A new accessible adaptable COVID-19 model

ObjectivesSophisticated epidemic models have been created to help governments and large healthcare organisations plan the necessary resources to manage the COVID-19 pandemic. Whilst helpful, current modelling systems are not widely accessible or easily adapted to different populations and circumstances. Our objective was to develop a widely applicable, easily accessible, adaptable model for projecting new COVID-19 infections and deaths that requires minimal expertise or resources to use. The model should be adaptable to different populations and able to accommodate social and pharmaceutical interventions as well as changes in the disease. DesignA Susceptible, Infected and Removed (SIR) infectious disease model was created using widely available Microsoft Excel(C) software. The model is deterministic, generating projections based on the available data and assumptions made. It uses a process of Monitored Forecasting through Visual Matching of predicated vs observed curves to improve accuracy and facilitate adaptability. A review of the COVID-19 literature was performed in order to produce an initial set of adjustable parameters on which to base the output of the model. SettingThis model can be adapted to different regions or countries for which the requisite input data (population size and number of deaths due to the disease) are available. This model has been successfully used with data from England, Sudan and Saudi Arabia. Data from NHS England were used for producing the illustrative results presented here. The model is a generic infectious disease forecast model which may be adapted to other epidemics. InterventionGovernments, public health organisations, pharmaceutical companies and other public institutions may introduce interventions that affect disease transmission or severity. Other unknown factors such as new variants of the infective agent may do the same. The effects of changes in disease transmission are identified by the model when predicted and observed curves deviate. By aligning the curves an evaluation of the effect of the changes can be made. Outcome MeasuresThe model graphically demonstrates projections for daily deaths, cumulative deaths, case mix (asymptomatic, symptomatic and severe infections requiring admission), hospital admissions and bed occupancy (ICU, general medical and total). ResultsThe model successfully produced projections for the outcome measures using NHS England data. Users can adapt and continuously update the model correcting its projections as further local data becomes available. The Microsoft Excel platform allows the model to be used without expensive health information systems or computing infrastructure. ConclusionWe present an SIR epidemic model that projects COVID-19 disease progression, is widely accessible, adaptable to different populations and environments as the disease progresses and is likely to be of benefit for identifying changing population healthcare needs.