From a logistical standpoint, we had 24 hours of lecture time, divided into two-hour blocks in which to meet our objectives, along with another 24 hours of lab (again in two-hour blocks). Two hours is enough time in which to present a fairly thorough introduction to an idea, and we decided that we would focus on about 7 basic ideas. These were
Three instructors rotated the lectures, based on individual areas of interest and expertise. A fourth instructor helped with the labs, and offered constructive, independent opinions on course direction. This resulted in a course which (depending on your perspective) was either ``more varied'' or ``less uniform''. We considered diversity, in both approach and presentation style, an advantage.
A good idea of our approach can be gleaned from a brief tour of our lecture topics. We began with an introduction to the central problems of the course, especially the epidemiologically interesting ones. This was followed by a lecture on GIS and Spatial Data handling, as GIS (while not central to the concepts of the course) would be important for the representation of the data discussed in the course. A lecture on Exploratory Spatial Data Analysis (ESDA) followed, in which both tools and strategy were discussed. We changed gears to discuss issues in disease surveillance, then jumped into the heart of Spatial Statistics. The next two lectures (``Disease Clusters'' and ``Designer Statistics'') focused on particular statistics, and illustrated their application using BioMedware software.
We had discussed pattern more than process up to this point in the course, so we began to rectify that with a lecture on Compartmental Models (where transmission of disease - i.e. the process - is modelled). The next topic (Geostatistical Methods and Models) was provided to give the students more expertise in map creation (generating maps from data). This topic also provides an application of spatial autocorrelation models, and illustrates the necessity of understanding spatial autocorrelation.
The final lectures explored the relationship between process and pattern. The first lecture was entitled ``Transmission and Exposure: implications for pattern and process''; this was followed by a guest lecture by Dr. Uriel Kitron which summarized the course very nicely: ``Applications of local and space-time spatial statistics to epidemiological data''. We ended with a lecture exploring the ``Unanswered questions'', featuring ``Flies in the Ointment'' and more on process and pattern.
Student Presentations followed this series of lectures, featuring applications of techniques from the lectures to data sets provided (for the most part) by the students themselves.