Exploratory Analysis of Disease Data

1
Exploratory Analysis of Disease Data
Introduction to UNCs GIS Reference Library

• Prepared originally by Kristen Hampton
• Updated and maintained by Ben Allshouse

2
Types of Investigations

• Disease mapping
• Summation of spatial and spatio-temporal
  variation in disease risk
• Geographical Correlation Analysis
• Relationship between heath outcomes and
  environmental risk factors
• Disease clustering
• Raised incidence of disease in an area
• Non-random spatial pattern of disease relative to
  non-cases

3
Disease Mapping
(Zucker, J.R. Changing Patterns of Autochthonous
Malaria Transmission in the United States A
Review of Recent Outbreaks. Emerging Infectious
Diseases. 19962(1)37-43.)
4
Geographical Correlation Analysis
(Glass, G.E. et al. Infectious Disease
Epidemiology and GIS A Case Study of Lyme
Disease. Geo Info Systems. 19922(10)65-69.)
5
Disease ClusteringGCA Point Source Study
John Snows 1854 Map of Cholera and the Broad
Street Pump

• Hypothesis of disease etiology
• Layers
• Roads
• Pumps
• Case event data

(McLeod, K.S. Our sense of Snow the myth of John
Snow in medical geography. Social Science
Medicine. 200050923-935.)
6
ID Application of GIS

• Continuum of diseases
• Highly applicable to environmental diseases,
  such as malaria and other vector-borne diseases
• Limited applicability to non-communicable
  diseases, such as MS, with weak or non-existent
  environmental components
• In between are HIV and other STDs with moderately
  strong links to the environment

(Tanser, F.C. et al. The application of
geographical information systems to important
public health problems in Africa. Int. J Health
Geographics. 200214.)
7
I-95 Syphilis Study
(Cook et al. Whats driving an epidemic? The
Spread of Syphilis Along an Interstate Highway in
Rural North Carolina. Am. J of Public Health.
1999 Mar89(3)369-373.)
8
I-95 Syphilis Study
(Cook et al. Whats driving an epidemic? The
Spread of Syphilis Along an Interstate Highway in
Rural North Carolina. Am. J of Public Health.
1999 Mar89(3)369-373.)
9
Issues

• Confounding
• Resolution / Aggregation level
• Patient confidentiality
• Spatial accuracy
• Ecological Fallacy (individual v. group)
• Temporal relationship between exposure and
  disease

10
Exploratory Analysis of HIV in North Carolina

• Screening and Tracing for Acute HIV Transmission
  (STAT)
• Collaboration between UNC-CH CFAR and NC DHHS
• Developed a system using HIV biomarkers to
  determine the time of transmission (incident
  cases)
• All HIV test samples sent to state labs are
  staged with the STAT protocol
• Records are mapped according to self-reported zip
  code of residence
• Sample data for 1 year
• 110,000 mappable records
• 500 prevalent cases
• 100 incident cases

11
Goal

• Given a table of patient records with disease X
  in North Carolina, create choropleth maps of
• i. case count,
• ii. density of cases per square mile, and
• iii. population density (persons per square
  mile)
• by zip code, shown with major highways.

12
Road Classification
Major Highway
No. of Cases
0
1
2
3
4 - 9
13
Information Needed

• Tabular Disease Data (Given)
• .dbf (dbase) or .txt (text) file formats
• GIS Data
• Polygon shapefile of NC zip codes
• Line shapefile of NC roads
• Attribute Data
• Zip code area and population
• All disease data introduced in this exercise
  is fictional and for demonstration purposes only.

14
Resources

• http//www.lib.unc.edu/reference/gis/
• Amanda Henley
• GIS Reference Librarian, Davis Library
• ahenley_at_refstaff.lib.unc.edu
• ArcGIS desktop Help

15
Data Confidentiality

• Health data is confidential
• Issue of privacy for health data
• As a researcher you must seek training about, and
  proactively protect data confidentiality
• Environmental data is confidential as well
• Privacy issue also applies to any environmental
  data that can be associated with an individual
• How will you protect data confidentiality in this
  course?
• Identify data that might be confidential
  (address, etc.)
• Take action to protect it (do not send it to
  others, do not discuss it or use it in a way that
  would allow others to identify/target
  individuals, erase the data after you use it,
  always ask yourself what else you can do)

16
Exercise

• Handout
• Searching UNCs GIS Database
• Additional Exercises
• Adding X,Y Coordinates of Polygon Centroids
• Creating Raster Density Maps from Point Layers
• Geocoding a Table of Addresses