5.7 Assignment for Unit 5

Activity: literature review of clustering studies in practice

Many of the cluster detection techniques reviewed in this unit were developed quite recently and are still unavailable in standard GIS software. This activity investigates how far the techniques reviewed in this unit – probability mapping, empirical Bayes estimation, global clustering tests, etc. – are used in practice.

Identify studies relevant to your topic (at least some of which should be in academic literature) that detect spatial or spatio-temporal clusters in human health data. The number of studies that you review will vary depending on the topic that you choose – some topics will be associated with a broader literature than others. The attached Excel spreadsheet contains a table of study characteristics. For each relevant study that you identify in your literature review, enter its characteristics into the Excel spreadsheet (as an example, characteristics have been entered into the spreadsheet for the article by Green et al. below). If you wish, you may modify the layout of the Excel spreadsheet and adapt it to or append your own study characteristics onto this sheet. Be sure to include the location of the study population as part of this spreadsheet.  Either by geocoding or by adding an attribute field (with a count of studies in that country) to a map of country boundaries, produce a map showing the locations of the populations participating in the relevant studies you have identified.  Please include the map in your report.

You can use this assignment to research one of several questions about cluster studies and you should choose a question you find interesting. Examples might be as follows:

• You could investigate cluster studies of a particular disease (e.g. breast cancer) and see how far people investigating that disease have followed the same method and used the same cluster detection techniques (e.g. have the methods people have used got more sophisticated over time?
• You could investigate one particular technique or cluster test (e.g. Cuzick and Edwards’ test) and explore the contexts and diseases where it has been used (e.g. is there a tendency for such studies to concentrate on certain types of disease, to use certain techniques, or to be based in certain parts of the world?)

In a short report of no more than 2,000 words, briefly describe how you searched for and identified the studies that you included in your review. In evaluating these studies, you should draw on your spreadsheet and use this to:

• consider how the nature of the data used affected the design and potentially findings of the study
• critically examine differences in the methods used to identify disease clusters

The key to doing this exercise well is not simply to report what each study found, but to provide your own critical appraisal and commentary on your chosen studies. Given what you have learnt elsewhere in this unit, what would you conclude about the way that cluster detection techniques are used in practice? You may wish to include your Excel spreadsheet or a similar table as part of your report, but you are not required to do so.

Useful starting points for this assignment:

Searching through academic literature can be a time-consuming and sometimes intimidating job. With many academic journals restricting access only to viewers who have paid to view articles, it can sometimes be difficult finding the full details of health GIS studies. However, there are quite a few health GIS articles on the web that anyone can view. Here are a couple of places where you might wish to start off your search for GIS and clustering studies:

Google Scholar: If you are not already familiar with Google Scholar, this is a great, freely available way of looking for academic publications – see the link below.

International Journal of Health Geographics: This is a relatively new on-line journal, which is freely accessible to anyone – see the link below.