Geospatial analyses are increasingly used in population oncology. We provide a first review of geospatial analysis in Canadian population oncology research, compare to international peers, and identify future directions. Geospatial-focused peer-reviewed publications from 1992–2020 were compiled using PubMed, MEDLINE, Web of Science, and Google Scholar. Abstracts were screened for data derived from a Canadian cancer registry and use of geographic information systems. Studies were classified by geospatial methodology, geospatial unit, location, cancer site, and study year. Common limitations were documented from article discussion sections. Our search identified 71 publications using data from all provincial and national cancer registries. Thirty-nine percent (N = 28) were published in the most recent 5-year period (2016–2020). Geospatial methodologies included exposure assessment (32.4%), identifying spatial associations (21.1%), proximity analysis (16.9%), cluster detection (15.5%), and descriptive mapping (14.1%). Common limitations included confounding, ecologic fallacy, not accounting for residential mobility, and small case/population sizes. Geospatial analyses are increasingly used in Canadian population oncology; however, efforts are concentrated among a few provinces and common cancer sites, and data are over a decade old. Limitations were similar to those documented internationally, and more work is needed to address them. Organized efforts are needed to identify common challenges, develop leading practices, and identify shared priorities.

Geospatial analysis in population oncology research utilizes spatial data to associate geographic information, cancer and relevant determinants (1). It is used across the cancer control continuum to increase understanding of cancer etiology, identify at-risk populations, and inform health service provision (2, 3). Recent advancements in geospatial methods have coincided with increases in geospatial applications to population oncology (1–3). Such advances include developments of analytic tools, methodologic approaches, and technologic capacity in geographic information systems (GIS; refs. 2, 3).

In the United States, geospatial applications are strongly supported by national agencies such as the NCI (2–4). The NCI has long supported efforts to develop and improve disease mapping related to cancer incidence, mortality, and survival. Web-based interactive mapping applications and geospatial tools and resources are publicly available through the NCI GIS Portal (5). Peer-reviewed publications using spatial analyses of U.S. data are also widely available and increasingly common (2). Similar initiatives have been demonstrated by international peers and organizations, such as the International Agency for Research on Cancer (IARC) and the International Association of Cancer Registries (IACR; refs. 6, 7). Specifically, IARC and IACR coordinate Global Cancer Observatory products, such as web-based data visualizations (both spatial and nonspatial) and reports that leverage data from population-based cancer registries around the world (8).

In Canada, cancer surveillance consists of 13 provincial and territorial cancer registries that collect information on all diagnosed tumors among Canadian residents, as well as patient demographics and geographic (residential) data (9). This availability of residential information permits spatial analysis of disease, risk factor and services locations, as well as etiology research (10, 11). A limited number of information products have come from geospatial analysis in Canadian population oncology, such as the Canadian Cancer Incidence Atlas, 2000–2006 (12). In this atlas, spatial statistics and geomapping were applied to examine cancer incidence from 2000 to 2006 across Canadian health regions (12). However, unlike U.S.-based research (2) the peer-reviewed literature on geospatial analysis in Canadian population oncology remains scattered and has not been systematically reviewed. The purpose of this article was to provide, to our knowledge, the first review of published, peer-reviewed geospatial analyses in Canadian population oncology research, draw comparisons to international counterparts, and inform future directions.

A comprehensive literature search was conducted in Google Scholar, PubMed, MEDLINE, and Web of Science in February 2020 to identify geospatial population oncology research in Canada. Search terms were combined (with “AND”) from three separate search constructs corresponding to geospatial terms, Canadian geographies, and cancer terms (Fig. 1). The following terms were used on Google Scholar: “Canada” OR “Cancer” OR “Registry” OR "Spatial analysis" OR "Medical Geography" OR geocoded OR GIS OR "Geographic Information Systems."

Figure 1.

Flow diagram of literature search.

Figure 1.

Flow diagram of literature search.

Close modal

Studies were considered if they were (i) included data from a cancer registry in Canada, (ii) included any spatial analysis, and (iii) published in a peer-reviewed journal. Ulrichsweb (13), a database of bibliographic and publisher information, was used to verify that articles were published in peer-reviewed journals. Studies that used nonspatial regression to identify associations between cancer outcomes and urban/rural status or by regional geographic classifications (e.g., region of residence at diagnosis), without use of any geospatial methodology listed in Table 1, were not considered. Studies examining national cancer rates in relation to international counterparts were also not considered. There were no filters applied for year of publication. Titles and abstracts were screened and relevant articles were reviewed and coded by a single reviewer. Titles of studies included in the reference sections of fully reviewed articles were also examined for any additional potential studies. Articles were classified according to geospatial methodology, cancer site, geospatial unit of analysis, study location (i.e., province or territory of cancer registry), study period, year of study, and software used for analysis.

Table 1.

Descriptions and examples of geospatial methodology categories used to classify articles.

