Rapid urbanization in Latin America has been one of the contributors of an increased burden of non-communicable diseases (NCDs) [1]. Changes in modes of daily transportation [2] with a major shift from public to individual means (e.g. cars) have also been documented [3]. Both of these trends have played a role in the increasing prevalence of physical inactivity in the region [4]. Studies focusing exclusively on sedentary behavior in Latin America have not been completed. However, a paper by Hallal et al. published on the recent Lancet series on physical activity reported that the overall prevalence of physical inactivity among adults in the Americas was 43.2% [4]. Data from Latin America was limited to Colombia, Ecuador, Brazil, Uruguay, Argentina and Paraguay.

Activity friendly environments have been identified as promising strategies to increase physical activity levels in the population [5-7]. Such understanding is based on evidence from systematic reviews and observational studies looking at the association between perceived environmental features (i.e. presence of sidewalks, proximity to parks, destinations and accessibility etc.) and likelihood of physical activity (e.g. walking for transport or leisure time) [8]. Although the majority of the studies on this topic have a cross-sectional design, previous systematic reviews have found that associations are fairly consistent and show that perceived traffic safety, neighborhood aesthetics, convenience of facilities for walking, accessibility of destinations [9], street scale strategies (relighting of streets, redesigning of streets, and improvements in street aesthetics) [7] are associated with higher physical activity. The limited evidence from low-income and middle-income countries (most of them from upper-middle income) indicates that perceived access to recreation facilities, density of exercise facilities and urbanization are positively associated with physical activity [10]. While association between safety (e.g. crime and traffic) and physical activity are less consistent [10].

Nonetheless, most of the evidence on this issue has been generated in high-income countries [10], thus limiting the generalizability of the findings [11] to other regions with stark differences in urban and social environments such as Latin America [12]. Latin America has one of the highest urbanization rates in the world with 80% of the population living in urban centers [3]. Taking into account this context, associations between perceived environmental attributes and physical activity may vary from those observed in high income countries [11,13,14]. For instance, Latin American cities are characterized by high land-mix use as well as higher density and connectivity [15-17]; these attributes have commonly been found to be associated with physical activity [18]. In addition, there are socio-economic inequalities, cultural aspects, and other contextual differences in the environment (crime rates unsafe, poor access to PA facilities and public transportation predominantly made by busses) unique to low and middle-income countries such as those found in Latin America, where physical activity is partially taken on a utilitarian instead of a recreational purpose. In this sense, people engage in high levels of transportation related physical activity versus leisure time.

Therefore, the objective of this systematic review was to identify which perceived environmental attributes are associated with physical activity in Latin America. Findings from this study could provide useful insights for decision makers and practitioners in order to prioritize limited resources, and have the best chance of success [19,20].

The electronic search yielded 12,963 articles from the selected databases (Pubmed, n = 1,716; Virtual Health Library, n = 2473; EBSCO, n = 30; Web of Science, n = 8744). After title screening, 537 abstracts were reviewed applying the inclusion criteria. Of excluded studies, 115 did not have a physical activity outcome or did not include perceived environment 219 were not conducted in Latin America, 54 were not original research studies (12 systematic reviews or meta-analysis, and 42 reviews), 15 were conducted in populations with clinical conditions, and 119 focused only on adolescents. Finally, 15 studies were included to undergo full abstraction (Figure 1).

The objective of this systematic review was to analyze the current evidence of the association between perceived environment and physical activity in Latin America. The findings highlight that few perceived environment attributes have been studied, and that their association has been explored mainly in two physical activity domains (i.e. Leisure-time and transport related). Furthermore, the studies found are concentrated in only a few countries within the region, namely Brazil and Colombia. Among the examined perceived environment correlates, findings were inconsistent across physical activity domains though more consistency was found for leisure-time as compared to other domains. The only attributes presenting a convincing association with physical activity were safety from crime during the day (leisure-time) and street lightning (transport related).

The convincing association between leisure time physical activity and safety during the day found in this study is similar from evidence gathered through systematic reviews [44,45] and meta-analysis [46] from high-income countries, where perceptions of safety seem to act as a barrier for physical activity. On the contrary for the transport related and walking domains, perceptions of personal and traffic safety were not found to be associated. This is partially explained by the fact that transport related activity represents a need rather than a choice in many countries from Latin America [16], thus is not undermined by perceptions of safety or aesthetics. However, street lighting, which could act as an indirect indicator of personal safety, reported a convincing association with transport related activity. Furthermore, the measures employed to assess safety are derived from high-income countries and they may fail to capture valuable information related to safety in Latin America. Regarding leisure-time physical activity and nearby facilities, classified as suggestive association in the expected direction, the evidence is consistent with findings from two systematic reviews on perceived environment (access to shops [46] and shops are in walking distance [47]) and physical activity. Similar findings were reported in a meta-analysis by Duncan et al. [48] in which shops and services within walking distance were associated with physical activity.

