Covid-19 SEIDRD Modelling for Pakistan with implementation of seasonality, healthcare capacity and behavioral risk reduction

Saadat, S.; Mansoor, S.; Fahim, A.

Introduction: December 2019 saw the origins of a new Pandemic which would soon spread to the farthest places of the planet. Several efforts of modelling of the geo-temporal transmissibility of the virus have been undertaken, but none describes the incorporation of effect of seasonality, contact density, primary care and ICU bed capacity and behavioural risk reduction measures such as lockdowns into the simulation modeling for Pakistan. We use above variables to create a close to real data curve function for the active cases of covid-19 in Pakistan. Objective: The objective of this study was to create a new computational epidemiological model for Pakistan by implementing symptomatology, healthcare capacity and behavioural risk reduction mathematically to predict of Covid-19 case trends and effects of changes in community characteristics and policy measures. Methods: We used a modified version of SEIR model called SEIDRD (Susceptible - Exposed Latent - Diagnosed as Mild or severe - Recovered - Deaths). This was developed using Vensim PLE software version 8.0. This model also incorporated the seasonal and capacity variables for Pakistan and was adjusted for behavioural risk reduction measures such as lockdowns. Results: The SEIDRD model was able to closely replicate the active covid-19 cases curve function for Pakistan until now. It was able to show that given current trends, though the number of active cases are dropping, if the smart lockdown measures were to end, the cases are expected to show a rise from 28th August 2020 onwards reaching a second peak around 28th September 2020. It was also seen that increasing the ICU bed capacity in Pakistan from 4000 to 40000 will not make a significant difference in active case number. Another simulation for a vaccination schedule of 100000 vaccines per day was created which showed a decrease in covid cases in a slow manner over a period of months rather than days. Conclusion: This study attempts to successfully model the active covid-19 cases curve function of Pakistan and mathematically models the effect of seasonality, contact density, ICU bed availability and Lockdown measures. We were able to show the effectiveness of smart lockdowns and were also to predict that in case of no smart lockdowns, Pakistan can see a rise in active case number starting from 28th of August 2020.