Geospatial methodologyDescriptionExampleStudies from this literature review
Identifying spatial associations Examines effect sizes between exposure(s) and outcomes, while accounting for spatial dependence in effect sizes using spatial regression modeling. A Bayesian spatial Poisson model was used to examine the effect of sociodemographic characteristics on non–small cell lung cancer incidence at a small geographic level (14). Saint-Jacques et al. (2018; ref. 41) 
   Singh et al. (2017; ref. 42) 
   Dawe et al. (2017; ref. 14) 
   Brown et al. (2016; ref. 43) 
   Saint-Jacques et al. (2016; ref. 47) 
   Torabi et al. (2015; ref. 44) 
   Torabi et al. (2014; ref. 45) 
   Jiang et al. (2014; ref. 62) 
   Hystad et al. (2013; ref. 34) 
   Torabi et al. (2012; ref. 63) 
   Torabi et al. (2011; ref. 64) 
   Holowaty et al. (2010; ref. 46) 
   Chen et al. (2008; ref. 65) 
   Van Leeuwen et al. (1999; ref. 66) 
   Walter (1992; ref. 67) 
Proximity-analysis Examines distance or travel time. Multivariable logistic regression was used to examine the effect of palliative radiotherapy utilization rates and travel time to the cancer center. Distance to cancer center was measured from postal code at diagnosis (15). Ahmed et al. (2019; ref. 68) 
   Chan et al. (2019; ref. 54) 
   Canale et al. (2018; ref. 69) 
   McCrate et al. (2018; ref. 70) 
   Payette et al. (2017; ref. 71) 
   Walker et al. (2017; ref. 72) 
   Liu et al. (2015; ref. 73) 
   Zhang et al. (2015; ref. 74) 
   Huang et al. (2014; ref. 15) 
   Helewa et al. (2013; ref. 75) 
   Groome et al. (2008; ref. 76) 
   Paszat et al. (1998; ref. 77) 
Cluster detection Identifies areas with higher/lower risk of disease. The scan statistics were used to examine areas at elevated risk in Ontario at the county and dissemination level (16). Ghazawi et al. (2019; ref. 78) 
   Ye et al. (2017; ref. 79) 
   Li et al. (2016; ref. 80) 
   Lofters et al. (2013; ref. 81) 
   Milewski (2012; ref. 82) 
   Luginaah et al. (2012; ref. 16) 
   Kulkarni et al. (2011; ref. 83) 
   Torabi et al. (2011; ref. 84) 
   Rosychuk et al. (2010; ref. 85) 
   Yiannakoulias (2009; ref. 86) 
   Walter (1992; ref. 87) 
Exposure-assessment Estimating levels of exposure that have an important spatial component. Self-reported residential histories were obtained from a cohort of lung cancer cases to recreate exposure history to air pollution from multiple ambient air pollutants, vehicle and industrial emissions (17). Ritonja et al. (2020; ref. 33) 
   Tabaczynski (2020; ref. 88) 
   Lagacé et al. (2019; ref. 89) 
   Pinault et al. (2017; ref. 90) 
   Auluck et al. (2016; ref. 91) 
   Trinh et al. (2016; ref. 92) 
   Labine et al. (2015; ref. 93) 
   Winters et al (2015; ref. 32) 
   Auluck et al. (2014; ref. 94) 
   Hystad et al. (2013; ref. 31) 
   Hwang et al. (2013; ref. 95) 
   Wanigaratne et al. (2013; ref. 36) 
   Hystad et al. (2012; ref. 17) 
   Pan et al. (2011; ref. 35) 
   Anderson et al. (2011; ref. 96) 
   Zhang-Salomons et al. (2006; ref. 97) 
   Borugian et al. (2005; ref. 98) 
   Gorey et al. (2003; ref. 55) 
   Gorey et al. (2000; ref. 99) 
   Mackillop et al. (2000; ref. 100) 
   Boyd et al. (1999; ref. 101) 
   Goel et al. (1997; ref. 102) 
   Gorey et al. (1997; ref. 97) 
Descriptive mapping Describes the spatial distribution of disease of a population. Cutaneous T-cell lymphoma incidence and mortality were mapped across Canada at the forward sortation area level (18). Larouche et al. (2020; ref. 103) 
   Darwich et al. (2019; ref. 104) 
   Ghazawi et al. (2019; ref. 105) 
   Ghazawi et al. (2019; ref. 106) 
   Ghazawi et al. (2019; ref. 107) 
   Le et al. (2019; ref. 108) 
   Ghazawi et al. (2018; ref. 109) 
   Ghazawi et al. (2017; ref. 18) 
   Walker et al. (2015; ref. 110) 
   Walter et al. (1994; ref. 111) 
Geospatial methodologyDescriptionExampleStudies from this literature review
Identifying spatial associations Examines effect sizes between exposure(s) and outcomes, while accounting for spatial dependence in effect sizes using spatial regression modeling. A Bayesian spatial Poisson model was used to examine the effect of sociodemographic characteristics on non–small cell lung cancer incidence at a small geographic level (14). Saint-Jacques et al. (2018; ref. 41) 
   Singh et al. (2017; ref. 42) 
   Dawe et al. (2017; ref. 14) 
   Brown et al. (2016; ref. 43) 
   Saint-Jacques et al. (2016; ref. 47) 
   Torabi et al. (2015; ref. 44) 
   Torabi et al. (2014; ref. 45) 
   Jiang et al. (2014; ref. 62) 
   Hystad et al. (2013; ref. 34) 
   Torabi et al. (2012; ref. 63) 
   Torabi et al. (2011; ref. 64) 
   Holowaty et al. (2010; ref. 46) 
   Chen et al. (2008; ref. 65) 
   Van Leeuwen et al. (1999; ref. 66) 
   Walter (1992; ref. 67) 
Proximity-analysis Examines distance or travel time. Multivariable logistic regression was used to examine the effect of palliative radiotherapy utilization rates and travel time to the cancer center. Distance to cancer center was measured from postal code at diagnosis (15). Ahmed et al. (2019; ref. 68) 
   Chan et al. (2019; ref. 54) 
   Canale et al. (2018; ref. 69) 
   McCrate et al. (2018; ref. 70) 
   Payette et al. (2017; ref. 71) 
   Walker et al. (2017; ref. 72) 
   Liu et al. (2015; ref. 73) 
   Zhang et al. (2015; ref. 74) 
   Huang et al. (2014; ref. 15) 
   Helewa et al. (2013; ref. 75) 
   Groome et al. (2008; ref. 76) 
   Paszat et al. (1998; ref. 77) 
Cluster detection Identifies areas with higher/lower risk of disease. The scan statistics were used to examine areas at elevated risk in Ontario at the county and dissemination level (16). Ghazawi et al. (2019; ref. 78) 
   Ye et al. (2017; ref. 79) 
   Li et al. (2016; ref. 80) 
   Lofters et al. (2013; ref. 81) 
   Milewski (2012; ref. 82) 
   Luginaah et al. (2012; ref. 16) 
   Kulkarni et al. (2011; ref. 83) 
   Torabi et al. (2011; ref. 84) 
   Rosychuk et al. (2010; ref. 85) 
   Yiannakoulias (2009; ref. 86) 
   Walter (1992; ref. 87) 
Exposure-assessment Estimating levels of exposure that have an important spatial component. Self-reported residential histories were obtained from a cohort of lung cancer cases to recreate exposure history to air pollution from multiple ambient air pollutants, vehicle and industrial emissions (17). Ritonja et al. (2020; ref. 33) 
   Tabaczynski (2020; ref. 88) 
   Lagacé et al. (2019; ref. 89) 
   Pinault et al. (2017; ref. 90) 
   Auluck et al. (2016; ref. 91) 
   Trinh et al. (2016; ref. 92) 
   Labine et al. (2015; ref. 93) 
   Winters et al (2015; ref. 32) 
   Auluck et al. (2014; ref. 94) 
   Hystad et al. (2013; ref. 31) 
   Hwang et al. (2013; ref. 95) 
   Wanigaratne et al. (2013; ref. 36) 
   Hystad et al. (2012; ref. 17) 
   Pan et al. (2011; ref. 35) 
   Anderson et al. (2011; ref. 96) 
   Zhang-Salomons et al. (2006; ref. 97) 
   Borugian et al. (2005; ref. 98) 
   Gorey et al. (2003; ref. 55) 
   Gorey et al. (2000; ref. 99) 
   Mackillop et al. (2000; ref. 100) 
   Boyd et al. (1999; ref. 101) 
   Goel et al. (1997; ref. 102) 
   Gorey et al. (1997; ref. 97) 
Descriptive mapping Describes the spatial distribution of disease of a population. Cutaneous T-cell lymphoma incidence and mortality were mapped across Canada at the forward sortation area level (18). Larouche et al. (2020; ref. 103) 
   Darwich et al. (2019; ref. 104) 
   Ghazawi et al. (2019; ref. 105) 
   Ghazawi et al. (2019; ref. 106) 
   Ghazawi et al. (2019; ref. 107) 
   Le et al. (2019; ref. 108) 
   Ghazawi et al. (2018; ref. 109) 
   Ghazawi et al. (2017; ref. 18) 
   Walker et al. (2015; ref. 110) 
   Walter et al. (1994; ref. 111) 

Geospatial methodologies were further classified as: identifying spatial associations, proximity analysis, cluster detection, exposure assessment, or descriptive mapping. These categories were chosen to enable comparisons with a previous review of geospatial population oncology research in the United States (2). Descriptions and examples for each category are listed in Table 1 (14–18). Studies were classified as descriptive mapping if the sole purpose of the study was to map disease estimates and no other geospatial methodology was applied. If a study included descriptive mapping and at least one other geospatial method, the study was not classified as a descriptive mapping study. Furthermore, if a study included more than one geospatial method, apart from descriptive mapping, the study was classified with the geospatial method most related to the study's primary objective. All analyses were conducted in R (19).

Cancer types included all cancers, specific cancer types, and multiple cancer sites. Geospatial units included various regional administrative health regions, postal codes, and Statistics Canada census area classifications (20). Studies that examined multiple geospatial units were assigned to the smallest area unit of analysis.

The common gaps and/or limitations identified by the authors were documented from the discussion sections of the relevant articles. Furthermore, data sources used to complement cancer registry data were noted.

A flow diagram of the literature search is shown in Fig. 1. The searches resulted in a total of 1,778 abstracts; the first 1,000 results of 5,912 total links were available on Google Scholar, 333 PubMed abstracts, 142 MEDLINE abstracts, and 303 Web of Science abstracts. Studies that did not include data from a cancer registry in Canada and use some form of geospatial methods were excluded (N = 1,718), leaving 60 peer-reviewed articles for full review. Eleven additional studies were found through the discussion and reference sections of the reviewed articles resulting in a total of 71 studies (Fig. 1).

