Built Environment, Selected Risk Factors and Major Cardiovascular Disease Outcomes: A Systematic Review

1Current global mortality rates from non-communicable diseases (NCDs) remain unacceptably high and are increasing. More than 70% of global cardiovascular disease (CVD), are attributable to modifiable risk factors. Rapidly globalization is accompanied by increasing urbanization, population growth and changes in demographics and promotes trends towards unhealthy lifestyles. The ecological model, however, states that an individual’s behaviour is influenced by multiple level factors such as social, neighborhood environment, and policy factors. One of these factors, the neighborhood environment, and its link to health have been the focus of an increasing number of studies in recent years. These studies are from a variety of disciplines, including urban planning and transportation planning.

2Despite increases in the number of studies on the relationship between the neighborhood environment and health, the potential impact of the neighborhood environment across a range of health outcomes has not been fully explored. For instance, existing studies have focused on specific CVD risk factors such as obesity, metabolic syndrome, physical activity and walking. In addition, a recent study reviewed obesity-related outcomes. Although Mayne et al. 2015 used quasi-experiment in their review, the study centered on obesity and related risk factors. Previously, the association between built environment and obesity has received wide publication. However, no study has broadly reviewed the relationship of neighborhood environment with major CVD outcomes and risk factors, while such a review is necessary to guide future research and policy formulation in this sector. Therefore, the purpose of this study is to synthesize the studies on the association between a number of neighborhood environment attributes and CVD risks.

3A comprehensive search was conducted to identify all research articles published from 2005 to 2015 that examine neighborhood environment, major CVD outcomes and selected risk factors (Table 1). English language articles were identified from the following databases: EBSCO (including: Academic Search, CINAHL, Global Health, Health Source: Nursing/academic and Medline) and Science Direct. Significant studies were identified using any of the following keywords: neighbourhood environment, perceived neighborhood environment, perceived built environment, land use mix diversity, physical activity, social environment, overweight or obesity, hypertension, diabetes mellitus, metabolic syndrome, coronary heart disease and myocardial infarction.

4Titles and abstracts of all identified articles were assessed for their potential eligibility. Full texts of potentially eligible articles were then retrieved and their eligibility was verified against the study eligibility criteria. Fig 1 (a flow chart of included studies; see appendix) represents the flow of the literature review conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, S1 Table (PRISMA 2009 checklist). Studies published in English were included if: 1) they used a Geographic Information System (GIS) or subjectively assessed one or more of the built environment factors categorized according to the validated and reliably tested ‘Neighborhood Environment Walkability Scale’ (NEWS) which is a better questionnaire to assess the local environment; 2) examined the relationship with any of the major CVD outcomes including myocardial infarction, coronary heart disease and stroke; 3) examined selected risk factors including physical activity (categorized in domains were considered), overweight or obesity, hypertension and diabetes mellitus; 4) were original reports on studies conducted among subjects aged 18 years and above; and 5) if the purpose of the studies were to explore the association between the variables of interest using multivariate analyses. Exclusion criteria were as follows: 1) Studies exclusively conducted on adolescents; 2) studies that employed a qualitative design; 3) systematic review papers; 4) publications from studies where subjects had difficulty with walking and 5) studies that did not meet the criteria for current review.

5The information extracted included the first authors’ name, publication year, the sample size, gender, age range of the subjects, country and city where the study was conducted, study design, study tool (assess neighborhood environment), exposure assessment (any of the neighborhood environment attributes), outcome assessment (CVD outcomes or risk factors), and measures of association. Data abstraction, classification, and quality assessment of each study were conducted by two reviewers independently. A third reviewer was consulted if there was disagreement.

6In order to assess the methodological quality for each study selected, the ‘Strengthening the reporting of observational studies in epidemiology’ (STROBE) checklist was adapted in accordance with the objectives of this study. For instance, this included the: sample size, setting, design, study tool (assessing neighborhood environment), exposure, outcome measure and association according to the area of this study. The final PRISMA checklist included I8 items that assessed the quality of this study. Each item scored one point if full reporting was met, or zero if not or partially reported.

