SISCOVID

The speaker introduces a project established in response to the pandemic in collaboration with various institutions and researchers, aimed at using systemic models and epidemiological intelligence to provide updated information to decision-makers regarding the crisis in five Colombian cities.

The project team is divided into several cohorts, including data acquisition, modelling, interventions, validation, and communication groups. They aim to provide tools for decision-makers to assess scenarios for mitigating and controlling the pandemic using open data sources and surveys.

They have developed two main models: a compartmental model used to estimate parameters, infection rates, predict mortality, and ICU occupancy, and an agent-based model that incorporates people's mobility in cities to simulate and evaluate interventions' efficacy.

The models have been used to provide summaries to decision-makers, detailing the potential impact of various interventions. A website, masdiscobit.com, was created to share predictions, model information, and results with the public and decision-makers.

Surveys have revealed public perceptions regarding COVID-19, showing high compliance with safety measures like mask-wearing, although worries about public spaces and the economy are prevalent. Analysis indicates strong public adaptation in safety practices.

Findings suggest Colombia has a significant proportion of its population adhering to safety measures like mask-wearing, although economic impacts are notable, with many reporting financial difficulties. Telework has helped reduce contacts, but it's not universally applicable.

Mobility data shows varying adherence to movement restrictions among different cities. Patterns observed correlate with learning processes, indicating that mobility alone does not necessarily lead to increased contacts as people adapt to maintaining precautions.

Proposed interventions include closing high-risk areas, staggered work schedules, and focused testing. An agent-based model simulates different scenarios, showing the impact of targeted interventions on infection rates and hospital occupancy.

The compartmental model, calibrated with data from hospitalizations and ICU use, aids in estimating asymptomatic cases, showing that infections might be underreported. Two waves predicted initially, but real-world data has affected this projection.

Now, with the reopening plans in place, models are adjusted continuously to reflect current trends and interventions' outcomes. The challenge remains in maintaining low infection rates as the population slowly returns to pre-pandemic levels of activity.

Data suggests that a considerable portion of the Colombian population might have already been exposed to the virus, impacting the observed trends and predictions. The focus is on gathering more data to refine these predictions further.