Scenarios and Projections for COVID-19 in Arizona

Tim Lant, PhD, MAS; Megan Jehn, PhD; Esma Gel, PhD; Anna Muldoon, MPH; Heather Ross PhD, DNP, ANP-BC

4/22/2020

Arizona State University

DRAFT materials prepared for the

Arizona Department of Health Services – Modeling Working Group

AZ Situation Update: Data and Modeling WG

Impossible to know if contribution to increased cases is due to increased testing or spread of disease.

Change in testing criteria on 3/28 – no longer testing symptomatic.

Increased severity of social-distancing measures 3/16, 3/21, and 3/31.

Estimates for undetected cases in the US are currently around 1 in 11 (9%-14%).

ADHS has assembled this modeling working group to prepare projections for state.

April 1, 2020 Situation Update

• Daily forecasts derived from ADHS and commercial lab testing data

• Monitor testing data and public health interventions as the basis of estimates

AZ Situation Update

• COVID-19 Testing Results for April 8

8-Apr Positive Negative Deaths Total

ADHS 2,726 31,838 80 34,564

COVID tracking 8-Apr 151 1,038 7 1,189

Total 2,726 31,838 80 34,564

Estimating Undetected Cases

Current estimates from J. Shaman (2020) and A. Perkins (2020) that 9% - 14% of infections are detected.

Detected Total

80 50

- 3,587

2,726 14,347

4,071

2,726 22,005

8-Apr-20

Deaths: 1.61%

Ventilator: 6.45%

ICU Bed 7.34%

Total Hospitalized Cases: 16.3%

Total Symptomatic but Unhospitalized Cases: 65.2%

Total Asymptomatic Cases: 18.5%

COVID-19 CLINICAL SEVERITY PYRAMID - Arizona

Credit: Tim Lant, Arizona State University

Epidemiology Signal

• Growth is rapid, but has slowed

• Doubling times • March 17-24: 1.7 days

• March 25-April 8: 5.3 days

• Does not include undetected cases

0

500

1000

1500

2000

2500

3000

3/4

/20

3/5

/20

3/6

/20

3/7

/20

3/8

/20

3/9

/20

3/1

0/20

3/1

1/20

3/1

2/20

3/1

3/20

3/1

4/20

3/1

5/20

3/1

6/20

3/1

7/20

3/1

8/20

3/1

9/20

3/2

0/20

3/2

1/20

3/2

2/20

3/2

3/20

3/2

4/20

3/2

5/20

3/2

6/20

3/2

7/20

3/2

8/20

3/2

9/20

3/3

0/20

3/3

1/20

4/1

/20

4/2

/20

4/3

/20

4/4

/20

4/5

/20

4/6

/20

4/7

/20

4/8

/20

Arizona COVID-19 Testing

Positive Deaths

4/8: Growth is slowing down. 3/16 limited gatherings; 3/20

bars and restraurants close; 3/31 stay-at-home.

4/8: ADHS data show positive confirmations are currently

7% of all tests.

Testing Signal

• Information about negative tests results released March 27. • Positive test result drops to

<10%.

• Range of 8%-10% is consistent with other US cities with community spread

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0

1000

2000

3000

4000

5000

6000

7000

8000

3/10/

20

3/12/

20

3/14/

20

3/16/

20

3/18/

20

3/20/

20

3/22/

20

3/24/

20

3/26/

20

3/28/

20

3/30/

20

4/1/2

0

4/3/2

0

4/5/2

0

4/7/2

0

Rolling weekly positive COVID-19 test rate

Tests Positive %

4/8: Positive test rate has increased to close

to 10% indicating higher burden of COVID-19 disease in population.

Transmission

• Early stochastic effects

• Fast exponential growth

• Slowing growth

0

1

2

3

4

5

6

7

8

9

3/6/2

0

3/7/2

0

3/8/2

0

3/9/2

0

3/10/

20

3/11/

20

3/12/

20

3/13/

20

3/14/

20

3/15/

20

3/16/

20

3/17/

20

3/18/

20

3/19/

20

3/20/

20

3/21/

20

3/22/

20

3/23/

20

3/24/

20

3/25/

20

3/26/

20

3/27/

20

3/28/

20

3/29/

20

3/30/

20

3/31/

20

4/1/2

0

4/2/2

0

4/3/2

0

4/4/2

0

4/5/2

0

4/6/2

0

4/7/2

0

4/8/2

0

AZ confirmed cases - log scale

Early contact based spread. Limited

transmission and contact tracing. Beginnings of community infection.

