ArticlePDF Available

Development of neighborhoods to measure spatial indicators of health

Authors:

Abstract

The literature on health inequalities demonstrates that where one lives impacts one's health. This report details the development of tools to investigate the spatial relationship between inequalities in neighborhood quality and health disparities. A combination of spatial statistics, geographic information system (GIS) concepts and capabilities, and community consultation provided a novel methodology to define neighborhood units and the context to spatially analyze the relations between neighborhood health indicators and socioeconomic status. Data sets from DMTI Spatial Inc., Statistics Canada, the City of Ottawa, the National Capital Commission, the Ottawa Real Estate Board, as well as QuickBird Satellite imagery, Canada 411 phone calls, corporation web sites, field-based observations, and expert knowledge, were utilized as the base data sets for defining natural neighborhood boundaries and defining and collecting data on indicator variables. These spatial health indicators take into account both the social component and the physical (environmental/contextual) component of the defined neighborhoods. The key to developing this quantitative set of indicators was the definition of neighborhoods in Ottawa. The methodologies established in this research are unique and transferable to similar research endeavors.
... The ecological elements of neighborhoods can be grouped into physical (e.g., air, noise), built (e.g., housing), service (e.g., education), sociocultural (e.g., ethnic composition), and reputational (e.g., perceptions of the area) kinds (Macintyre and Ellaway 2003;Cutchin et al. 2011). These elements can also be expanded into a long list, such as the 44 indicators listed by Parenteau et al. (2008) which were organized into five domains to correspond to Maslow's hierarchy of needs (Maslow 1968). Further understandings of neighborhood effects can be obtained by reviewing the theoretical models that explain the contextual effects on human welfare (e.g., Shinn and Toohey 2003) or by examining classifications of neighborhood effects to distinguish various ways individuals are influenced by contextual/ecological factors (e.g., Dietz 2002, among a huge multi-year literature on neighborhood effects). ...
... All these complications in GIS modeling are relevant in the neighborhood mapping context. For example, many defining attributes of neighborhoods (e.g., Parenteau et al. 2008;Milbrath and DeGuzman 2015) are characterized not as regular GIS fields that are independent existences (such as elevation), but as density fields or field objects: the mean household income, property value, educational level, and crime rate are usually calculated based on census or other administrative areal units of certain scales; proximity to parks is a continuous field, but it can be calculated only when the (discrete) parks are specified. Sliding neighborhoods (see Sect. 2.2), on the other hand, are similar to object fields. ...
... The definition of pre-defined zones did not incorporate sufficient consideration of social processes (Leventhal and Brooks-Gunn 2000;Huie 2001;Parenteau et al. 2008), hence providing boundaries that are arbitrary and meaningless in many social applications (e.g., Sampson et al. 2002). A reasonable opinion, however, may be that census units are for general purposes to map all, instead of a narrow band of sociodemographic themes (e.g., Clapp and Wang 2006). ...
Article
Full-text available
This paper examines complications in neighborhood mapping and corresponding challenges for the GIS community, taking both a conceptual and a methodological perspective. It focuses on the social and spatial dimensions of the neighborhood concept and highlights their relationship in neighborhood mapping. Following a brief summary of neighborhood definitions, five interwoven factors are identified to be origins of neighborhood mapping difficulties: conceptual vagueness, uncertainty of various sources, GIS representation, scale, and neighborhood homogeneity or continuity. Existing neighborhood mapping methods are grouped into six categories to be assessed: perception based, physically based, inference based, preexisting, aggregated, and automated. Mapping practices in various neighborhood-related disciplines and applications are cited as examples to demonstrate how the methods work, as well as how they should be evaluated. A few mapping strategies for the improvement of neighborhood mapping are prescribed from a GIS perspective: documenting simplifications employed in the mapping procedure, addressing uncertainty sources, developing new data solutions, and integrating complementary mapping methods. Incorporation of high-resolution data and introduction of more GIS ideas and methods (such as fuzzy logic) are identified to be future opportunities.
... At the local level, how place is conceptualised in health research is considered as the weakest theoretical aspects of health studies (Matthews, 2008, Parenteau & Sawada, 2011. Apart from the ONS, only a few Canadian studies are based on neighbourhoods as oppose to administrative areas (Parenteau & Sawada, 2011 (Parenteau et al., 2008). ...
