Modelando casos COVID-19 para estimar la oferta de PCR y camas hospitalarias en servicios de salud

Objective: Determine the installed capacity of services to attend the COVID-19 pandemic based on statistical and mathematical techniques. Methods: Describes four phases for modeling: 1. number of COVID cases, 2. use of historical information on the phenomenon, 3. non-linear statistical modeling, 4. mathematical modeling. For the latter, Gompertz was used as a model with which the projections of confirmed cases were made. The main data source was the data provided by the National Institute of Health of Colombia. The parameters to estimate the supply of services were based on the projection of cases and the parameters established according to the INS guidelines. Results: The best fit models for the different stages of the pandemic were: in the exponential, the cubic nonlinear regression model with a coefficient of determination of 99.7%; for the exponential growth and stabilization stages, the Gompertz mathematical model. The combination of models made it possible to predict for August 26 a cumulative of 548,956 confirmed cases for Colombia. These predictions made it possible to establish the supply of services for a target population of six million (6,000,000) inhabitants in 26,226 Polymerase Chain Reaction tests - PCR for the month of August, 225 beds in Intensive Care Units - ICU and 660 hospital beds to that date. Conclusions: The combination of simple and robust statistical and mathematical techniques allowed reliable estimates to project the needs of services related to CRP, hospital beds and ICU beds.