Overall, data from the Ontario cancer registry was the most commonly used among studies (38.0%, n = 27). This was followed by cancer registries from Alberta and British Columbia (each at 21.1%, n = 15), Manitoba (16.9%, n = 12), Quebec (14.1%, n = 10), the Nova Scotia and Saskatchewan (each at 11.3%, n = 8), Newfoundland and Labrador (9.9%, n = 7), Prince Edward Island (8.5%, n = 6), and New Brunswick (1.4%, n = 1). Sixteen studies (22.5%) included cancer registry data from the national Canadian Cancer Registry (CCR).

The studies were also examined by geospatial methodology and year of publication (Fig. 2), as well as by cancer site, geospatial unit of analysis, and software use (Fig. 3AC). Exposure assessment was the most commonly used methodology (32.4%, n = 23), followed by identifying spatial associations (21.1%, n = 15), proximity analysis (16.9%, n = 12), cluster detection (15.5%, n = 11), and descriptive mapping (14.1%, n = 10; Fig. 1). The earliest publication was in 1992 and the most recent was in 2020. The majority of studies were published in the most recent 5-year period 2016–2020 (39.4%, n = 28), and 2011–2015 (35%, n = 25), with the rest being published before 2010 (25.4%, n = 18; Fig. 2). Overall, the cancer data used in these studies ranged from 1966 to 2015. The median number of data years included in each study was 13 years (25th–75th percentiles: 5–19 years) with a minimum of 1 year and maximum of 30 years. After separating the range of data years examined into 5-year periods, the most common periods represented were 1996–2000 (n = 43) followed by 2001–2005 (n = 41) and 1991–1995 (n = 39). Across all studies, 62% (n = 44) published disease maps.

Figure 2.

Number of studies published by 5-year period and category, 1992–2020.

Figure 2.

Number of studies published by 5-year period and category, 1992–2020.

Close modal
Figure 3.

Proportion of publications by most common cancers* (A), geographic units** (B), and analysis software*** (C). *Other cancer types = kidney, leukemia, melanoma, childhood leukemias, central and nervous system, liver, non-Hodgkin lymphoma, oral cavity, stomach, thyroid, cervix, ovary, and pancreas. **Other geographic units = county, health region, census division, enumeration area, municipal, province, regional health authority, health service delivery area, and neighborhood. ***Other software = Excel, GeoDa, BOA, EPIDAT, FleXScan, Google Maps, GraphPad, INLA, JoinPoint, MapPoint, an online distance calculator, Orius, PRIMER, QGIS, Rapid Inquiry facility, S-PLUS, and SPACESTAT.

Figure 3.

Proportion of publications by most common cancers* (A), geographic units** (B), and analysis software*** (C). *Other cancer types = kidney, leukemia, melanoma, childhood leukemias, central and nervous system, liver, non-Hodgkin lymphoma, oral cavity, stomach, thyroid, cervix, ovary, and pancreas. **Other geographic units = county, health region, census division, enumeration area, municipal, province, regional health authority, health service delivery area, and neighborhood. ***Other software = Excel, GeoDa, BOA, EPIDAT, FleXScan, Google Maps, GraphPad, INLA, JoinPoint, MapPoint, an online distance calculator, Orius, PRIMER, QGIS, Rapid Inquiry facility, S-PLUS, and SPACESTAT.

Close modal

The most common cancers examined were breast (23.9%, n = 17), lung (16.9%, n = 12), and colorectal (15.5%, n = 11; Fig. 3A). Five studies examined multiple cancer sites (each study examined greater than ten sites). The most common geographic units used for analysis was six-digit postal code (19.7%, n = 14), followed by dissemination area and forward sortation area (each at 12.7%, n = 9), and home address (9.9%, n = 7; Fig. 3B). Software use was plotted by study category in Fig. 2C. The most common computer software programs used were ArcGIS (23.9%, n = 17), SAS (19.7%, n = 14), and R (12.7%, n = 9. ArcGIS, R, and SAS were used for a variety of geospatial methods. Some programs were specific to a certain set of methods (e.g., WinBUGS was used specifically for identifying spatial associations modeling and cluster detection; Fig. 3C).

Commonly cited gaps and limitations included confounding (n = 42), ecologic fallacy (n = 24), exposure misclassification (n = 22), and small case counts/population sizes (n = 16), among others (Table 2). Common data sources used in addition to cancer registries are described in Supplementary Table S1.

Table 2.

Common gaps and limitations by authors in the discussion sections of articles.

LimitationCountsaExample 1Example 2
Accounting for confounders and residual confounding 42 Risk factors were not included in the analysis limiting causal inference (16). Cancer rates may be confounded by distributions of unknown or unmeasured factors related to cancer (85). 
Ecologic fallacy 24 Results should be interpreted only in the context of area of residence (42). Ecologic studies are limited for causal inference but are often the first step to determine whether an association exists (16). 
Exposure misclassification including those related to changes in environmental and residential histories. 22 Spatial accuracy of postal codes vary between rural and urban areas in Canada. Exposure to vehicle and industrial emissions based on proximity will be more accurate in urban areas (17). Current postal code may not reflect past exposures that may have led to the development of breast cancer, which may take many years to develop (16). 
Small cases and population sizes 16 Unable to stratify by subtypes of cancer due to small number of cases (44). Some geographic units had small population denominators, reducing statistical power (41). 
Lack of information at the individual-level on sociodemographics and health behaviors Individual-level data would support more robust models for evaluating the associations; however, individual-level risk factor data were not available (14). 
Changes in cancer reporting and geographic boundaries over time Diagnosis and reporting changes may have occurred during the period of analysis (85). Postal code and geographic boundaries changed during the study period (79). 
Using covariate data from a single year to represent multiple years. Only a single year of data was used for covariates, which may not be reflective of early (or latter) time periods (42, 72). 
Self-reported measurements, recall bias and other information biases (excluding misclassification of exposure) Self-reported measures of physical activity can be subject to recall biases and error (92). 
Modifiable area unit problem Results may be influenced by the choice of geographic unit (65). 
Selection bias  
Multiple testing Numerous statistical comparisons were made without adjustment for multiple testing, and there may be chance findings (92). 
Generalizability  
LimitationCountsaExample 1Example 2
Accounting for confounders and residual confounding 42 Risk factors were not included in the analysis limiting causal inference (16). Cancer rates may be confounded by distributions of unknown or unmeasured factors related to cancer (85). 
Ecologic fallacy 24 Results should be interpreted only in the context of area of residence (42). Ecologic studies are limited for causal inference but are often the first step to determine whether an association exists (16). 
Exposure misclassification including those related to changes in environmental and residential histories. 22 Spatial accuracy of postal codes vary between rural and urban areas in Canada. Exposure to vehicle and industrial emissions based on proximity will be more accurate in urban areas (17). Current postal code may not reflect past exposures that may have led to the development of breast cancer, which may take many years to develop (16). 
Small cases and population sizes 16 Unable to stratify by subtypes of cancer due to small number of cases (44). Some geographic units had small population denominators, reducing statistical power (41). 
Lack of information at the individual-level on sociodemographics and health behaviors Individual-level data would support more robust models for evaluating the associations; however, individual-level risk factor data were not available (14). 
Changes in cancer reporting and geographic boundaries over time Diagnosis and reporting changes may have occurred during the period of analysis (85). Postal code and geographic boundaries changed during the study period (79). 
Using covariate data from a single year to represent multiple years. Only a single year of data was used for covariates, which may not be reflective of early (or latter) time periods (42, 72). 
Self-reported measurements, recall bias and other information biases (excluding misclassification of exposure) Self-reported measures of physical activity can be subject to recall biases and error (92). 
Modifiable area unit problem Results may be influenced by the choice of geographic unit (65). 
Selection bias  
Multiple testing Numerous statistical comparisons were made without adjustment for multiple testing, and there may be chance findings (92). 
Generalizability  

aStudies often cited multiple limitations, and the total of counts do not add to 71.