... In the second revision of the neighbourhood delimitations, 97 neighbourhoods were defined 9 . To ensure data reliability, the majority of neighbourhoods are comprised of at least 4,000 persons and are formed by an aggregation of disseminations areas or postal codes (Parenteau et al., 2008). ...
... Neighbourhood SES was based on Kristjansson's index of socio-economic disadvantage as described in Parenteau et al. (2008). This SES index consists of the following measures: percent of residents with less than a high school education, percent of single-parent families, percent of unemployed residents, percent of households below the low income cut-off and average household income. ...
... In the ONS, neighbourhoods were defined based on natural boundaries, similarity in SES and demographics, Ottawa Multiple Listing Services maps, and participatory mapping feedback from community members and experts [15]. Objectively measured environmental data were collected from 2006 to 2008 using the following data and methods: 1) 2006 Canadian census household data; 2) GIS data from DMTI Spatial Inc., the City of Ottawa, and the National Capital Commission; 3) telephone contact with businesses; 4) web-based research; 5) team knowledge of local resources; and 6) field research and validation. ...
... Objectively measured environmental data were collected from 2006 to 2008 using the following data and methods: 1) 2006 Canadian census household data; 2) GIS data from DMTI Spatial Inc., the City of Ottawa, and the National Capital Commission; 3) telephone contact with businesses; 4) web-based research; 5) team knowledge of local resources; and 6) field research and validation. A further in-depth description of methods related to the ONS and its variables is available elsewhere [15]. ...
... The REC environment was based on a REC Index that was created using principal components analysis; this served as a measure of the density of facilities for recreation in a neighbourhood [15]. The REC index includes meters of bike and walking paths per person, meters-squared of park space per person, and recreation facilities per thousand people per neighbourhood. ...
Article
Full-text available
Background: There is limited Canadian research examining whether directly measured physical activity (PA) and body mass index (BMI) differ between neighbourhoods with different objectively measured socioeconomic (SES) and recreation (REC) environments. Purpose: To determine whether mean adult PA levels, sedentary time and BMIs were different across four neighbourhoods with contrasting SES and REC environments in Ottawa, Canada. Methods: This study employed a cross-sectional design to collect pilot data of objectively measured height, weight and PA (using accelerometry) and self-reported covariates in 113 adults (≥18 years). Four contrasting neighbourhoods (high REC/high SES, high REC/low SES, low REC/high SES, and low REC/low SES) were selected based on data collected as part of the Ottawa Neighbourhood Study. Analysis of covariance and logistic regression were used to perform neighbourhood comparisons for PA, sedentary time and BMI, adjusting for age, sex and household income and possible interactions. Post-hoc comparisons using Tukey’s test were performed. Results: Significant neighbourhood-group effects were observed for light intensity PA and sedentary time. Post-hoc tests identified that the low REC/high SES neighbourhood had significantly more minutes of light PA than the low REC/low SES (Mdiff = 56.05 minutes·day, Tukey p = 0.01). Unadjusted BMI differed between the four neighbourhoods, but the differences were not significant after controlling for age, sex and household income. Conclusions: This study demonstrates that light PA and sedentary time differ between neighbourhoods of varying REC and SES environments after controlling for differences in age, sex and household income. Findings also suggest that other area-level factors may explain these neighbourhood differences.
... Built and social environment characteristics were collected by the ONS; a large study of neighbourhoods and health outcomes in Ottawa, Canada. Briefly, neighbourhoods were defined based on natural barriers, similarity in socioeconomics and demographics, Ottawa Multiple Listing Services (real estate) maps, and participatory mapping feedback from community members and experts [26]. Most neighbourhoods contained >4000 people. ...
... Objectively measured built environment data were collected from 2006 to 2008 using the following methods: 1) 2006 Canadian census household data; 2) GIS data from DMTI Spatial Inc., the City of Ottawa, and the National Capital Commission (NCC); 3) telephone contact with businesses; 4) web-based research (e.g., Canada 411, websites, Google Maps); 5) team knowledge of local resources; and 6) field research and validation (e.g., car, walking, bicycle). A further in-depth description of methods related to the ONS and its derived variables is available elsewhere [26]. ...
