research

This research is investigating algorithms for retaining confidentiality and allowing granular-level data analysis. The goal is to develop a novel transformation that incorporates all three aspects of the epidemiological triad. The epidemiological real-world domain consisting of the person, place, and time dimensions will be transformed into an epidemiological research domain, where the triad is “mapped” differently so as to preserve confidentiality, and simultaneously preserve relationships in space and time. The research will focus on the space component; since relative space is “mouldable”, existing only with reference to the spatial entities and processes under consideration, health events can therefore be interpreted as patterns and processes within specific contexts. This relative approach focuses on the health events as the subject matter, with space being measured as relationships between objects. This is an inversion of the absolute approach to space typically used in epidemiological studies as described thus far, where space becomes a fixed, underlying geometry, and therefore the subject matter in which the health events are located.

While it is anticipated that a core set of essential features and relationships will be defined, it is expected that these will both be context and event-specific. The generalisability of these is unknown at this point, and may be assessed by the research. All transformations will be evaluated through a comparison with non-transformed data. The value of the proposed research is that it sets out to overcome the barriers posed by privacy and confidentiality constraints on public health research in many parts of the world. The concepts and overall findings need not be limited to any particular country or health event, and have the potential to promote further research into more complex and comprehensive functional analyses involving the complete epidemiological triad. This research is unique and innovative in that it takes a holistic epidemiological approach whilst building on existing and novel technologies and concepts. The study will also engage public health professionals in Canada and the United Kingdom.

Publications

A Method for Managing Re-Identification Risk from Small Geographic Areas in Canada
(BMC Med Inform Decis Mak, 2010)

Musings on Privacy Issues in Health Research Involving Disaggregate Geographic Data About Individuals (Int J Health Geogr, 2009)

Evaluating Predictors of Geographic Area Population Size Cut-offs (JAMIA, 2009)

The Perceived Impact of Location Privacy
(BMC Public Health, 2008)