Skip Ribbon Commands
Skip to main content

COVID-19 Projection Model FAQs


More About Recent COVID-19 Projections for Mecklenburg County

  1. How were the projections developed?
    These models were developed using the below model inputs, based on local experiences and data, and assumptions in the University of Pennslyvania CHIME Model:

    • The Mecklenburg County healthcare facilities deliver 85% of acute care for residents of Mecklenburg, Cabarrus, Iredell, Lincoln, Gaston, and Union counties in North Carolina and York and Lancaster counties in South Carolina (a total population of about 2.3 million).
    • The first COVID-19 patient from this region was hospitalized in Mecklenburg County around 3/11/20.
    • On April 13, there were approximately 90 patients hospitalized in acute care facilities in Mecklenburg County – about 1 in 4 were ventilated.
    • The model assumes that 1.2% of infected patients will be hospitalized with an average length of stay of 8 days. This assumption attempts to account for the underestimates in actual cases due to testing limitations.
    • The model assumes 30% of hospitalized patients will require and intensive care unit stay, with an average of 10 days per stay, and 75% of those will require ventilation for an average of 8 days.

    For more about the modeling approach please visit the Institute for Health Metrics and EvaluationCOVID-19 Model FAQ.

  2. Why do these projections differ from previous projections or projections from other modeling approaches?
    In partnership with Atrium and Novant Health’s data science teams, we tested multiple modeling approaches. We fully recognize that all modeling approaches provide differing results, based on design and underlying assumptions. Commonly cited models prepared for the U.S. and each state by the Institute for Health Metrics and Evaluation at the University of Washington, differ from the CHIME model in that it relies on information about COVID-related deaths instead of incidence, to project future demand. This provides a more conservative estimate and appears to be tracking well with some states. However, this modeling approach is not widely available at the county or regional level and it remains unclear how state-level trajectories align with local trajectories. For example, there are currently twice as many reported cases per capita in Mecklenburg County compared to entire state of North Carolina.

  3. What do the different color curves really mean?
    These projections are an estimate of expected demand on our healthcare system under several scenarios. The scenarios assume varying levels of social distancing – 30%, 45%, 60% are achieved over the model period. These scenarios were chosen based on currently available data from Google Mobility analytics indicating that mobility in the county has declined by 30 to 70%, depending on the type of activity. We did not model 0% or >60% scenarios as there is no evidence that supports abruptly ending all social distancing measures and we know complete social distancing is not realistic, even if only considering people accessing essential services.

  4. How much have we flattened the curve?
    We also provided 14-day projections based on the trajectory of reported cases - which helps us to see how well we are "flattening the curve". Because we do not yet have broad testing capabilities (e.g. testing for everyone regardless of risk), our case counts are underestimates of the true burden of disease and these projections are limited by our testing capabilities. By some estimates – these reported cases may be as little as 5-10% of actual infections.

  5. Why is the peak later now?
    When we "flatten the curve" we undoubtedly extend the peak out further, but it's a much lower peak -- and in fact, may result in the same number of people being infected overall -- over a longer period of time. This allows the county and the healthcare systems to adequately prepare for extended capacity.

  6. Does the Stay at Home Order end before or after the peak?
    The "peak" dates refer to the estimated maximum hospital demand – it informs but does NOT determine the date we implement or suspend interventions, like the Stay at Home Order. Having a high and low bar and tentative dates, allows us to make informed decisions about when and to what degree we should act. These models are volatile and will be updated daily. These are only one type of data informing decisions about the Stay at Home Order.