Undetected disease transmission,

sustained community spread, and cases increase in severity.

Reduced transmissiondue (hopefully) to

social distancing and increased testing.

S E

IA

IP IS

Ihosp D

R

Ihome

2 days 3 days

3 days

6 days

6 days

2 days

0.1=0.45*pD => pD=22%

Results in an overall mortality of 2% among symptomatic individuals

18.5%

20%

ICU

ICU

REC

8 days

6 days

45% 22%

Model1 Details

Transmission Dynamics: Beta_A = 0.55 * Beta_S Beta_P = 0.55 * Beta_S Beta_S = 0.30 Beta_home = 0.20 * Beta_S Beta_hosp = 1 * Beta_S Beta_ICU1 = 0.20 * Beta_S Beta_hosp = 0.20 * Beta_S

Pyramid of Disease Severity • The assumed parameters in the

model are all sourced from recent results

• The top of the pyramid implies significant healthcare resource requirements

TOTAL ASYMPTOMATIC CASES: 18.5%

TOTAL SYMPTOMATIC BUT NONHOSPITALIZED CASES: 65.2%

TOTAL HOSPITALIZED CASES: 16.3%

VENTILATOR: 6.45% IN ICU: 7.34%

DEATHS:1.61%

Assumptions & Parameters

Assumptions & Parameters

Scenarios and projections • We considered five scenarios to provide a range of projections on

• Total number infected – includes asymptomatic and pre-symptomatic

• Total symptomatic patients – includes all patients who are non-hospitalized

• Hospitalized patients – patients in regular hospital beds and ICU

• Patients in ICU

• Patients on a ventilator

Scenario ID 𝜷𝑺

Assumed Total Infected Individuals on 4/8/20

Assumed Summer Effect

Scenario#1 0.30 20,383 On May 15, 𝛽𝑆 reduces to 0.15

Scenario#2 0.25 20,383 On May 15, 𝛽𝑆 reduces to 0.15

Scenario#3 0.30 1,853 On May 15, 𝛽𝑆 reduces to 0.15

Scenario#4 0.25 1,853 On May 15, 𝛽𝑆 reduces to 0.15

Scenario#5 0.30 20,383 On May 15, 𝛽𝑆 reduces to 0.05

Scenarios: Scenario Description

Scenario 1. Assumes all infections are known based on a reporting rate of 9% (18530 initial unreported cases, and 1853 reported cases) and ``moderate” (modeled by transmission rate for symptomatic patients, betaS ~ 0.30) social distancing. The estimate of unreported cases obtained by an estimate provided by Shaman et. al. 2020. Assumes no additional mitigation Summer effect is modeled by reducing betaS by half on May 15.

Scenario 2. Assumes a reporting rate of 9% (18530 initial unreported cases and 1853 reported cases) and ``maximal” social distancing (modeled by transmission rate for symptomatic patients, betaS ~ 0.25). Assumes ongoing mitigation Summer effect is modeled by reducing betaS by half on May 15.

Scenario 3. Assumes that the current reported cases reflect the actual number of infected individuals as of 4/8/20 (1853 initial infected) and moderate social distancing (modeled by transmission rate for symptomatic patients, betaS ~ 0.30). Summer effect is modeled by reducing betaS by half on May 15. Assumes no additional mitigation

Scenario 4. Assumes that the current reported cases reflect the actual number of infected individuals as of 4/8/20 (1853 initial infected) and maximal social distancing (modeled by transmission rate for symptomatic patients, betaS ~ 0.25), Summer effect is modeled by reducing betaS by half on May 15. Assumes no ongoing mitigation

Scenario 5.

Same as Scenario 1 with extreme summer-time transmission effects (heat or distancing); reduced transmission rate, betaS to 0.05 after May 15. Assumes no additional mitigation for social distancing.

