UAMS released updated forecasting models for Covid-19 in Arkansas on July 7. Watch the video posted above to hear the high points from the people behind the models. And click here to read the full report.

Several things stood out to me about the updates. First, the mean-case and worst-case estimates for total active infections at our peak decreased by 20% and 25% respectively from the June 19 model. Sounds good, right? Well, even with the updates, we’re still on a path toward 150,000 active cases by the end of October.

Let me clarify what an active case is. Those are the people who are currently sick at any one given time. For the past couple of weeks, Arkansas has hovered around 6,000 active cases. And the UAMS model is projecting 150,000 less than four months from now.

Why the 2,400% increase? This model isn’t just projecting the active cases we know about. It also includes those we know are out there but we aren’t identifying through testing – the mildly symptomatic who might not seek out or qualify for a test but who could be quietly spreading the virus.

To quote the summary: “Note that the model predicts total infections, not confirmed cases, which are likely to be only a small fraction of total infections.”

The forecast includes both short-term and long-term models. While the long-term models decreased somewhat (as noted above), the short term models reflect the rapid growth in cases in late June. The June 19 model forecasted 16,000 cumulative cases by the end of the month. Instead, we surpassed 20,000 cases at the end of June.

From the summary: “The curve now suggests that the epidemic in Arkansas is growing at a much faster rate than the majority of April, May and June.” 

The report is very clear in what forecasters believe to be behind the rapid growth – relaxation of regulations.

“The increase in actual and predicted cases is due to the relaxing of social distancing regulations. Previously, we thought the number of increased cases was associated with increased testing. As shown on the previous page, predicted deaths from Covid have increased off trend as of this reporting period, suggesting a real increase in coronavirus infections and resulting deaths.”

What are these models projecting?

The Time-Series Short-term model, a more conservative model, projects that we will be at 30,000 cumulative cases and 375 deaths by July 12. As of July 7, Arkansas was at 24,512 cumulative cases and 301 deaths.

The eSIR Long-term model gives different scenarios based on various behavior modifications. If we don’t do anything, we’re on track to see 20,000 new cases per day by September 30.

If we mitigate spread to a modest degree (example, some people wear masks), that 20,000 per day drops to 12,000 new cases a day.

If everyone wears a mask when out in public? The new cases drop to 6,000 a day.

The model predicts that cases in Arkansas will peak on October 30, with:

  • 150,000 active cases
  • 2,794 Arkansans hospitalized
  • 838 Arkansans in the ICU
  • 586 Arkansans on a ventilator

The model doesn’t predict the state will exceed its capacity for hospital beds, ICU beds and ventilators – however, that’s considering only Covid-19 patients. It doesn’t include heart attacks, car crashes, or other patients that require ICU beds.

Let’s dig deeper into those numbers. Arkansas currently has a capacity 8,917 hospital beds, and 6,005 were occupied on July 6 by non-Covid patients. If 2,794 Covid patients are hospitalized on top of the current non-Covid number, that would be 8,799 occupied hospital beds – 118 beds shy of full occupancy.

The ICU picture looks even more dire. Arkansas currently has 982 ICU beds, 649 of which were occupied on July 6 by non-Covid patients. If 838 Covid patients were added to the current non-Covid number, that would be 1,487 ICU beds – 505 more beds than state capacity.

Next up, ventilators. The state has 964 ventilators, 250 of which were in use on July 6 by non-Covid patients. If 586 Covid patients were added to the current non-Covid number on ventilators,  that’s 836 ventilators – 128 below the state’s capacity.

From a regional perspective as opposed to a statewide perspective, things look different. I plan to deep dive into capacity by region in a future post, but sufficeth to say that the worst-case scenario projection for Northwest Arkansas has it alone needing 924 ICU beds on Oct. 9. That’s 94% of all the beds available in the state. One thing we can know for sure, Covid cases will not be spread out equally at all times throughout the state.

 

Regional Scenarios

Mean-Case Scenario

Northwest Arkansas

Peak: Oct. 16
Hospitalized: 1,977
ICU: 652
On a ventilator: 457

August 1
Hospitalized: 291
ICU: 96
On a ventilator: 67

Northeast Arkansas

Peak: November 24
Hospitalized: 348
ICU: 115
On a ventilator: 80

August 1
Hospitalized: 30
ICU: 10
On a ventilator: 7

Central Arkansas

Peak: October 21
Hospitalized: 1,426
ICU: 471
On a ventilator: 329

August 1
Hospitalized: 115
ICU: 38
On a ventilator: 27

Southwest Arkansas

Peak: November 24
Hospitalized: 265
ICU: 87
On a ventilator: 61

August 1
Hospitalized: 291
ICU: 96
On a ventilator: 67

Southeast Arkansas

Peak: November 20
Hospitalized: 350
ICU: 115
On a ventilator: 81

August 1
Hospitalized: 29
ICU: 10
On a ventilator: 7

Worst-Case Scenario

Northwest Arkansas

Peak: Oct. 9
Hospitalized: 2,801
ICU: 924
On a ventilator: 647

August 1
Hospitalized: 394
ICU: 130
On a ventilator: 91

Northeast Arkansas

Peak: November 11
Hospitalized: 734
ICU: 242
On a ventilator: 170

August 1
Hospitalized: 48
ICU: 16
On a ventilator: 11

Central Arkansas

Peak: October 13
Hospitalized: 2,412
ICU: 796
On a ventilator: 557

August 1
Hospitalized: 178
ICU: 59
On a ventilator: 41

Southwest Arkansas

Peak: November 11
Hospitalized: 519
ICU: 171
On a ventilator: 120

August 1
Hospitalized: 39
ICU: 13
On a ventilator: 9

Southeast Arkansas

Peak: November 11
Hospitalized: 612
ICU: 202
On a ventilator: 141

August 1
Hospitalized: 44
ICU: 14
On a ventilator: 10

About the models

The forecasts are just that – forecasts. They illuminate the path that we’re on right now. That doesn’t mean the path won’t change as behaviors change or societal events occur. The chart below shows the change in new cases around holidays and school reopening, as well as various scenarios based on behavior.

I have more to post about these models, but I’ll leave it with this for now:
The next 2-3 weeks are predetermined through previous actions and acquired infections. We can’t change that. But we can change next month. We can change October. It’s up to our leaders and us as individuals to determine which path we walk.