Summary of main findings

To our knowledge, this study is the first review of Canadian population oncology geospatial research. The Canadian literature was limited to 71 studies, although the number of studies has increased substantially just in the past 5 years. The studies employed a diverse set of geospatial methodologies. The most common were exposure assessment, identifying spatial associations, proximity analysis, cluster detection, and disease mapping. Studies generally used small geographic units, including home address, six-digit postal code (a six-character code maintained by Canada Post Corporation for the purpose of sorting and delivering mail; ref. 21), and dissemination areas (areas based on population size of nearly 400 to 700 persons; ref. 22). Among proximity analysis studies, authors typically calculated distances between points of interest (e.g., cancer center) and location at residence characterized by the center of postal codes. Current efforts were concentrated among a few provinces, focused on the common cancer sites, and typically made use of data over a decade old. Common limitations cited by authors included confounding, the ecologic fallacy, not accounting for residential mobility and exposure misclassification, and small case and population sizes.

Comparison with U.S. geospatial cancer research

Similar to this study, U.S.-based geospatial cancer research has increased over time (2), although in comparison, the U.S. literature is larger and more developed. In an analysis of U.S.-based research, Korycinski and colleagues found 725 publications related to geospatial research published after 2000. Most of these studies were published after 2010, coinciding with growth in geospatial tools and methods (2). Nearly half of U.S. geospatial cancer studies were published by five cancer research centers (2). Similarly, this study found that the majority of studies used data from three Canadian cancer registries; with the Ontario cancer registry the most common represented in over a third of all studies. As posited by Korycinski and colleagues, concentrations of geospatial research among cancer research institutions may be related to differences in the availability of funding, resources, expertise, and priorities (2).

Among methods examined in this study, the most common approaches were similar to those found in the United States, albeit in a slightly different order. In the United States, identifying spatial associations was the most common approach, followed by exposure assessment and proximity studies (2). The presence of multiple methods, geospatial units, and software highlight the diversity in research questions and the geospatial tools available. However, it may also imply differences in resources, expertise, and data availability that may limit what can be done in some regions more so than in others.

Common limitations and examples of novel practices

Common challenges identified by this study were also cited within the international literature (2, 3, 23–29). Most geospatial studies focus on place of residence to assign exposures. However, this approach does not account for the fact that people spend a significant amount of time away from their home location (2, 3, 23). Furthermore, residence is often taken at a single point in time, even though nearly 40% of Canadians change their home residence within a 5-year period (30). These issues result in exposure misclassification as they fail to account for cumulative exposure and changes in residential history (2, 3, 17, 24). Only seven studies in the present review accounted for changes in residential history (17, 31–36). Many of these studies used data from the National Enhanced Cancer Surveillance System, a collaborative effort between Health Canada and Canadian Provincial cancer registries to strengthen and facilitate the evaluation of environmental–cancer concerns (37). In these studies, self-reported residential histories were collected through research questionnaires mailed to cancer cases identified from provincial cancer registries. Residential histories were then used to recreate historical exposures. In the United States, Wheeler and colleagues used a public records database to reconstruct residential histories; however, the usefulness of these histories depended on the period of interest and geographic area (24). Following Schootman and colleagues, we recommend routine geocoding of health data and collection of time-varying contextual information in addition to residential locations, such as health services and neighborhoods perception data (3). Longitudinal census-based and linked health survey cohorts from Statistics Canada, such as the Canadian Census, Health, and Environment Cohort (CanCHEC) or the mortality-linked Canadian Community Health Survey Cohort, may provide the data quality required to overcome such exposure misclassification (38, 39).

Most studies in the present review were cross-sectional and ecologic, consistent with the U.S. literature (2, 3). Such study designs limit causal inference between spatially relevant exposures and cancer outcomes. As suggested by Schootman and colleagues, we recommend incorporating geocoded data within longitudinal cohort studies and examining exposure locations before and after diagnosis (3). Longitudinal census-based and linked health survey cohorts from Statistics Canada can also address this limitation (39, 40). Novel statistical approaches that enable better examination of causality have also been suggested (3). Croon and colleagues proposed micro–macro regression models, a type of multilevel model, to predict aggregate-level outcomes from individual-level variables (25). Multilevel models can reduce the risk of the ecologic fallacy by modeling relationships at different levels of a hierarchy (25). This approach has not been well explored in population oncology (3).

Examining small geographic areas can help better identify disease patterns given local variations in risk factors, screening, and service use (26). However, unstable risk estimates due to small case counts or population sizes were commonly cited in this review. Statistical approaches such as Bayesian hierarchical modeling may be used in these situations to stabilize risk estimates (41–47). Saint-Jacques and colleagues used the Bayesian Besag-York-Mollie (BYM) model to quantify the risk of bladder or kidney cancer in relation to arsenic exposure (41). Holowaty and colleagues used the BYM model to map standardized incidence ratios at the dissemination area level for numerous cancers (46). The BYM model is a spatial autoregressive model that can analyze and map incidence data among small areas and low population density (14, 26, 41, 46, 48). However, it can be limited by oversmoothing and low sensitivity for detecting areas with excess risk (28, 46). An alternative approach is to stabilize local risk estimates through empirical Bayesian modeling (27), which has recently been integrated into the ArcGIS (a common GIS software; ref. 49) Rate Stabilizing Tool.

In Canada, geospatial analysis of cancer data may miss important aspects of population heterogeneity, including ethnicity, immigration, and Indigenous status. Recent efforts in Canadian cancer control led by Indigenous people are working towards establishing culturally responsive data governance and filling the evidence gap in cancer information among Indigenous Peoples (50–53). In this review, four geospatial studies included Indigenous status (14, 44, 54, 55). Torabi and colleagues examined the relationship between childhood leukemia incidence and sociodemographic characteristics using Bayesian spatial Poisson regression. The proportion of people with Indigenous Status per areal unit was included in the model (44). Dawe and colleagues used a similar approach to model non–small cell lung cancer incidence by census subdivision (14). Given recent efforts in Indigenous cancer control, increased accuracy of digital geographic information (56), and the increasing trend of geospatial methods applied to Canadian population oncology research, there is a need to examine issues related to privacy, interpretation and information governance. Furthermore, given relatively small population sizes of Indigenous peoples in Canada (57), there is a need to explore opportunities for supporting Indigenous-led efforts using geospatial methods.

The impact of nationally coordinated GIS efforts in population oncology

GIS and spatial data were previously underutilized in the United Kingdom despite the availability of geographic information in most government data (58). This was attributed to a lack of a clear strategy around GIS and inclusion of GIS terminology in strategic documents (58). In the United States, the NCI, among other institutions, have long supported GIS with funding opportunities, methods and tool development, and fostering collaboration. Language concerning GIS-related analysis are also consistently included in strategic documents (1–3, 58). Nationally coordinated efforts have had a strong impact on the advancement of geospatial methods in cancer control research in the United States (2, 3, 58). In Canada, there are few examples at the national level of applying geospatial methods to population oncology (12). Instead, geospatial applications are typically led by provincial cancer and health organizations (59–61). Given the increasing use of geospatial methods in Canada, nationally organized efforts could help identify common challenges, develop leading practices, and identify shared priorities moving forward.

