1
GIS and Satellite Imaging of Animal-Related
Human Disease

• Gregory Glass, Ph.D.
• W. Harry Feinstone Department of Molecular
  Microbiology and Immunology
• and
• Department of Epidemiology
• Johns Hopkins Bloomberg School of Public Health

2
Goals

• Role of Environment in Patterns of Infectious
  Diseases
• What are Zoonotic Vector-borne (Animal-Related)
  Diseases
• How do We Combine Environmental Health Data
• Can We Forecast Our Changing Patterns of Risk

3
Infectious Diseases are Environmental Diseases,
too

• Disease system group of interacting populations
  in environment, at least one of which produces a
  phenotype in members of another population that
  is considered pathological

ENVIRONMENT
Human
Pathogen
4
Vector-borne/Zoonoses

• Vector-borne diseases Infectious agents
  transmitted to humans through action of another
  species
• Many vector-borne diseases are transmitted by
  arthropods
• There may be non-human reservoir species
• Up to 75 of EIDs are zoonoses

5
Extent of the Problem I. Zoonotic Diseases and
Rats

• Baltimore, Maryland

6
Human Zoonotic Agents in Wild Norway Rats
Agent Prevalence
Hep - E virus 74
Hantavirus 58
Leptospira interrogans 65
Bartonella elizabethae 34
Calodium hepatica 88
Rickettsia typhi 7
LCMV 0
Easterbrook, et al. 2007a (Epi. Inf. In press)
7
Extent of the Problem II. Policy Implications
Serology x-reacts to R. prowazekii
8
Extent of the Problem Determined by our Ability
to Detect Willingness to Monitor

• Leptospira interrogans previously reportable
  but removed because of few cases observed

Sporadic Urban Leptospirosis Joseph M.
Vinetz, MD Gregory E. Glass, PhD Charles E.
Flexner, MD Paul Mueller, MD and David C.
Kaslow, MD Arch. Int. Med. (1996) 125 794-798
9
How We Relate to Environment Affects our Risk
How do we monitor the environment as related to
VBZD?
10
Reflected Energy in Different Parts EM Spectrum
Tell Us About the Environment
Blue portion EM
Near IR portion EM
Baltimore, MD
11
Lyme Disease
Etiologic agent Borrelia burgdorferi
Vector Ixodes scapularis (Black-legged tick)
Reservoir Peromyscus leucopus (White-footed
mouse)
12
How do we monitor the environment? How do we do
it as the mice see it?

• How do the target species see the environment?
• We infer a model of how perception works e.g.
  Borrelia burgdorferi is associated with forest
  0/1 classifier
• (Dont go near the woods)

13
Satellite Data Merged with Health Data tells us
where People are at Risk

• People in high risk areas 20 x more likely to get
  disease
• High risk for Lyme disease is limited
• Target areas for vaccine, education, tick control

14
Wouldnt it be Nice to Forecast when and where?

• Don't let mosquitoes spoil your outdoor plans.
  Find out how active you can expect mosquitoes to
  be

Enter zip code or city
Go
15
Aedes sollicitans
Eastern salt marsh mosquito Major pest species
along the Atlantic and Gulf coasts of the United
States and Canada Vector of Eastern Equine
Encephalitis, Venezuelan Equine Encephalitis, and
Dirofilaria immitis (Leidy) Tested positive for
West Nile virus and is moderately susceptible
laboratory vector (Turell et al. 2001)
16
Goals

• Characterize responsiveness of Aedes sollicitans
  to environmental factors
• Determine when these climatic factors influence
  mosquito abundances
• Use the relationship between meteorological
  conditions and mosquito populations to forecast
  changes

17
Number of Female Ae. sollicitans
MD Dept Agriculture
18
Jug Bay Wetlands Sanctuary fresh water swamp
flooded woodlands
Parkers Creek Wetlands Sanctuary salt marsh
19
(No Transcript)
20
freshwater swamp
flooded woodlands
salt marsh
21
Statistical Analysis Paradigm
Goal Understand mosquito dynamics as a function
of meteorological conditions.
Data y(s,t) y denotes trap
counts of female mosquitoes s trap
site location t varies daily over
several trapping season years
Site Specific Model y(t)
Pois(?(t)) log(?(t)) ß0
ß1X1(t) . . . ßpXp(t)
22
Daily Abundances Based on Leading Environmental
Conditions
Recent rain Lowest minimum daily temperature Wind
speed Average Relative humidity
23
Used to Predict other Times
24
Used to Predict other Places
25
Zoonotic Example Hantavirus
Samet
Schmaljohn
Samet
ASM
26
Where Are the Mice? How do they see the
Environment?
Low Risk Few P. maniculatus
High Risk Many P. maniculatus
27
Would We Classify These as the Same Type of
Habitat?
28
But the Satellite Sees the Environment Differently
Case site
Control sites
29
Satellite imagery finds the mice
Hantavirus reservoir abundance U.S. Southwest
Trap Success/1000 trap nights
Time (months) 2004
30
(No Transcript)
31
What is the Ecological Mechanism?
Trophic Cascade Hypothesis
Increased Moisture
Vegetative Growth
2-3 litters/year
Terrestrial Arthropod Popn Increases
Increase HPS
Rodent Popn Increases
32
What is the Forecast?

• Numbers of cases HPS dropped off during past 5
  years due to regional drought
• 2004-05 Good snows improved precipitation in
  spring
• Will there be an outbreak?

33
In 2005
34
2006
6-10cases
1 case
35
Evaluation in Chile
Andes virus in O. longicaudatus
Marquet et al 2003
NIH ICIDR
36
Human Cases are in Close Proximity to Areas
Suitable for Infected O. lonigicaudatus
Human Cases
Distance (m)
37
Summary

• Infectious diseases also are environmental
  diseases
• VBZDs are influenced by the environment and how
  people interact with vectors/reservoirs
• Disease outbreaks follow changes in contact
  between people and animals
• Leading environmental conditions can affect
  animal populations
• These relationships can be used to predict
  changing patterns of risk in time to intervene