Background: Valle del Cauca is the region with the fourth-highest number of COVID-19 cases in Colombia (> 50,000 on September 7 , 2020). Due to the lack of anti-COVID-19 therapies, decision-makers require timely and accurate data to estimate the incidence of disease and the availability of hospital resources to contain the pandemic .
Methods: We adapted an existing model to the local context to forecast COVID-19 incidence and hospital resource use assuming different scenarios: (1) the implementation of quarantine from September 1st to October 15th (average daily growth rate of 2 %); (2-3) partial restrictions (at 4% and 8% growth rates); and (4) no restrictions, assuming a 10% growth rate . Previous scenarios with predictions from June to August were also presented . We estimated the number of new cases, diagnostic tests required, and the number of available hospital and intensive care unit (ICU) beds (with and without ventilators) for each scenario .
Results: We estimated 67,700 cases by October 15th when assuming the implementation of a quarantine , 80,400 and 101,500 cases when assuming partial restrictions at 4% and 8% infection rates, respectively, and 208,500 with no restrictions . According to different scenarios, the estimated demand for reverse transcription-polymerase chain reaction tests ranged from 202,000 to 1,610,600 between September 1st and October 15th . The model predicted depletion of hospital and ICU beds by September 20th if all restrictions were to be lifted and the infection growth rate increased to 10% . Conclusion: Slowly lifting social distancing restrictions and reopening the economy is not expected to result in full resource depletion by October if the daily growth rate is maintained below 8% . Increasing the number of available beds provides a safeguard against slightly higher infection rates . Predictive models can be iteratively used to obtain nuanced predictions to aid decision-making.