modeling of COVID-19 Infection: Development, Testing, and Implementation on the Population of the Benelux Union

An investigation into the impact of temporality on COVID-19 infection and mortality predictions: new perspective based on Shapley Values

INTRODUCTION

Machine learning models have been employed to predict COVID-19 infections and mortality, but many models were built on training and testing sets from different periods. The purpose of this study is to investigate the impact of temporality, i.e., the temporal gap between training and testing sets, on model performances for predicting COVID-19 infections and mortality. Furthermore, this study seeks to understand the causes of the impact of temporality.

METHODS

This study used a COVID-19 surveillance dataset collected from Brazil in year 2020, 2021 and 2022, and built prediction models for COVID-19 infections and mortality using random forest and logistic regression, with 20 model features. Models were trained and tested based on data from different years and the same year as well, to examine the impact of temporality. To further explain the impact of temporality and its driving factors, Shapley values are employed to quantify individual contributions to model predictions.

RESULTS

For the infection model, we found that the temporal gap had a negative impact on prediction accuracy. On average, the loss in accuracy was 0.0256 for logistic regression and 0.0436 for random forest when there was a temporal gap between the training and testing sets. For the mortality model, the loss in accuracy was 0.0144 for logistic regression and 0.0098 for random forest, which means the impact of temporality was not as strong as in the infection model. Shapley values uncovered the reason behind such differences between the infection and mortality models.

CONCLUSIONS

Our study confirmed the negative impact of temporality on model performance for predicting COVID-19 infections, but it did not find such negative impact of temporality for predicting COVID-19 mortality. Shapley value revealed that there was a fixed set of four features that made predominant contributions for the mortality model across data in three years (2020-2022), while for the infection model there was no such fixed set of features across different years.

How Does Vaccine-Induced Immunity Compare to Infection-Acquired Immunity in the Dynamics of COVID-19?

Five years into the COVID-19 pandemic, the availability of effective vaccines has substantially reduced new cases, hospitalizations, and mortality. However, the waning of immunity has been a topic of particular interest in relation to disease control. The objective of this study is to investigate the impact of the decline in vaccine-induced immunity (ω1) and infection-acquired immunity (ω2) on disease dynamics. For this purpose, we use a compartmental model with seven compartments that accounts for differential morbidity, vaccination, and waning immunity. A compartmental model divides a population into distinct groups depending on their disease status. The temporal changes in the compartments are represented through ordinary differential equations (ODEs). The model is mathematically analyzed to show that a backward bifurcation (i.e., a perverse outcome) may occur when the vaccinated reproduction number (Rv) is equal to unity. Both local and global sensitivity analysis on the reproduction number reveal that the vaccine efficacy, waning of vaccine-induced immunity, vaccine coverage rate, coefficients of transmissibility, and the recovery rate for mild infections are the most sensitive parameters. The global sensitivity analysis on the cumulative number of infections shows that ω1 and ω2 are both pivotal parameters, while ω2 has a higher influence. Simulations on infections and mortality suggest that the changes in ω2 result in dynamics that are more pronounced compared to the dynamics resulting from the changes in ω1, thus indicating the importance of the duration of infection-acquired immunity in disease spread.

Investigating the impact of vaccine hesitancy on an emerging infectious disease: a mathematical and numerical analysis

Throughout the last two centuries, vaccines have been helpful in mitigating numerous epidemic diseases. However, vaccine hesitancy has been identified as a substantial obstacle in healthcare management. We examined the epidemiological dynamics of an emerging infection under vaccination using an SVEIR model with differential morbidity. We mathematically analyzed the model, derived R 0 , and provided a complete analysis of the bifurcation at R 0 = 1 . Sensitivity analysis and numerical simulations were used to quantify the tradeoffs between vaccine efficacy and vaccine hesitancy on reducing the disease burden. Our results indicated that if the percentage of the population hesitant about taking the vaccine is 10%, then a vaccine with 94% efficacy is required to reduce the peak of infections by 40%. If 60% of the population is reluctant about being vaccinated, then even a perfect vaccine will not be able to reduce the peak of infections by 40%.

