Severe Acute Respiratory Syndrome

1
Severe Acute Respiratory Syndrome
SARS

• What can we do to help?
• Ed Fredkin
• CMU -- MIT
• ef_at_cmu.edu fredkin_at_media.mit.edu 4 April 2003

2
The Problem of Time

• This is a problem of exponential growth
• 10 per day, doubles each week, and grows to a
  million times greater in 20 weeks!
• Whatever is done, doing it one day later
• Might not matter
• Might result in a few more deaths
• Might result in 100,000 more deaths
• Its the nature of the problem!

3
Not Well Prepared for This!

• Started on Saturday, 29 March, PM
• Spent a few days to gather digest info
• Current from CDC (www.cdc.gov), WHO (www.who.int)
  and from the Web at large
• Historical info on 1918 Spanish flu -- from the
  Web
• Made an Excel spreadsheet Model
• Decided on a plan of action

4
The CDC (Centers for Disease Control and
Prevention) is the primary US agency. They are
at www.cdc.gov
The Cumulative (by state) number of cases is
posted almost everyday http//www.cdc.gov/od/
oc/media/sars.htm
5
WHO, The World Health Organization (United
Nations) is the international agency most
concerned with the SARS epidemic www.who.int
This is the cumulative number of reported SARS
cases updated almost every day The chart for
April 4th is at www.who.int/csr/sarscountry/2003_0
4_04/en/
6
WHO data
7
Hong Kong is in the Forefront
8
Some See No Problem
9
China had not updated on a daily basis
10
SARS 1 issue after April 2nd!
www.cdc.gov
11
When China didnt update 806 reported cases, the
growth rate was underestimated at 6.6 (because
of the constant 806 cases)
Blue is day to day, red is day to day averaged
from day 1
12
The top group of 3 lines is the raw data for the
previous slide. The top row is the total SARS
cases counting the 806 China data. The row is
the day to day growth and the next row is the
daily growth computed from the first day
(smoothed growth data)
The following is the data for the next slide,
excluding the China data to give a more accurate
measure of growth rate 8.8/day.
13
Excluding the constant 806 data, the actual
growth rate is 8.8
14
Millions, for each week of the epidemic Small
epidemic
Less than 106 deaths
Sick with SARS Fatalities/week
Cumulative Deaths
15
Millions, for each week of the epidemic. Big
epidemic
Sick with SARS Fatalities Cumulative
Deaths
16
Millions, for each week of the epidemic Bigger
epidemic
Sick with SARS Fatalities/week
Cumulative Deaths
17
Statistics from the 1918 Spanish Flu
Epidemic October, 1918, Kansas The solid line
is all deaths Dashed is Flu deaths The doubling
time constant is the same as observed for SARS
One Week!
18
Major Issues

• At some early point there may be no further help
  because the number sick will be too, too many!
• No more room in any hospital
• No rooms in any temporary hospitals
• No available Doctors, nurses or medics
• No life saving equipment
• No medication or other profession care
• The death rate may then increase

19
What to Do?

• Think rationally
• Make good models
• Use them and make them better!
• Using Excel
• Transparent easy to understand and use
• All independent parameters changeable
• Clear and understandable

20
Maximize Utility

• Compare the Utilities of various options
• Show what can be controlled what cant
• Show the benefit of early intervention
• Show the consequences of unnecessary delay
• Show the consequences of indecision
• Show advantages of world-wide cooperation
• Show how lives can be saved or lost

21
Who can best do this?

• Just people who have nothing better to do? No!
• Some top computer programmers others must stop
  doing other important things and work on this!
• It might be a waste of time
• When its obviously important it may be too
  late!
• Maybe it is too late already, maybe its not too
  late.
• The stakes are bigger than anything ever before.

22
We are proceeding!

• This project is underway
• We will make contacts at the CDC and WHO
• So what we do will be as useful as possible
• We will need general and organizational support
  from MIT CMU to proceed swiftly efficiently
• We need to encourage very serious and capable
  workers to implement and deploy the best possible
  tools
• What cannot be tolerated is delay!