Towards a methodological shift in the representation of neighbourhood for vulnerable urban population studies: The example of Montréal injection drugs users

Studies on local neighbourhoods and the health of vulnerable urban populations (e.g., injection drug users, prostitutes, or unemployed adults) most often utilise administrative data and represent neighbourhood as the aggregated characteristics of residents of census tracts or like groupings. These large, administratively fixed boundary areas may be too broad to adequately capture immediate relevant environmental conditions which for specific vulnerable populations can vary at smaller spatial scales. Administratively defined boundaries also present no meaningful basis for coding exposures as relevant area features can be near dwellings yet on the other side of artificial boundaries. Local areas determined using buffer radii centred on an individuals' dwelling place offer a potentially more meaningful way to represent the immediate functional environments of specific vulnerable urban populations. Such individuals often spend large amounts of time in public or private spaces in the vicinity of their dwelling place. As an illustration of this method, we report a recent study on neighbourhood characteristics and injection risk behaviours among injection drug users (IDUs) in Montréal. Neighbourhoods were represented as 500 metre radius buffer zones centred on the IDU's usual dwelling place. The high magnitude of influence of neighbourhood characteristics on injection risk behaviours in Montréal, compared to results from other published studies using different methods, suggests the utility of defining neighbourhood as a circular buffer around the IDU dwelling place. For vulnerable populations such as IDUs, a viable approach to conceptualising local neighbourhood is centring the individual within their immediate local area.