A generalized multinomial probabilistic model for SARS-COV-2 infection prediction and public health intervention assessment in an indoor environment

SARS-CoV-2 Omicron and its sub-lineages have become the predominant variants globally since early 2022. As of January 2023, over 664 million confirmed cases and over 6.7 million deaths had been reported globally. Current infection models are limited by the need for large datasets or calibration to specific contexts, making them difficult to apply to different settings. This study aims to develop a generalized multinomial probabilistic model of airborne infection to assist public health decision-makers in evaluating the effectiveness of public health interventions (PHIs) across a broad spectrum of scenarios. The proposed model systematically incorporates group characteristics, epidemiology, viral loads, social activities, environmental conditions, and PHIs. Assumptions about social distance and contact duration that estimate infectivity during short-term group gatherings have been made. The study is differentiated from earlier works on probabilistic infection modeling in the following ways: (1) predicting new cases arising from more than one infectious person in a gathering, (2) incorporating additional key infection factors, and (3) evaluating the effectiveness of multiple PHIs on SARS-CoV-2 infection simultaneously. Although the results show that limiting group size has an impact on infection, improving ventilation has a much greater positive health impact. The proposed model is versatile and can flexibly accommodate other scenarios or airborne diseases by modifying the parameters allowing new factors to be added.

Fitting Early Phases of the COVID-19 Outbreak: A Comparison of the Performances of Used Models

The COVID-19 outbreak involved a spread of prediction efforts, especially in the early pandemic phase. A better understanding of the epidemiological implications of the different models seems crucial for tailoring prevention policies. This study aims to explore the concordance and discrepancies in outbreak prediction produced by models implemented and used in the first wave of the epidemic. To evaluate the performance of the model, an analysis was carried out on Italian pandemic data from February 24, 2020. The epidemic models were fitted to data collected at 20, 30, 40, 50, 60, 70, 80, 90, and 98 days (the entire time series). At each time step, we made predictions until May 31, 2020. The Mean Absolute Error (MAE) and the Mean Absolute Percentage Error (MAPE) were calculated. The GAM model is the most suitable parameterization for predicting the number of new cases; exponential or Poisson models help predict the cumulative number of cases. When the goal is to predict the epidemic peak, GAM, ARIMA, or Bayesian models are preferable. However, the prediction of the pandemic peak could be made carefully during the early stages of the epidemic because the forecast is affected by high uncertainty and may very likely produce the wrong results.

Sources, diffusion and prediction in COVID-19 pandemic: lessons learned to face next health emergency

Scholars and experts argue that future pandemics and/or epidemics are inevitable events, and the problem is not whether they will occur, but when a new health emergency will emerge. In this uncertain scenario, one of the most important questions is an accurate prevention, preparedness and prediction for the next pandemic. The main goal of this study is twofold: first, the clarification of sources and factors that may trigger pandemic threats; second, the examination of prediction models of on-going pandemics, showing pros and cons. Results, based on in-depth systematic review, show the vital role of environmental factors in the spread of Coronavirus Disease 2019 (COVID-19), and many limitations of the epidemiologic models of prediction because of the complex interactions between the new viral agent SARS-CoV-2, environment and society that have generated variants and sub-variants with rapid transmission. The insights here are, whenever possible, to clarify these aspects associated with public health in order to provide lessons learned of health policy that may reduce risks of emergence and diffusion of new pandemics having negative societal impact.

Beyond COVID-19: Do biothermodynamic properties allow predicting the future evolution of SARS-CoV-2 variants?