Limitations

This review was limited to a set of keywords and MeSH terms in the search criteria. Some relevant geospatial studies were potentially not captured, although this study used similar keywords from a previous review of the U.S. literature to ensure comparability of methods and findings (2). The focus of this study was peer-reviewed literature and did not capture geospatial methods applied to government and nonacademic settings. Google Scholar was included to capture a broader scope of peer-reviewed literature. Abstracts were screened and articles were reviewed and coded by a single reviewer. Therefore, potential articles may have been incorrectly screened or included articles misclassified. However, strict inclusion criteria and methodologies were chosen and defined a priori based off of a published review (2) to ensure a systematic approach to reviewing the literature.

Conclusions

Geospatial methods are applied across the cancer control continuum and support the advancement and understanding of cancer etiology, identifying at-risk populations, and informing health services (2, 3). To our knowledge, this study is the first to examine geospatial research in Canadian population oncology. Despite a relatively small number of studies, geospatial research using data from Canadian cancer registries is increasing, although it is concentrated in a few regions and typically makes use of data over a decade old. Multiple methods, geospatial units, and software were reported, implying diversity among research questions but also, a lack of a standardized approaches to geospatial research. Evaluating consistency across multiple scales and approaches could indicate real effects, as opposed to confounded or biased results. Studies shared common limitations that coincide with those cited in the international literature. Novel approaches were highlighted in this review but more work is needed to address limitations that are common throughout the geospatial setting. Organized efforts across Canadian population oncology organizations are needed to establish leading practices in geospatial research, develop solutions to common challenges, and determine priorities moving forward.

M.C. Otterstatter is a consultant for Canadian Partnership Against Cancer. No potential conflicts of interest were disclosed by the other authors.

