COVID-19 Predictive Analytics April 8th, 2020 Predictive Analytics - - PowerPoint PPT Presentation

COVID-19

Predictive Analytics

April 8th, 2020

Case Trajectories

Various projection scenarios are provided to inform COVID-19 management and health system planning. Goals: 1) Assess provincial and regional trends of COVID-19. 2) Determine whether public health measures have had an impact on flattening the curve. 3) Forecast health system demand and compare to capacity.

Predictive Analytics Focus Areas

Clinical Health Consequences Health System Capacity Forecasting

2

Focus of Today’s Presentation

3

• Current Situation
• Compare the trajectory of COVID-19 cases in NL to other provinces
• Examine local trends and impact of ongoing public health measures
• Discuss characteristics of the NL population affected with COVID-19
• Future projections for COVID-19
• Discuss challenges and limitations in projecting cases
• Display short term needs for acute and critical care beds
• Present projections results from two possible scenarios

4

Current Situation

Understanding the Data

COVID-19 Positive Cases by Province (Total Number of Cases / 100K population

5

6

Cumulative NL Cases (All, Funeral Home, Other)

7

0.0 5.0 10.0 15.0 20.0 25.0 30.0 Under 20 20-44 45-54 55-64 65-74 75-84 85+

Percent Age

Age Distribution for COVID-19 Population, NL

8 10 20 30 40 50 60

Under 20 20-44 45-54 55-64 65-74 75-84 85+

Percent Age

Age Distribution for COVID-19 Population, by Hospitalization Status, NL

Hospitalized ICU Non-hospitalized

9 10 20 30 40 50 60 70 80 90

Heart Failure Hypertension IHD AMI Asthma COPD Stroke Diabetes

Percent Additional illness

COVID-19 Population with Selected Comorbid Illnesses, by Hospitalization Status, NL

Hospitalized ICU only Non-hospitalized

10

Forecasting

Projecting COVID-19 Cases

• Modelling
• Conceptually helpful
• Numerous underlying assumptions, frequently misleading
• Challenges in NL - Cluster effect, small number of deaths
• Our Approach
• Let the data drive the model
• Short term projections
• Look for consistency with other models
• Refine model frequently based on the data

11

Current and “worst case” scenarios for acute care hospitalizations till April 30th

12

Current and “worst case” scenarios for critical care beds till April 30th

13

Health Care System Capacity for Province

(Scenario 1 - 32% of NL Infected with COVID-19)

14

Health Care System Capacity for Province

(Scenario 2 – 51 % of NL Infected with COVID-19)

15

Summary

• Public Health Measures have clearly had a significant impact in

flattening the curve

• Need to continue and further improve our efforts to achieve the best

possible outcome

• Future outcome for COVID-19 is highly dependent on the collective

effect of our individual actions to stop its spread.

16

Thank You