Article
Background In Canada, there is limited research examining the effects of objectively measured neighbourhood environments on physical activity (PA) and obesity. Purpose To determine the relationships between variables from built and social environments and PA and overweight / obesity across 86 Ottawa neighbourhoods. Methods Individual-level data including self-reported leisure-time PA, height and weight were examined using a sample of 4727 adults from four combined cycles (years 2001/2003/2005/2007) of the Canadian Community Health Survey. Data on neighbourhood characteristics were obtained from the Ottawa Neighbourhood Study; a large study of neighbourhoods and health in Ottawa. Binomial multivariate multilevel models were used to examine the relationships of environmental and individual variables with PA and overweight / obesity using population weights. Results Approximately 75% of adults were inactive (<12.5 kJ/kg/day) while half were overweight / obese. Results of the multilevel models suggest that higher numbers of convenience stores and fast food outlets in a neighbourhood were associated with increased odds of being overweight / obese, while a larger number of restaurants was associated with lower odds. Season of data collection was significantly associated with PA in men and women with the odds of PA in winter being half that of summer. Intraclass coefficients were low, and identified that the models explained a small proportion of the neighbourhood-level variance in PA and overweight / obesity. Conclusions Findings from this sample identified that recreation and social environments did not exert significant influences on PA or overweight / obesity, however, food outlets did show a significant influence on female overweight / obesity. The impact of individual-level characteristics to the model was modest.
... The natural neighbourhoods were delineated through a semi-automatic approach with the purpose of being used as the geography of reference for this research as well as for the Ottawa Neighbourhood Study (ONS) project [39]. The objective was to delineate homogeneous units in terms of socioeconomic status (SES), which has been linked to health outcomes, which would also maximize external heterogeneity. ...
... This iterative boundary delineation work was achieved through consultations with the City of Ottawa and leaders of local grassroots organizations and public input. Details on the methodological approach are published elsewhere [39]. The aggregated structure was created through the grouping of census tracts using ArcGIS 9.2 [42] with a simple contiguity constraint. ...
Article
Full-text available
Many Canadian population health studies, including those focusing on the relationship between exposure to air pollution and health, have operationalized neighbourhoods at the census tract scale. At the same time, the conceptualization of place at the local scale is one of the weakest theoretical aspects in health geography. The modifiable areal unit problem (MAUP) raises issues when census tracts are used as neighbourhood proxies, and no other alternate spatial structure is used for sensitivity analysis. In the literature, conclusions on the relationship between NO(2) and health outcomes are divided, and this situation may in part be due to the selection of an inappropriate spatial structure for analysis. Here, we undertake an analysis of NO(2) and respiratory health in Ottawa, Canada using three different spatial structures in order to elucidate the effects that the spatial unit of analysis can have on analytical results. Using three different spatial structures to examine and quantify the relationship between NO(2) and respiratory morbidity, we offer three main conclusions: 1) exploratory spatial analytical methods can serve as an indication of the potential effect of the MAUP; 2) OLS regression results differ significantly using different spatial representations, and this could be a contributing factor to the lack of consensus in studies that focus on the relation between NO(2) and respiratory health at the area-level; and 3) the use of three spatial representations confirms no measured effect of NO(2) exposure on respiratory health in Ottawa. Area units used in population health studies should be delineated so as to represent the a priori scale of the expected scale interaction between neighbourhood processes and health. A thorough understanding of the role of the MAUP in the study of the relationship between NO(2) and respiratory health is necessary for research into disease pathways based on statistical models, and for decision-makers to assess the scale at which interventions will have maximum benefit. In general, more research on the role of spatial representation in health studies is needed.
... Validation was performed on food outlets listed in EPOI files within 12 CTs in Montreal. To our knowledge, no quantified validation study has been devoted to this sub-dataset, beyond minor reports of inconsistencies, missing data, or misclassifications [23]. ...