This study was supported by the Canadian Institutes of Health Research through the Canada Graduates Scholarships Doctoral Award. The scholarship provides financial support for students in doctoral studies. J. Simkin is a recipient of this award. The Canadian Institutes of Health Research has no involvement in development of this manuscript. The authors acknowledge the Canadian Institutes of Health Research for their support through the Canada Graduates Scholarships Doctoral Award.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Richardson
DB
,
Volkow
ND
,
Kwan
M-P
,
Kaplan
RM
,
Goodchild
MF
,
Croyle
RT
. 
Spatial Turn in Health Research
.
Science
2013
;
339
:
1390
2
.
2.
Korycinski
RW
,
Tennant
BL
,
Cawley
MA
,
Bloodgood
B
,
Oh
AY
,
Berrigan
D
. 
Geospatial approaches to cancer control and population sciences at the United States cancer centers
.
Cancer Causes Control
2018
;
29
:
371
7
.
3.
Schootman
M
,
Gomez
SL
,
Henry
KA
,
Paskett
ED
,
Ellison
GL
,
Oh
A
, et al
Geospatial approaches to cancer control and population sciences
.
Cancer Epidemiol Biomarkers Prev
2017
;
26
:
472
5
.
4.
Pickle
LW
. 
A history and critique of U.S. mortality atlases
.
Spat Spatiotemporal Epidemiol
2009
;
1
:
3
17
.
5.
National Cancer Institute
. 
NCI GIS Portal - Geographic Information Systems & Science
. Available from
: https://gis.cancer.gov/.
6.
World Health Organization
. 
International Agency for Research on Cancer
. Available from
: https://www.iarc.fr/.
7.
World Health Organization
. 
International Association of Cancer Registries
. Available from
: http://www.iacr.com.fr/.
8.
Ferlay
F
,
Ervik
M
,
Lam
F
,
Colombet
M
,
Mery
L
,
Piñeros
M
, et al
Global Cancer Observatory: Cancer Today
.; 
2018
. Available from
: https://gco.iarc.fr/today.
9.
Statistics Canada
. 
Surveys and statistical programs - Canadian Cancer Registry (CCR)
. Available from
: http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3207#a2.
10.
Jerrett
M
,
Burnett
R
,
Goldberg
M
,
Sears
M
,
Krewski
D
,
Catalan
R
, et al
Spatial analysis for environmental health research: concepts, methods, and examples
.
J Toxicol Environ Health A
2003
;
66
:
1783
810
.
11.
Fornango
RJ
. 
When space matters: spatial dependence, diagnostics, and regression models
.
J Crim Justice Educ
2010
;
21
:
117
35
.
12.
Public Health Agency of Canada
.
Canadian cancer incidence atlas
.
Vol. 2
.
Ottawa (Canada)
:
2000–2006/Centre for Chronic Disease Prevention
; 
2014
. Available from: https://www.canada.ca/en/public-health/services/chronic-diseases/cancer/highlights-canadian-cancer-incidence-atlas-volume-2-2000- 2006.html.
13.
ProQuest LLC
. 
Ulrichsweb
. Available from
: http://ulrichsweb.serialssolutions.com/login
14.
Dawe
DE
,
Singh
H
,
Wickramasinghe
L
,
Pitz
MW
,
Torabi
M
. 
Geographical variation and factors associated with non-small cell lung cancer in manitoba
.
Can Respir J
2017
;
2017
:
1
9
.
15.
Huang
J
,
Wai
ES
,
Lau
F
,
Blood
PA
. 
Palliative radiotherapy utilization for cancer patients at end of life in British Columbia: retrospective cohort study
.
BMC Palliat Care
2014
;
13
:
49
.
16.
Luginaah
IN
,
Gorey
KM
,
Oiamo
TH
,
Tang
KX
,
Holowaty
EJ
,
Hamm
C
, et al
A geographical analysis of breast cancer clustering in southern Ontario: generating hypotheses on environmental influences
.
Int J Environ Health Res
2012
;
22
:
232
48
.
17.
Hystad
P
,
Demers
PA
,
Johnson
KC
,
Brook
J
,
van Donkelaar
A
,
Lamsal
L
, et al
Spatiotemporal air pollution exposure assessment for a Canadian population-based lung cancer case-control study
.
Environ Heal
2012
;
11
:
22
.
18.
Ghazawi
FM
,
Netchiporouk
E
,
Rahme
E
,
Tsang
M
,
Moreau
L
,
Glassman
S
, et al
Comprehensive analysis of cutaneous T-cell lymphoma (CTCL) incidence and mortality in Canada reveals changing trends and geographic clustering for this malignancy
.
Cancer
2017
;
123
:
3550
67
.
19.
R Core Team
. 
R Foundation for Statistical Computing
.
R: a language and environment for statistical computing
.
Vienna, Austria
.
Available from
: https://www.r-project.org.
20.
Statistics Canada
. 
Chapter 12 - Census Geography
. Available from
: https://www12.statcan.gc.ca/census-recensement/2016/ref/98-304/chap12-eng.cfm.
21.
Statistics Canada
. 
Dictionary, Census of Population, 2016 - Postal code
. Available from
: https://www12.statcan.gc.ca/census-recensement/2016/ref/dict/geo035-eng.cfm.
22.
Statistics Canada
. 
Dictionary, Census of Population, 2016
.
Available from
: https://www12.statcan.gc.ca/census-recensement/2016/ref/dict/index-eng.cfm.
23.
Hurvitz
PM
,
Moudon
AV
,
Kang
B
,
Fesinmeyer
MD
,
Saelens
BE
. 
How far from home? The locations of physical activity in an urban U.S. setting
.
Prev Med
2014
;
69
:
181
6
.
24.
Wheeler
D
,
Wang
A
. 
Assessment of residential history generation using a public-record database
.
Int J Environ Res Public Health
2015
;
12
:
11670
82
.
25.
Croon
MA
,
van Veldhoven
MJPM
. 
Predicting group-level outcome variables from variables measured at the individual level: a latent variable multilevel model
.
Psychol Methods
2007
;
12
:
45
57
.
26.
Johnson
GD
. 
Small area mapping of prostate cancer incidence in New York State (USA) using fully Bayesian hierarchical modelling
.
Int J Health Geogr
2004
;
3
:
29
.
27.
Quick
H
,
Tootoo
J
,
Li
R
,
Vaughan
AS
,
Schieb
L
,
Casper
M
, et al
The rate stabilizing tool: generating stable local-level measures of chronic disease
.
Prev Chronic Dis
2019
;
16
:
180442
.
28.
Jarup
L
,
Best
N
,
Toledano
MB
,
Wakefield
J
,
Elliott
P
. 
Geographical epidemiology of prostate cancer in Great Britain
.
Int J Cancer
2002
;
97
:
695
9
.
29.
Boulos
DNK
,
Ghali
RR
,
Ibrahim
EM
,
Boulos
MNK
,
Abdelmalik
P
. 
An eight-year snapshot of geospatial cancer research (2002–2009): clinico-epidemiological and methodological findings and trends
.
Med Oncol
2011
;
28
:
1145
62
.
31.
Hystad
P
,
Demers
PA
,
Johnson
KC
,
Carpiano
RM
,
Brauer
M
. 
Long-term residential exposure to air pollution and lung cancer risk
.
Epidemiology
2013
;
24
:
762
72
.
32.
Winters
N
,
Goldberg
MS
,
Hystad
P
,
Villeneuve
PJ
,
Johnson
KC
,
McCrate
F
, et al
Exposure to ambient air pollution in Canada and the risk of adult leukemia
.
Sci Total Environ
2015
;
526
:
153
76
.
33.
Ritonja
J
,
McIsaac
MA
,
Sanders
E
,
Kyba
CCM
,
Grundy
A
,
Cordina-Duverger
E
, et al
Outdoor light at night at residences and breast cancer risk in Canada
.
Eur J Epidemiol
2020
;
1
:
3
.
34.
Hystad
P
,
Carpiano
RM
,
Demers
PA
,
Johnson
KC
,
Brauer
M
. 
Neighbourhood socioeconomic status and individual lung cancer risk: evaluating long-term exposure measures and mediating mechanisms
.
Soc Sci Med
2013
;
97
:
95
103
.
35.
Pan
SY
,
Morrison
H
,
Gibbons
L
,
Zhou
J
,
Wen
SW
,
DesMeules
M
, et al
Breast cancer risk associated with residential proximity to industrial plants in Canada
.
J Occup Environ Med
2011
;
53
:
522
9
.
36.
Wanigaratne
S
,
Holowaty
E
,
Jiang
H
,
Norwood
TA
,
Pietrusiak
MA
,
Brown
P
. 
Estimating cancer risk in relation to tritium exposure from routine operation of a nuclear-generating station in Pickering, Ontario
.
Chronic Dis Inj Can
2013
;
33
:
247
56
.
37.
Johnson
KC
,
Mao
Y
,
Argo
J
,
Dubois
S
,
Semenciw
R
,
Lava
J
. 
The national enhanced cancer surveillance system: a case-control approach to environment-related cancer surveillance in Canada
.
Environmetrics
1998
;
9
:
495
504
.
38.
Tjepkema
M
,
Christidis
T
,
Bushnik
T
,
Pinault
L
. 
Cohort profile: the Canadian Census Health and Environment Cohorts (CanCHECs)
.
Heal Reports
2019
;
30
:
18
26
.
39.
Statistics Canada
. 
