Thread by @CamThompsonCBS6

Virginia administration and UVA officials about to hold a tele-news conference on modeling for #COVID19/#coronavirus in the state.

@JakeBurnsCBS6 will be covering for @CBS6

Health Secretary Dr. Dan Carey kicks off the news conference.

Carey: We hope to have an informative discussion on models we're looking at for COVID-19 response in Virginia. Northam has made it clear every decision we make must be based on data.

Carey: This is not a crystal ball and the indications and predictions from these models change.

Carey: We'll hear from the RAND Corporation, then UVA on their model.

Carter Price, a senior mathematician from the RAND Corporation now speaking.

Price: Our analysis has focused primarily on the UVA, IHME, and CHIME models

Price: We'll continue to refine this work as new models/data become available.

Price now going into what is considered a suitable model:

Price says the models fall into two types of models. Statistical (IHME) and Systems Dynamics (UVA and CHIME):

Price now covering the suitability of the different model types:

Overview of CHIME model from Penn Medicine:

More on the CHIME model:

An overview of the UVA model. Says its use of county data allows for detailed projections.

The IHME model takes the trajectory of confirmed COVID-19 deaths and then projects it forward.

Uses data not just from Virginia, but other locations.

Price's key takeaways.

CHIME model may push the peak too far in the future, IHME may have peak too soon.

UVA/IHME models should get more accurate w/ more data.

Now two people from the UVA Biocomplexity Institute are presenting. Says 85 people have been tracking the virus since January.

Bryan Lewis is the primary presenter:

Overview of their presentation, will include a projection of infections through the summer and consider some "what-if" scenario:

Lewis: Key takeaways -- current social distancing is working, Virginia as whole should have enough medical resources for the next couple months, and lifting social distancing too soon could case a second wave.

This is a visualization of the simulation engine that was used:

Lewis going "into the weeds" of how the model configured. Says they assume a 15% detection rate.

Explored five different scenarios.

Overview of the five scenarios:

Lewis: An analogy for the above chart -- unmitigated growth is like if you are pedal to the metal in a car; slowing is like your foot is on the pedal, but not going crazy. Pausing is like you're at cruising speed on cruise control.

How a projection is calibrated:

Lewis: Data showing that people are doing a great job of social distancing.

A slide showing their short-term projections of daily confirmed cases to May 1 on those five different scenario types:

And here's the long-term projections for #COVID19/#coronavirus in Virginia under the five scenarios:

And here's the effect on hospital capacity:

Lewis is done. Carey now speaking.

Carey: Key takeaways. 1 -Social distancing is working. We're in it for the long haul.

Carey: 2. We feel confident on the ability of our hospitals to handle this.

Carey: 3. We have to be cautious on how we ease restrictions and make sure we don't trigger a second wave.

Q&A.
Q - Why did the Virginia specific model take longer than others? What were the challenges?

Lewis: We had to wait and see the impact of social distancing had to make these projections.

Q - Are models incorporating seasonality?
Lewis: The seasonality has been plaguing our community for some time. This novel coronavirus doesn't. Evidence shows there may be some reductions due to weather. Hard to detangle if that's just weather or societal response.

Q - Can you explain your infection range of 3-5 days. Thought people could be infected up to 14?
Lewis: That's covering when people are infectious and participating in spreading the virus/in public. Assumes people are removed from public life (isolating/hospitalized).

Q - How could extensive testing change the projections?
Price: Better estimates for the input value. How many people have it now, less uncertainty in the models. But hopefully, this info would also be used to change policies/behaviors (such as self-isolation).

Q: What can the average Virginian take away from this?
Barrett: No specific guidance, but as this info becomes better understood and we better understands how these actions can move us forward. It will let use better understand what we can do without reigniting the virus.

Carey, closing out: Average Virginian should take away -- we are in this together. This is complex. This is not one and done. We're going to work out what those steps are to get back to normal.

Carey: We have to ease restrictions in the most responsible way possible.

News conference over.