Modeling and simulation for the second wave of COVID-19 in Pakistan

Purpose: Owing to an unexpected rate of COVID-19 spread around the globe and to have safer health in the future, it is crucial to predict exposure from different useful predictive modeling techniques. We modeled cumulative and daily (diagnosed and deaths) cases in Pakistan by using two modeling frameworks. Two models GLEAMviz and FB-Prophet were employed. The estimated results were compared with actual data and then evaluated the models on evaluation metrics. Methods: A modified susceptible-exposed-infected-recovered (SEIR) model was implemented in GLEAMviz to simulate the spread of COVID-19 in seven different compartments. The following parameters were used: reproductive number (Ro), detection probability of severe infection, rate of total diagnosed-undiagnosed cases, detection probability of mild infection, and effective contact rate (β). The FB-Prophet model implemented in Python was used to decompose the input time series into four components. The performance was evaluated by the parameters MAE, RMSE, RRSE, and MAPE. Results: We observed a respectable agreement between real and projected values and finally, we estimated peak dates, and decay rates of approximately 95%, and correspondingly diagnosed cumulative and death cases for every province. It was concluded that under the present circumstances, the peak occurs in the second week of December 2020 estimated at around 3000 daily diagnosed cases. According to FB-Prophet’s prediction trend that existed till 30th December 2020 (assuming conditions remain similar), we will have 504,672 cumulative diagnosed cases and 11,329 cumulative deaths, by the end of February 2021. Conclusion: The FB-Prophet model proved to offer better simulations with low errors. It can be used for modeling and estimation at the provincial scale in Pakistan. Our predicting model was in good agreement with the actual data on the comparison. We can safely use the FB-Prophet model for future forecasting of any epidemic. © The Author(s), under exclusive licence to The Brazilian Society of Biomedical Engineering 2024.