Canadian Population Health Survey data linked to mortality, hospitalization and historical postal codes
. Available from
: https://www.statcan.gc.ca/eng/rdc/data/cphs.
40.
Sanmartin
C
,
Decady
Y
,
Trudeau
R
,
Dasylva
A
,
Tjepkema
M
,
Finès
P
, et al
Linking the Canadian Community Health Survey and the Canadian Mortality Database: an enhanced data source for the study of mortality
.
Health Reports
2016
;
27
:
10
8
.
41.
Saint-Jacques
N
,
Brown
P
,
Nauta
L
,
Boxall
J
,
Parker
L
,
Dummer
TJB
. 
Estimating the risk of bladder and kidney cancer from exposure to low-levels of arsenic in drinking water, Nova Scotia, Canada
.
Environ Int
2018
;
110
:
95
104
.
42.
Singh
H
,
Nugent
Z
,
Decker
K
,
Deniers
A
,
Samaddar
J
,
Torabi
M
. 
Geographic variation and factors associated with colorectal cancer incidence in Manitoba
.
Can J Public Heal
2017
;
108
:
e558
64
.
43.
Brown
P
,
Jiang
H
,
Ezzat
S
,
Sawka
AM
. 
A detailed spatial analysis on contrasting cancer incidence patterns in thyroid and lung cancer in Toronto women
.
BMC Public Health
2016
;
16
:
950
.
44.
Torabi
M
,
Singh
H
,
Galloway
K
,
Israels
SJ
. 
Geographical variation in the incidence of childhood leukaemia in Manitoba
.
J Paediatr Child Health
2015
;
51
:
1121
6
.
45.
Torabi
M
,
Green
C
,
Nugent
Z
,
Mahmud
S
,
Demers
A
,
Griffith
J
, et al
Geographical variation and factors associated with colorectal cancer mortality in a universal health care system
.
Can J Gastroenterol Hepatol
2014
;
28
:
191
7
.
46.
Holowaty
EJ
,
Norwood
TA
,
Wanigaratne
S
,
Abellan
JJ
,
Beale
L
. 
Feasibility and utility of mapping disease risk at the neighbourhood level within a Canadian public health unit: an ecological study
.
Int J Health Geogr
2010
;
9
:
21
.
47.
Saint-Jacques
N
,
Lee
JSW
,
Brown
P
,
Stafford
J
,
Parker
L
,
Dummer
TJB
. 
Small-area spatio-temporal analyses of bladder and kidney cancer risk in Nova Scotia, Canada
.
BMC Public Health
2016
;
16
:
1
17
.
48.
Besag
J
,
York
J
,
Mollie
A
. 
Bayesian image restoration, with two applications in spatial statistics
.
Ann Inst Stat Math
1991
;
43
:
1
20
.
49.
Environmental Systems Research Institute
. 
About ArcGIS | Mapping & Analytics Platform
. Available from
: https://www.esri.com/en-us/arcgis/about-arcgis/overview.
50.
Canadian Partnership Against Cancer
.
First Nations Cancer Control in Canada Baseline Report
.
Toronto (Canada)
:
Canadian Partnership Against Cancer
; 
2013
.
51.
McGahan
CE
,
Linn
K
,
Guno
P
,
Johnson
H
,
Coldman
AJ
,
Spinelli
JJ
, et al
Cancer in First Nations people living in British Columbia, Canada: an analysis of incidence and survival from 1993 to 2010
.
Cancer Causes Control
2017
;
28
:
1105
16
.
52.
Decker
KM
,
Kliewer
EV
,
Demers
AA
,
Fradette
K
,
Biswanger
N
,
Musto
G
, et al
Cancer incidence, mortality, and stage at diagnosis in First Nations living in Manitoba
.
Curr Oncol
2016
;
23
:
225
.
53.
First Nations Information Governance Centre
. 
The First Nations Principles of OCAP®
. Available from
: www.FNIGC.ca/OCAP.
54.
Chan
J
,
Polo
A
,
Zubizarreta
E
,
Bourque
J-M
,
Sheppard
AJ
,
Olson
R
, et al
Disparities in access to RT for regions inhabited by a higher proportion of FNIM populations in Canada
.
Appl Radiat Oncol
2019
;
8
:
34
40
.
55.
Gorey
KM
,
Kliewer
E
,
Holowaty
EJ
,
Laukkanen
E
,
Ng
EY
. 
An international comparison of breast cancer survival: Winnipeg, Manitoba and Des Moines, Iowa, metropolitan areas
.
Ann Epidemiol
2003
;
13
:
32
41
.
56.
Jacquez
GM
,
Essex
A
,
Curtis
A
,
Kohler
B
,
Sherman
R
,
El Emam
K
, et al
Geospatial cryptography: enabling researchers to access private, spatially referenced, human subjects data for cancer control and prevention
.
J Geogr Syst
2017
;
19
:
197
220
.
57.
Statistics Canada
. 
Aboriginal Peoples Highlight Tables, 2016 Census: Aboriginal identity population by both sexes, total - age, 2016 counts, Canada, provinces and territories, 2016 Census–25% Sample data
. Available from
: https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/hlt-fst/abo-aut/Table.cfm?Lang=Eng&T=101&S=99&O=A.
58.
Kamel Boulos
MN
. 
Towards evidence-based, GIS-driven national spatial health information infrastructure and surveillance services in United Kingdom
.
Int J Health Geogr
2004
;
3
:
1
.
59.
Government of British Columbia
. 
Community Health Atlas
. Available from
: http://maps.gov.bc.ca/ess/hm/cha/.
60.
Cancer Care Ontario
. 
Ontario Cancer Profiles - high bandwidth (detailed age groups) - 2018 live dashboard
. Available from
: https://dashboards.instantatlas.com/viewer/report?appid=740454012ef34e59be97991f85ddf498.
61.
CancerControl Alberta
. 
The 2019 Report on Cancer Statistics in Alberta - CancerControl AB | Tableau Public
. Available from
: https://public.tableau.com/profile/cancercontrol.ab#!/vizhome/The2019ReportonCancerStatisticsinAlberta/Highlights?publish=yes.
62.
Jiang
H
,
Brown
PE
,
Rue
H
,
Shimakura
S
. 
Geostatistical survival models for environmental risk assessment with large retrospective cohorts
.
J R Statist Soc A
2014
;
177
:
679
95
.
63.
Torabi
M
,
Rosychuk
RJ
. 
Hierarchical bayesian spatiotemporal analysis of childhood cancer trends
.
Geogr Anal
2012
;
44
:
109
20
.
64.
Torabi
M
,
Rosychuk
RJ
. 
Spatio-temporal modelling using B-spline for disease mapping: analysis of childhood cancer trends
.
J Appl Stat
2011
;
38
:
1769
81
.
65.
Chen
Y
,
Yi
Q
,
Mao
Y
. 
Cluster of liver cancer and immigration: a geographic analysis of incidence data for Ontario 1998–2002
.
Int J Health Geogr
2008
;
7
:
28
.
66.
Van Leeuwen
JA
,
Waltner-Toews
D
,
Abernathy
T
,
Smit
B
,
Shoukri
M
. 
Associations between stomach cancer incidence and drinking water contamination with atrazine and nitrate in Ontario (Canada) agroecosystems, 1987–1991
.
Int J Epidemiol
1999
;
28
:
836
40
.
67.
Walter
SD
. 
The analysis of regional patterns in health data
.
Am J Epidemiol
1992
;
136
:
742
59
.
68.
Ahmed
S
,
Baig
T
,
Iqbal
N
,
Le
D
,
Chalchal
H
,
Zaidi
A
, et al
Effects of social and contextual factors including marital status and children on the use of palliative chemotherapy in metastatic colorectal cancer
.
Am J Clin Oncol
2019
;
42
:
363
6
.
69.
Canale
TD
,
Cho
HK
,
Cheung
WY
. 
A population-based analysis of urban–rural disparities in advanced pancreatic cancer management and outcomes
.
Med Oncol
2018
;
35
:
116
.
70.
McCrate
F
,
Dicks
E
,
Powell
E
,
Chafe
J
,
Roome
R
,
Simmonds
C
, et al
Surgical treatment choices for breast cancer in Newfoundland and Labrador: a retrospective cohort study
.
Can J Surg
2018
;
61
:
377
84
.
71.
Payette
E
,
Sarker
S
,
Chalchal
H
. 
Impact of travel distance on access to treatment and survival in patients with metastatic colorectal cancer prescribed bevacizumab plus chemotherapy
.
Can J Rural Med
2017
;
22
:
148
52
.
72.
Walker
BB
,
Schuurman
N
,
Auluck
A
,
Lear
SA
,
Rosin
M
. 
Socioeconomic disparities in head and neck cancer patients' access to cancer treatment centers
.
Rural Remote Health
2017
;
17
:
4210
.
73.
Liu
E
,
Santibáñez
P
,
Puterman
ML
,
Weber
L
,
Ma
X
,
Sauré
A
, et al
A quantitative analysis of the relationship between radiation therapy use and travel time
.
Int J Radiat Oncol Biol Phys
2015
;
93
:
710
8
.
74.
Zhang
H
,
Dziegielewski
PT
,
Jean Nguyen
TT
,
Jeffery
CC
,
O'Connell
DA
,
Harris
JR
, et al
The effects of geography on survival in patients with oral cavity squamous cell carcinoma
.
Oral Oncol
2015
;
51
:
578
85
.
75.
Helewa
RM
,
Turner
D
,
Wirtzfeld
D
,
Park
J
,
Hochman
DJ
,
Czaykowski
P
, et al
Geographical disparities of rectal cancer local recurrence and outcomes
.
Dis Colon Rectum
2013
;
56
:
850
8
.
76.
Groome
PA
,
Schulze
KM
,
Keller
S