7Due to differences in research questions, exposure measurements, outcome measurements and methods across studies, a formal meta-analysis was not possible. Thus, the current review applied a semi-quantitative procedure. The aim of this semi-quantitative procedure was to allow a rapid assessment of the strength of the evidence of an association between the exposure and the outcomes of interest by reducing a range of results from heterogeneous analytical designs to two binary questions: a) did the study under review show a positive or negative association between the built environmental attributes and the outcome of interest? b) and, if so, was this finding statistically significant (p<0.05)? Hence, estimates of associations between neighborhood environment attributes, CVD risk factors and major outcomes were extracted from the eligible studies according to their substantive relevance and methodological findings and results summarized (Table 2). However, to take into account potential publication bias, we did not limit our analysis on papers published in peer-reviewed journals. References of finally included records were additionally checked. Built environment studies assessing relationship with CVD risks and outcomes are relatively recent. Therefore, this study restricted the search for a specific time period and database. Contrary, no quantitative assessment for risk of bias in individual studies was performed. However, in each study sample size, number of observations per built environment and total number of considered CVD risks and outcomes were checked, because small sample sizes result in biased effect estimates.

8An overview of the types of the articles selected is provided in Table 2, highlighting the author, country, gender, age, sample size, study design, study tools (assess neighborhood environment), exposure measures, outcome measures and their associations. The electronic search yielded 565 articles from the selected databases; MasterFile Premier = 118, CINAHL = 71, Science Direct = 323, EBSCO (including; Academic Search, CINAHL, Global Health, Health Source: Nursing/academic, and Medline) = 47, manual search = 6. After title/abstract screening, 525 articles were excluded for not meeting inclusion criteria. Of the excluded articles, 510 articles were unrelated to neighborhood environmental attributes, CVD risk and CVD outcomes, 5 were systematic reviews, 6 were conducted in a population with clinical conditions (disability), and another 4 were duplicates. The abstracts of 40 citations were then obtained and retrieved. Out of these abstracts, 11 were excluded since 4 were qualitative design and 7 were conducted among adolescents. Thus, 29 full text articles were assessed for eligibility. Of these, 11 were excluded as 7 did not use NEWS, 2 were conducted among adolescents and another 2 did not meet the objective of the review to measure BE (S2 Table, excluded articles). Therefore only 18 articles were finally eligible for inclusion in the current review. The flow chart in Fig 1 shows the process leading to the number of included articles for the review.

9Table 2 depicts the descriptive characteristics of the included studies. The year of study ranged between 2005 and 2015, with 27.8% (n = 5) being published in 2012. Sample sizes varied across studies, ranging from 102 to 4,319,674. In all, 55.5% (n = 10) of the studies were conducted in urban areas as compared to rural, suburban and urban/suburban/rural. Community based studies constituted 22.2% (n = 4) compared to one institution based study. The reported ages of the participants ranged from 18 to 80 years. Most studies included females and males (88.9%; n = 16) with only 11.1% (n = 2) being in females only. Sixteen studies (88.9%) were conducted in high-income countries, 11.1% (n = 2) in middle income countries and 38.9% (n = 7) were conducted in the USA alone. Of all included studies, 94.4% (n = 17) were cross-sectional with one being longitudinal.

10Of the 18 studies reviewed, 44.4% focused on physical activity, 16.7% on body mass index, 5.6% on blood pressure, 5.6% on diabetes mellitus and 16.7% on metabolic syndrome. Furthermore, 16.7% of studies focused on coronary heart disease, stroke and heart failure, Table 2.

11The majority of the studies (66.7%) used GIS to assess neighbourhood environment attributes, while 33.3% used NEWS questionnaires (Table 2).

12The majority of the reported associations of neighborhood environmental attributes with CVD risk factors and outcomes were statistically significant (p < 0.05) with effects estimates in the expected direction, and only two studies with mixed results, comparing neighborhood environmental attributes with transport related physical activity and hypertension respectively, reported no significant association, Table 2. Forty four percent of studies reported variety of neighborhood environmental attributes associated with physical activity domains. Conversely, 11.1% of studies reported neighborhood environmental attributes were associated with body mass index and blood pressure. In addition, 16.6% studies reported metabolic syndrome and only one study indicated diabetes mellitus to be related with Built environment attributes. Similarly, 16.6% of studies showed a significant association between neighborhood environmental attributes and myocardial infarction, coronary heart disease, congestive heart failure, angina and stroke, Table 2.

