AGENT-BASED MODELING OF THE SPREAD OF THE 1918-1919 SPANISH FLU

[Sally Furniss]

AGENT-BASED MODELING OF THE SPREAD OF THE 1918-1919 SPANISH FLU
IN THREE CANADIAN FUR TRADING COMMUNITIES

http://edt.missouri.edu/Winter2006/Thesis/AhillenC-042806-T5077/research.pdf



CHAPTER
1. INTRODUCTION.???????????????????? 1

2. INFLUENZA AND THE 1918-1919 INFLUENZA PANDEMIC...... 5
Influenza and the Influenza virus
The Making of a Pandemic
The 1918-1919 Influenza Pandemic
The Identity of Influenza

3. THE STUDY COMMUNITIES AND THE 1918-1919 FLU
PANDEMIC?.................................................................................. 18
Life in Central Manitoba
The Influence of the Fur Trade
The Study Communities
The 1918-1919 Flu Comes to the Canadian Subarctic

4. EPIDEMIC MODELING AND MODELS OF THE 1918-1919
FLU EPIDEMIC IN THE NORWAY HOUSE REGION OF
CANADA ?????????????????????..... 31
Models and Modeling
Models of the 1918-1919 Flu Epidemic in the Norway House
Region of Canada
iii
TABLE OF CONTENTS (continued)
The Norway House Ordinary Differential Equation
(NHODE) Model
The Norway House Agent-Based Model
Objectives of This Project

5. DESIGN AND CONSTRUCTION OF THE NHOHGL MODEL?... 44
Design of the NHOHGL Model
The NHOHGL Model Program
Running the NHOHGL Model
The Input Parameters
The Model Output
Advantages of the New NHOHGL Model

6. RESULTS OF THE NHOHGL TEST SIMULATIONS...????... 78
The Replication Study
Replication Test for the Standard Parameter Scenario
Replication Tests for the Winter and Summer Scenarios
The Sensitivity Analysis
Between Post Move Probabilities
Steps-Per-Day
Camp Stay
Population Size and Geographic Distribution
Proportion of Males, Females, and Children/Elders
iv
TABLE OF CONTENTS (continued)
Number of Family Groups
Probability of Contact Within a Family Group
Probability of Contact Outside of a Family Group
Time of First Infection
Location of First Infection
Disease Probability
Death Probability
Recovery Time
Death Time
Turning Off the Displays

7. INVESTIGATION OF THE RESEARCH QUESTIONS
AND ASSESSMENT OF THE NHOHGL MODEL?.???..? 162

Question 1: What impact might the traditional Cree seasonal
movement and settlement patterns have had on the spread
of the 1918-1919 influenza epidemic in the Norway House
region of Canada?

Corollary to Question 1: Do mobility rates or settlement
patterns play a larger role in the observed differences
between the winter and summer simulations?

Corollary to Question 1: How do the findings of the
NHOHGL model regarding the impact of seasonality
compare to the data from the original Norway House
model?

Question 2: Why did the 1918-1919 influenza epidemic never
reach Oxford House or God?s Lake?
v
TABLE OF CONTENTS (continued)

Corollary to Question 2: What would it take for the
simulated epidemic to routinely reach Oxford House
and God?s Lake?

Corollary to Question 2: Would the flu epidemic have
been more or less likely to spread if it was introduced
at Oxford House or God?s Lake instead of at
Norway House?

Question 3: How do the results of the agent-based simulations
compare to those of Sattenspiel?s mathematical model of the
1918-1919 flu epidemic?

Corollary to Question 3: Is the difference in modeling
technique directly responsible for the observed
differences between the data from the NHOHGL
model and the NHODE model?
Assessment of the NHOHGL Model: Is the model a good
representation of the 1918-1919 flu epidemic in this region
of Canada?

8. DISCUSSION AND CONCLUSIONS...?????????...?.. 194

APPENDIX
A THE NHOHGL MODEL PROGRAM???????????.... 202
B THE EXCEL MACRO??????????????????. 258