Total Infected • Total infected includes asymptomatic and pre-symptomatic individuals, who

may be transmitting the disease

• The sharp decline in Scenario #5 due to the reduction in transmission rate due to summer effect • Assumes May 15 for reduction

In transmission

• Summer effects not yet known

Symptomatic Infections

• A large number of the symptomatic infections will recover at home • Due to social distancing measures, we assumed that these individuals with

transmit the disease at a lower rate

Hospitalized Infections • A portion of the hospitalized infections are in ICU, which we track separately due to the

significant resources need to care for ICU patients

• Under our mid-range scenario (Scenario #2), the number of hospitalized patients hit 13,091 on May 23

• Scenario #4 estimates a max of

1258 patients on May 23, similar to

IHME estimates of 1203

on April 22

Patients in ICU • ICU resources can be critical to save lives

• In particular, several sources have pointed to longer ICU stays by patients that eventually recover

• ICU stays can be as long as 14+ days for these patients

Patients on Ventilator

• A significant fraction of patients (~88%) need mechanical ventilators in ICU

• Rate of mortality among patients on mechanical ventilator is higher than other causes of ARDS (~67%)

Maximum Daily Counts: All scenarios

Low High Mid

Peak Infected 6,875 175,695 88,466

Peak Hopitalized

(Daily) 1,259 31,670 15,428

Peak ICU (Daily) 591 14,981 7,126

Peak Ventilators

(Daily) 520 13,183 6,270

Projected Infections: Low, medium, high

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

4/8/2

0

4/15/20

4/22/20

4/29/20

5/6/2

0

5/13/20

5/20/20

5/27/20

6/3/2

0

6/10/20

6/17/20

6/24/20

7/1/2

0

7/8/2

0

7/15/20

7/22/20

7/29/20

Projected Infections: High, Med, Low(Estimate Made 4/20)

Low Interventions Hi Interventions

current

Projected Hospitalizations: Low, medium, high

0

200

400

600

800

1,000

1,200

1,400

4/8/20 4/15/20 4/22/20 4/29/20 5/6/20 5/13/20 5/20/20 5/27/20 6/3/20 6/10/20 6/17/20 6/24/20 7/1/20 7/8/20 7/15/20 7/22/20 7/29/20

Projected Hospitalizations: High, Med, Low(Estimate Made 4/20)

Low Interventions Hi Interventions Min interventions

Current

Projected ICU visits: Low, medium, high

0

100

200

300

400

500

600

700

800

4/8/2

0

4/15/20

4/22/20

4/29/20

5/6/2

0

5/13/20

5/20/20

5/27/20

6/3/2

0

6/10/20

6/17/20

6/24/20

7/1/2

0

7/8/2

0

7/15/20

7/22/20

7/29/20

Projected ICU visits: High, Med, Low(Estimate Made 4/20)

Low Interventions Hi Interventions

current

Projected Ventilator Use: Low, medium, high

Assumes 88% ventilator utilization for ICU patients

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4/8/20

4/15/20

4/22/20

4/29/20

5/6/20

5/13/20

5/20/20

5/27/20

6/3/20

6/10/20

6/17/20

6/24/20

7/1/20

7/8/20

7/15/20

7/22/20

7/29/20

Projected Ventilation Use : High, Med, Low(Estimate Made 4/20)

Low Interventions Hi Interventions

current

Model Comparison: All scenarios

• Our model predicts infections will peak around the middle of May

• Model is highly-sensitive to social distancing and increased temperature

• A wide range (1-2 order of magnitude) in outcomes is still feasible with uncertainty in undetected cases

1

10

100

1,000

10,000

100,000

1,000,000

15-May 16-May 17-May 18-May 19-May 20-May 21-May 22-May 23-May 24-May 25-May

Forecasts of peak week and peak resources - COVID-19 Arizona

Total Infected Hospital Beds ICU Vents

Recommendations: 1. Adopt a baseline planning scenario

with “low” and “high” excursions.

2. Discuss & reach consensus on importance of predicting peak week.

3. Update forecasts based on new information weekly(?)

4. Prioritize additional analysis 0

100,000

200,000

300,000

400,000

500,000

600,000

700,000

4/2/20 4/9/20 4/16/20 4/23/20 4/30/20 5/7/20 5/14/20 5/21/20 5/28/20 6/4/20 6/11/20 6/18/20 6/25/20 7/2/20 7/9/20 7/16/20 7/23/20 7/30/20

Axi

s Ti

tle

COVID-19 Current Infections (estimate made 4/1/20)