Article
Full-text available
Background Validation studies of secondary datasets used to characterize neighborhood food businesses generally evaluate how accurately the database represents the true situation on the ground. Depending on the research objectives, the characterization of the business environment may tolerate some inaccuracies (e.g. minor imprecisions in location or errors in business names). Furthermore, if the number of false negatives (FNs) and false positives (FPs) is balanced within a given area, one could argue that the database still provides a “fair” representation of existing resources in this area. Yet, traditional validation measures do not relax matching criteria, and treat FNs and FPs independently. Through the field validation of food businesses found in a Canadian database, this paper proposes alternative criteria for validity. Methods Field validation of the 2010 Enhanced Points of Interest (EPOI) database (DMTI Spatial®) was performed in 2011 in 12 census tracts (CTs) in Montreal, Canada. Some 410 food outlets were extracted from the database and 484 were observed in the field. First, traditional measures of sensitivity and positive predictive value (PPV) accounting for every single mismatch between the field and the database were computed. Second, relaxed measures of sensitivity and PPV that tolerate mismatches in business names or slight imprecisions in location were assessed. A novel measure of representativity that further allows for compensation between FNs and FPs within the same business category and area was proposed. Representativity was computed at CT level as ((TPs +|FPs-FNs|)/(TPs+FNs)), with TPs meaning true positives, and |FPs-FNs| being the absolute value of the difference between the number of FNs and the number of FPs within each outlet category. Results The EPOI database had a "moderate" capacity to detect an outlet present in the field (sensitivity: 54.5%) or to list only the outlets that actually existed in the field (PPV: 64.4%). Relaxed measures of sensitivity and PPV were respectively 65.5% and 77.3%. The representativity of the EPOI database was 77.7%. Conclusions The novel measure of representativity might serve as an alternative to traditional validity measures, and could be more appropriate in certain situations, depending on the nature and scale of the research question.
Article
Full-text available
How can the efforts of philanthropic agencies be better supported to promote citywide community well-being? To address this question, this article uses a citywide, engagement-driven initiative led by local practitioners that took place in Detroit between 2016 and 2019. This article is based on the initiative’s two main outcomes, namely the identification of seven elements of an effective community development system and a vitality framework designed to measure community progress and success. The author of this article conducted participant observation, interviews, and a literature review, as well as site visits to, and case studies on, four best-practice cities, and then used the outcomes to validate the results of the initiative. Informed by the outcomes of the initiative and the research, the article suggests how best to utilize the seven community development system elements and the vitality success framework effectively to support philanthropy that promotes community well-being. This article focuses on theory-building in Detroit and calls for empirical research to further validate the findings. The article provides useful insights into the benefits of community-based research and citywide engagement as essential components of an effective community development system that can coordinate philanthropic practice more effectively to promote community well-being.
Article
The spatial representation of a city is typically formed by top‐down jurisdictional boundaries. A parallel approach would be to consider representing a city based on platial characteristics, that is, a bottom‐up landscape created through individual and collectively derived representations. This study contributes to this discourse through the exploratory examination of the ecology notions of home range and habitat applied to humans in an urban context. Using spatial data collected through a WebGIS platform, we employ a spatial definition of sense of place and social capital to understand the platial nature of the city and, simultaneously, defining home range and habitat as platial notions. We found spatial variability among individual home range and habitat and the difficulty of traditional administrative boundaries to represent these areas. This research defines and presents home range and habitat to partially describe the emergent nature of platial theory and explores their operationalization at the urban level.
Article
Full-text available
Urban health indicator (UHI) tools provide evidence about the health impacts of the physical urban environment which can be used in built environment policy and decision-making. Where UHI tools provide data at the neighborhood (and lower) scale they can provide valuable information about health inequalities and environmental deprivation. This review performs a census of UHI tools and explores their nature and characteristics (including how they represent, simplify or address complex systems) to increase understanding of their potential use by municipal built environment policy and decision-makers. We searched seven bibliographic databases, four key journals and six practitioner websites and conducted Google searches between January 27, 2016 and February 24, 2016 for UHI tools. We extracted data from primary studies and online indicator systems. We included 198 documents which identified 145 UHI tools comprising 8006 indicators, from which we developed a taxonomy. Our taxonomy classifies the significant diversity of UHI tools with respect to topic, spatial scale, format, scope and purpose. The proportions of UHI tools which measure data at the neighborhood and lower scale, and present data via interactive maps, have both increased over time. This is particularly relevant to built environment policy and decision-makers, reflects growing analytical capability and offers the potential for improved understanding of the complexity of influences on urban health (an aspect noted as a particular challenge by some indicator producers). The relation between urban health indicators and health impacts attributable to modifiable environmental characteristics is often indirect. Furthermore, the use of UHI tools in policy and decision-making appears to be limited, thus raising questions about the continued development of such tools by multiple organisations duplicating scarce resources. Further research is needed to understand the requirements of built environment policy and decision-makers, public health professionals and local communities regarding the form and presentation of indicators which support their varied objectives. Electronic supplementary material The online version of this article (10.1007/s11524-018-0228-8) contains supplementary material, which is available to authorized users.
