a. General Limitations of GIS in Health Research
There is a wide variety of limitations in regards to GIS and health research. Richards et al. (1999) lists several limitations to GIS technology from a public health perspective. These limitations include:

• The underdeveloped niche for GIS in health applications
  
• Constant need for up-to-date information
  
• Inadequate training and support for users of GIS in the health field
  
• Inadequate technology to protect patient confidentiality
  
• Upgrading demands of GIS software

b. Limitations of this Study
The supplied DAD had no meta-data; therefore we relied on the column headings to assume the context of each field. Limitations to this project include:

1. Within the field ‘MCC’ there were many records that were unclear as to the type of injury sustained by an individual.
   
2. Within the census data from Canada Census Analyzer some of the fields were only based on a 20% sampling.
   
3. Census data was from the 2001 Canada Census as the 2006 data was not yet available.
   
4. DAD postal code data was based on residential addresses, therefore the project focused on burn injuries in residential areas (this limitation stems from that fact that many burn injuries occur in the workplaces and not only in the home) .
   
5. GVRD land use layer (from 2001) was used to clip the burn instances. Since 2001, it is acceptable to assume that zoning in Vancouver has changed and many more residential areas are present.
   
6. Postal code data was not as accurate as expected and points were manually placed into DAs)
   
7. The modifiable area unit problem (MAUP), as it relates to this project, is concerned with point data being represented and included within a polygon location (i.e. burn instances in a DA). Where a point was located on the border between two DAs it could only be included in one.

c. Implications of the Limitations
DAD had no meta-data: reliance on the column headings to assume the context of each field.
Census Analyzer: This could affect the results of the SES values derived for each DA and consequently used for VANDIX calculation.
2001 Census Data: This limits the data used in that demographics across Metropolitan Vancouver could have changed in 5 years.
DAD postal code data based on residential addresses: The project focused only on burn injuries in residential areas.
2001 GVRD land use layer: Records from the DAD were from 2002 to 2006. Many of these records were excluded simply because the land use layer was out of date.
Inaccurate postal code data: The imprecision of the postal code data has left an element of vagueness to the actual geographical location of the instances of burn injuries in the Lower Mainland.
MAUP: It is possible that the MAUP has caused some points to be inaccurately included within some DAs.

Source: http://www.stillsecureafteralltheseyears.com/photos/uncategorized

< >