,
Mackillop
WJ
. 
Demographic differences between cancer survivors and those who die quickly of their disease
.
Clin Oncol
2008
;
20
:
647
56
.
77.
Paszat
LF
,
Mackillop
WJ
,
Groome
PA
,
Zhang-Salomons
J
,
Schulze
K
,
Holowaty
E
. 
Radiotherapy for breast cancer in Ontario: rate variation associated with region, age and income
.
Clin Investig Med
1998
;
21
:
125
34
.
78.
Ghazawi
FM
,
Le
M
,
Cyr
J
,
Netchiporouk
E
,
Rahme
E
,
Alakel
A
, et al
Analysis of acute myeloid leukemia incidence and geographic distribution in Canada from 1992 to 2010 reveals disease clusters in Sarnia and other industrial US border cities in Ontario
.
Cancer
2019
;
125
:
1886
97
.
79.
Ye
X
,
Torabi
M
,
Lix
LM
,
Mahmud
SM
. 
Time and spatial trends in lymphoid leukemia and lymphoma incidence and survival among children and adolescents in Manitoba, Canada: 1984–2013
.
PLoS One
2017
;
12
:
e0175701
.
80.
Li
J
,
Adilmagambetov
A
,
Mohomed Jabbar
MS
,
Zaïane
OR
,
Osornio-Vargas
A
,
Wine
O
. 
On discovering co-location patterns in datasets: a case study of pollutants and child cancers
.
Geoinformatica
2016
;
20
:
651
92
.
81.
Lofters
AK
,
Gozdyra
P
,
Lobb
R
. 
Using geographic methods to inform cancer screening interventions for South Asians in Ontario, Canada
.
BMC Public Health
2013
;
13
:
395
.
82.
Milewski
I
. 
Identifying at-risk communities for action on cancer prevention: a case study in new Brunswick (Canada) communities
.
New Solut
2012
;
22
:
79
107
.
83.
Kulkarni
K
,
Stobart
K
,
Witol
A
,
Rosychuk
RJ
. 
Leukemia and lymphoma incidence in children in Alberta, Canada: a population-based 22-year retrospective study
.
Pediatr Hematol Oncol
2011
;
28
:
649
60
.
84.
Torabi
M
,
Rosychuk
RJ
. 
An examination of five spatial disease clustering methodologies for the identification of childhood cancer clusters in Alberta, Canada
.
Spat Spatiotemporal Epidemiol
2011
;
2
:
321
30
.
85.
Rosychuk
RJ
,
Witol
A
,
Stobart
K
. 
Childhood cancer trends in a western Canadian province: a population-based 22-year retrospective study
.
Pediatr Blood Cancer
2010
;
55
:
1348
55
.
86.
Yiannakoulias
N
. 
Using population attributable risk to understand geographic disease clusters
.
Health Place
2009
;
15
:
1142
8
.
87.
Walter
SD
. 
The analysis of regional patterns in health data: II. The power to detect environmental effects
.
Am J Epidemiol
1992
;
136
:
742
59
.
88.
Tabaczynski
A
,
Strom
DA
,
Wong
JN
,
McAuley
E
,
Larsen
K
,
Faulkner
GE
, et al
Demographic, medical, social-cognitive, and environmental correlates of meeting independent and combined physical activity guidelines in kidney cancer survivors
.
Support Care Cancer
2020
;
28
:
43
54
.
89.
Lagacé
F
,
Ghazawi
FM
,
Le
M
,
Rahme
E
,
Savin
E
,
Zubarev
A
, et al
Analysis of incidence, mortality trends, and geographic distribution of breast cancer patients in Canada
.
Breast Cancer Res Treat
2019
;
178
:
683
91
.
90.
Pinault
L
,
Bushnik
T
,
Fioletov
V
,
Peters
CE
,
King
WD
,
Tjepkema
M
. 
The risk of melanoma associated with ambient summer ultraviolet radiation
.
Health Reports
2017
;
28
:
3
11
.
91.
Auluck
A
,
Walker
BB
,
Hislop
G
,
Lear
SA
,
Schuurman
N
,
Rosin
M
. 
Socio-economic deprivation: a significant determinant affecting stage of oral cancer diagnosis and survival
.
BMC Cancer
2016
;
16
:
569
.
92.
Trinh
L
,
Larsen
K
,
Faulkner
GE
,
Plotnikoff
RC
,
Rhodes
RE
,
North
S
, et al
Social-ecological correlates of physical activity in kidney cancer survivors
.
J Cancer Surviv
2016
;
10
:
164
75
.
93.
Labine
M
,
Green
C
,
Mak
G
,
Xue
L
,
Nowatzki
J
,
Griffith
J
, et al
The geographic distribution of liver cancer in Canada does not associate with cyanobacterial toxin exposure
.
Int J Environ Res Public Health
2015
;
12
:
15143
53
.
94.
Auluck
A
,
Walker
BB
,
Hislop
G
,
Lear
SA
,
Schuurman
N
,
Rosin
M
. 
Population-based incidence trends of oropharyngeal and oral cavity cancers by sex among the poorest and underprivileged populations
.
BMC Cancer
2014
;
14
:
316
.
95.
Hwang
E
,
Johnson-Obaseki
S
,
McDonald
JT
,
Connell
C
,
Corsten
M
. 
Incidence of head and neck cancer and socioeconomic status in Canada from 1992 to 2007
.
Oral Oncol
2013
;
49
:
1072
6
.
96.
Anderson
LN
,
Cotterchio
M
,
Kirsh
VA
,
Knight
JA
. 
Ultraviolet sunlight exposure during adolescence and adulthood and breast cancer risk: a population-based case-control study among Ontario women
.
Am J Epidemiol
2011
;
174
:
293
304
.
97.
Zhang-Salomons
J
,
Qian
H
,
Holowaty
E
,
Mackillop
WJ
. 
Associations between socioeconomic status and cancer survival: choice of SES indicator may affect results
.
Ann Epidemiol
2006
;
16
:
521
8
.
98.
Borugian
MJ
,
Spinelli
JJ
,
Mezei
G
,
Wilkins
R
,
Abanto
Z
,
McBride
ML
. 
Childhood leukemia and socioeconomic status in Canada
.
Epidemiology
2005
;
16
:
526
31
.
99.
Gorey
KM
,
Holowaty
EJ
,
Fehringer
G
,
Laukkanen
E
,
Richter
NL
,
Meyer
CM
. 
An International comparison of cancer survival: metropolitan Toronto, Ontario, and Honolulu, Hawaii
.
Am J Public Health
2000
;
90
:
1866
72
.
100.
Mackillop
WJ
,
Zhang-Salomons
J
,
Boyd
CJ
,
Groome
PA
. 
Associations between community income and cancer incidence in Canada and the United States
.
Cancer
2000
;
89
:
901
12
.
101.
Boyd
C
,
Zhang-Salomons
JY
,
Groome
PA
,
Mackillop
WJ
. 
Associations between community income and cancer survival in Ontario, Canada, and the United States
.
J Clin Oncol
1999
;
17
:
2244
.
102.
Goel
V
,
Olivotto
I
,
Gregory Hislop
T
,
Sawka
C
,
Coldman
A
,
Holowaty
EJ
, et al
Patterns of initial management of node-negative breast cancer in two Canadian provinces
.
CMAJ
1997
;
156
:
25
35
.
103.
Larouche
V
,
Toupin
A-K
,
Lalonde
B
,
Simonyan
D
,
Jabado
N
,
Perreault
S
. 
Incidence trends in pediatric central nervous system tumors in Canada: a 15 years report from Cancer and Young People in Canada (CYP-C) registry
.
Neurooncol Adv
2020
;
2
:
vdaa012
.
104.
Darwich
R
,
Ghazawi
FM
,
Rahme
E
,
Alghazawi
N
,
Zubarev
A
,
Moreau
L
, et al
Epidemiology of ophthalmic lymphoma in Canada during 1992–2010
.
Br J Ophthalmol
2019
:
1
5
.
105.
Ghazawi
FM
,
Cyr
J
,
Darwich
R
,
Le
M
,
Rahme
E
,
Moreau
L
, et al
Cutaneous malignant melanoma incidence and mortality trends in Canada: a comprehensive population-based study
.
J Am Acad Dermatol
2019
;
80
:
448
59
.
106.
Ghazawi
FM
,
Le
M
,
Lagacé
F
,
Cyr
J
,
Alghazawi
N
,
Zubarev
A
, et al
Incidence, mortality, and spatiotemporal distribution of cutaneous malignant melanoma cases across Canada
.
J Cutan Med Surg
2019
;
23
:
394
412
.
107.
Ghazawi
FM
,
Darwich
R
,
Le
M
,
Rahme
E
,
Zubarev
A
,
Moreau
L
, et al
Uveal melanoma incidence trends in Canada: a national comprehensive population-based study
.
Br J Ophthalmol
2019
;
103
:
1872
6
.
108.
Le
M
,
Ghazawi
FM
,
Alakel
A
,
Netchiporouk
E
,
Rahme
E
,
Zubarev
A
, et al
Incidence and mortality trends and geographic patterns of follicular lymphoma in Canada
.
Curr Oncol
2019
;
26
:
e473
81
.
109.
Ghazawi
FM
,
Netchiporouk
E
,
Rahme
E
,
Tsang
M
,
Moreau
L
,
Glassman
S
, et al
Distribution and clustering of cutaneous t-cell lymphoma (Ctcl) cases in canada during 1992 to 2010
.
J Cutan Med Surg
2018
;
22
:
154
65
.
110.
Walker
BB
,
Schuurman
N
,
Auluck
A
,
Lear
SA
,
Rosin
M
. 
Suburbanisation of oral cavity cancers: evidence from a geographically-explicit observational study of incidence trends in British Columbia, Canada, 1981–2010
.
BMC Public Health
2015
;
15
:
758
.
111.
Walter
SD
,
Marrett
LD
,
Martin Tnylor
S
,
Derek King
B
. 
An analysis of the geographic variation in cancer incidence and its determinants in Ontario
.
Can J Public Health
1999
;
90
:
104
8
.