Chapter
The concept of neighborhood is at the forefront of place and health research as an appropriate scale of study (Sampson 2003; Riva et al. 2007). What actually constitutes a neighborhood is subjective at best, but an underlying idea of a geographic unit of limited size with some degree of social interaction is generally acknowledged (Weiss et al. 2007). The last two decades have seen significant evidence that the neighborhood influences health beyond individual characteristics in the developed and developing world (Pickett and Pearl 2001; Sampson et al. 2002; Stafford and Marmot 2003; Montgomery and Hewett 2005; Perera et al. 2009; Diez-Roux 1998; Robert 1999). While there are a number of issues for health and place research, establishing geographical boundaries that define ‘place’ has been the subject of significant debate in the field (Diez Roux 2001; Entwisle 2007; Gauvin et al. 2007).
Article
This article is a commentary on a paper by Galster, Metzger and Waite submitted in the Housing and Neighbourhood Domain Workshop of the Metropolis Third National Conference held in Vancouver, British Columbia in January 1999. After reviewing the major elements of Metzger et al.'s contribution, the authors discuss the methodological difficulties associated with the neighbourhood concept. They argue that the analysis of the complex socio-spatial reality of the neighbourhood requires a multidimensional approach involving quantitative and qualitative methods. In this paper, the neighbourhood concept is discussed from a sociological standpoint, with arguments and conclusions based on several case studies on immigration and neighbourhood life.
Article
Most earlier models of residential sorting employ a 'featureless plain', paying little attention to cities' physical environments. The empirical question of physical features mitigating neighbour externalities remains largely unexplored. This article adds to the literature by considering the environmental aspects of group boundaries. Physical barriers that mitigate the externality of neighbours' characteristics should be expected to have important differential effects on urban land use patterns. This hypothesis is tested for the percentage of Black people in Chicago in 2000. Some features (such as, parks, railroads, major roads) have strong barrier effects. Despite the limitations of this approach, the findings appear robust to spatial dependence in the data. The findings hold important implications for future research into residential location decisions and the planning of public amenities and infrastructure.
Article
Abraham Maslow's Hierarchy of Needs theory proposed that there is a hierarchy of human needs, which range from the basics of survival to creativity. Building from Maslow's idea, this paper examines the first level of human needs and empirically constructs a basics of living (BOL) index for every United States county from 28 candidate factors. The final BOL Index includes nine measures of air, water, and income. This study examined whether county-level health would be poor if the BOL were not met. Findings showed a positive correlation between the BOL index and all cause mortality, the fewer BOL, the worse the all cause mortality. Strikingly, poor BOL is most often found in small jurisdictions (<25,000 population) and less often in larger jurisdictions (>150,000 population). In the mapping of the BOL index, areas of greatest need can be observed and in multivariate modeling, highest overall mortality is regional. The BOL index predicts a substantial amount of variation observed in county overall mortality and is independent of county poverty and quality of life.
Article
To identify spatial patterns of changes in infant mortality rates and proportions of low-birth-weight live births observed in 1994. Canada. Live births and infant deaths in Canada between 1987 and 1994. Data for Newfoundland were unavailable for 1987 through 1990. Annual infant mortality rates (crude and after excluding live newborns weighing less than 500 g); proportion of live births by low-birth-weight category (500-2499 g). Nova Scotia, New Brunswick, Quebec and Manitoba had lower crude and adjusted infant mortality rates in 1994 than in 1993. Newfoundland, Saskatchewan, Alberta and British Columbia had higher rates in 1994 than in 1993. The crude rate in Ontario was lower, and the adjusted rate higher, in 1994 than in 1993. A downward trend in the proportion of low-birth-weight live births was observed in Quebec (chi(2) for trend = 29.2, p < 0.01). Conversely, an upward trend was observed in Ontario (chi(2) for trend = 241.3, p < 0.01). However, the increase may have been due to data errors, especially in 1993 and 1994, involving truncation of ounces in 2 digits to 1 digit (e.g., 5 pounds 10 ounces became 5 pounds 1 ounce). Although the marginal increases in infant mortality observed in several provinces could be the result of random variation, future trends should be closely monitored. The proportion of low-birth-weight live births in Canada (excluding Ontario) appears to be stable, with Quebec showing significant reductions. The errors in data for Ontario need to be corrected before trends can be estimated for that province and for Canada as a whole.