During the COVID-19 pandemic, many statistical and epidemiological studies have been published, trying to predict the future development of the SARS-CoV-2 pandemic. However, it would be beneficial to have a specific, mechanistic biophysical model, based on the driving forces of processes performed during virus-host interactions and fundamental laws of nature, allowing prediction of future evolution of SARS-CoV-2 and other viruses. In this paper, an attempt was made to predict the development of the pandemic, based on biothermodynamic parameters: Gibbs energy of binding and Gibbs energy of growth. Based on analysis of biothermodynamic parameters of various variants of SARS-CoV-2, SARS-CoV and MERS-CoV that appeared during evolution, an attempt was made to predict the future directions of evolution of SARS-CoV-2 and potential occurrence of new strains that could lead to new pandemic waves. Possible new mutations that could appear in the future could lead to changes in chemical composition, biothermodynamic properties (driving forces of new virus strains) and biological properties of SARS CoV-2 that represent a risk for humanity.

Combining and comparing regional SARS-CoV-2 epidemic dynamics in Italy: Bayesian meta-analysis of compartmental models and global sensitivity analysis

During autumn 2020, Italy faced a second important SARS-CoV-2 epidemic wave. We explored the time pattern of the instantaneous reproductive number, R ₀(t), and estimated the prevalence of infections by region from August to December calibrating SIRD models on COVID-19-related deaths, fixing at values from literature Infection Fatality Rate (IFR) and average infection duration. A Global Sensitivity Analysis (GSA) was performed on the regional SIRD models. Then, we used Bayesian meta-analysis and meta-regression to combine and compare the regional results and investigate their heterogeneity. The meta-analytic R ₀(t) curves were similar in the Northern and Central regions, while a less peaked curve was estimated for the South. The maximum R ₀(t) ranged from 2.15 (South) to 2.61 (North) with an increase following school reopening and a decline at the end of October. The predictive performance of the regional models, assessed through cross validation, was good, with a Mean Absolute Percentage Error of 7.2% and 10.9% when considering prediction horizons of 7 and 14 days, respectively. Average temperature, urbanization, characteristics of family medicine and healthcare system, economic dynamism, and use of public transport could partly explain the regional heterogeneity. The GSA indicated the robustness of the regional R ₀(t) curves to different assumptions on IFR. The infectious period turned out to have a key role in determining the model results, but without compromising between-region comparisons.

Analytical investigation of compartmental models and measure for reactions of governments

The ongoing coronavirus disease 2019 (COVID-19) pandemic poses a major threat to the worldwide health care. In this context, epidemic modelling is an integral part of containment strategies. Compartmental models are typically used for this purpose. Analytical solutions of the two distinct but connected Susceptible-Infectious-Recovered-Deceased (SIRD) and Susceptible-Infectious-Quarantine-Recovered (SIQR) models are presented in this study. Furthermore, the behaviour at the start of a disease outbreak is derived. This analysis shows that a combination of transmission, recovery and isolation rates dominates the behaviour at the start of an epidemic. In addition, the loss occurring due to quarantine and lockdown measures is investigated, where it can be observed that quarantine procedures lead to a smaller loss in comparison with lockdown regulations. Within this framework, optimized strategies that lead to a constant epidemic peak or a minimized loss are presented.

Analysis of multi-strain infection of vaccinated and recovered population through epidemic model: Application to COVID-19

In this work, an innovative multi-strain SV EAIR epidemic model is developed for the study of the spread of a multi-strain infectious disease in a population infected by mutations of the disease. The population is assumed to be completely susceptible to n different variants of the disease, and those who are vaccinated and recovered from a specific strain k (k ≤ n) are immune to previous and present strains j = 1, 2, ⋯, k, but can still be infected by newer emerging strains j = k + 1, k + 2, ⋯, n. The model is designed to simulate the emergence and dissemination of viral strains. All the equilibrium points of the system are calculated and the conditions for existence and global stability of these points are investigated and used to answer the question as to whether it is possible for the population to have an endemic with more than one strain. An interesting result that shows that a strain with a reproduction number greater than one can still die out on the long run if a newer emerging strain has a greater reproduction number is verified numerically. The effect of vaccines on the population is also analyzed and a bound for the herd immunity threshold is calculated. The validity of the work done is verified through numerical simulations by applying the proposed model and strategy to analyze the multi-strains of the COVID-19 virus, in particular, the Delta and the Omicron variants, in the United State.