13This review has shown that a variety of neighborhood environmental attributes are associated with physical activity. Furthermore, density of fast food restaurants, supermarkets/grocery stores and high walkable neighborhood environments were associated with body mass index, blood pressure, diabetes mellitus and metabolic syndrome. In addition, high density traffic, road proximity and high density of fast food restaurants were associated with major CVD outcomes.

14Our results are consistent with other studies. In particular, physical activity was associated with safe footpaths and recreational facilities and walking. The results indicate that urban attributes such as street connectivity, residential density, recreational facilities and availability of traffic devices improves neighborhood walkability which may promote walking, leisure and transport related to physical activity which, consequently, lowers the incidence of CVDs. For instance, environmental attributes are thought to increase active transportation and lessen the need for private automobile use to accomplish daily tasks, which, in turn, lowers body mass index.

15This review found that neighborhood environmental attributes such as fast-food restaurants and high walkable neighborhood environment were associated, either positively or negatively with body mass index, blood pressure and metabolic syndrome risk. Previous studies have reported similar results on the association between food environment and BMI or blood pressure. Greater accessibility to fast food restaurants may encourage people to make food choices at odds with ‘healthy’ dietary recommendations by making these choices easier. Another explanation is that limited access to supermarkets may incentivize visits to convenience stores or fast food restaurants outlets thereby increasing the chance of consuming unhealthy foods, with consequential increases in individual body mass indices and blood pressure levels.

16Living in high walkable neighborhoods was associated with a lower prevalence of high body mass index, diabetes mellitus and metabolic syndrome risk. Similar results have been reported elsewhere. Neighbourhood environmental attributes may increase an individual’s active transportation related to the physical activity needed to accomplish daily tasks and thus lower the. For example, a higher population density may support increased recreational opportunities and supermarkets offering a better supply of healthy foods, and so explaining associations between body mass index and metabolic syndrome risk. Moreover, high walkable neighborhood environments are associated with promoting recreational and transport related physical activity, participation in which eventually assists in lowering the prevalence of obesity or metabolic syndrome risks. Furthermore, an increase in intersection density in the neighborhood may promote walking through providing more route options and may regulate traffic.

17Our study also observed that major CVD outcomes are related to built environment attributes. Specifically, a study has reported similar results on proximity to traffic. Environmental attributes include proximity to stores, and access to supermarkets and non-fast food stores which may, consequently, affect the extent to which individuals walk and the food choices they make, which governs their diet and thus links to CVDs. Likewise, high traffic volumes have been associated with noise and air pollution which are linked to major CVDs. In addition, road proximity has been linked with low individual and neighborhood socioeconomic status, both of which have been shown to be associated with CVDs.

18One limitation of this study is the paucity of primary research on the association between neighborhood environmental attributes and CVD risk and major CVDs in an African context. Almost all publications included in the review were cross-sectional, thus causal inferences in the relationships could not be determined. The exclusion of studies not conducted in English also detracts from this study. In addition, this study reviewed few CVD risk factors with selected CVDs. Furthermore, we did not perform meta-analysis to derive pooled estimates of the association across studies. This was due to the much heterogeneity in measures of associations used across included studies, as well as the wide range of outcomes examined across studies. Future studies should explore any association between CVDs and other environmental attributes such as tobacco use, alcohol use and air pollution in order to have a broader understanding of other moderating effects. To our knowledge, this is the first review to document the associations between both objectively and subjectively measured built environment attributes and selected CVD risk and major CVDs. Methods of classification and categorization of the findings in this study follow those of other similar studies, facilitating comparisons. Moreover, this study further contributes to illustrating that studies from developed countries use comparable methodologies to studies from less well developed countries, such as this one.

19This study shows that both objective and perceived neighborhood environmental attributes are linked to CVD and its risk factors. The information gathered here from studies that explored neighborhood environmental attributes and their association with CVD risks and major CVD outcomes will help guide policy makers on the neighborhood environmental, transportation, health and education to improve intervention programs by local government and for people at a ‘grass-roots’ level. Future studies should further explore the associations of CVD risk and CVD outcomes with a broad set of neighborhood attributes using a longitudinal approach to better understand the direction of effects.