One of the main contributions from the present study is that it studies and stratifies associations according to physical activity domains, there is evidence that particular characteristics of built and perceived environments are more relevant for some domains versus others, and may also vary accordingly [32,49,50]. A recent systematic review found that the association between several environmental variables varied by domain, particularly between leisure-time and transport related activity, which is similar to the findings from this review [32]. Prior recommendations for the region of Latin America have suggested the use of the leisure-time and transportation sections of IPAQ as they have shown higher reliability and validity among this population [51].

Another finding of the study is that many of the highest quality studies found null associations (see Table 2). This may suggest that many of the tools and instruments used are not capturing the environmental perceptions that matter most for physical activity in Latin America. The lack of consistency on findings may be due to the small number of selected studies found in the electronic search, which indicates the lack of evidence in this area in Latin America. This area of research is relatively new in the region and only until recently have studies begun to emerge and few of them make it to the indexed literature, possibly due to lack of capacity and resources in the region [52]. Within region variability of the perceived environment context may also help to explain this inconsistency, although this is a less likely explanation, as most countries from Latin America follow similar urban designs from those of early European settlements. For instance, the higher density of Latin American cities, resulting in more proximity and accessibility to destinations, could positively influence physical activity, but the perception of personal and traffic safety hinder this association [53]. In addition, higher levels of density and connectivity usually found in Latin American cities, limits variability in study measures and the possibility of finding significant associations. The social and economic characteristics of the population from Latin America may also mediate and moderate the association between perceived environment and physical activity. It has been reported that low income populations are more likely to engage in transport related physical activity (meeting public health recommendations of at least 150 minutes per week) versus leisure time physical activity [54,55]. In large cities from Latin America the use of public transportation is more of a necessity rather than an option, lower rates of household car ownership (ranging between 20% and 80%), high prices of gas, and high accessibility to public transportation infrastructure, help explain this phenomenon. Findings may also vary by gender and according to the quality of some environmental attributes.

To our knowledge, this is the first systematic review documenting the associations between perceived environment and physical activity in Latin America and the first focusing upon findings from developing countries. The methodology used to classify and categorize the evidence has been used in similar studies, which allows establishing comparisons with other regions from the world. In addition, this approach contributes to the potential generalization of recommendations and the use of this methodology in future studies in different regions of the globe, particularly in developing countries.

Some limitations of this review need to be acknowledged. First, all 15 articles included in the review were cross-sectional studies, thus causal inference in the relationship between perceived environment and physical activity cannot be evaluated. However, the review by the US Community Guide faced the same challenge [5]. Because most of the selected studies reported only adjusted associations, the analysis was focused only on the adjusted results in order to allow consistency. This might have biased the results, however, studies adjusted their results by SES, age and education, which are common confounders in the association between perceived environment and physical activity. Prospective studies of environmental factors and physical activity are needed as well as evidence from intervention studies in order to advocate for policy changes and large-scale environmental interventions. This review has the usual limitations derived from perceptions about the environment, which may differ from objective measures, and is more sensitive to reporting bias. Lastly, most of the instruments used to assess perceived environment characteristics are derived from research conducted in North America, Australia and Europe and may not be sensible to key environmental features within the socio-cultural environment from Latin America. The majority of the studies were originated from Brazil followed by Colombia, with no existing evidence from other Latin American countries, in part due to the infancy of the field in Latin America [51,52]. However, the built environment from many Latin American countries has followed similar urbanization patterns and development as developed countries [56], and they also share common socio-cultural characteristics, making them more homogenous, which could facilitate the extrapolation of findings. A summary net effect was limited considering the wide variety of outcome measures used in the selected studies, situation that has been observed in other systematic reviews [57].

This study shows that, perceived environmental attributes and their relationship with physical activity appears to be domain, and context specific. In addition, findings from this study show inconsistencies with the information gathered from high-income countries, specifically as it relates to personal safety, while others are in accordance and can be used to advance a common agenda for the design of activity friendly environments. However, the lack of associations or the lack of studies in many of the perceived environment characteristics shows a gap in the literature. The information gathered through studies exploring perceived environment associations with physical activity can help guide the decision-making process of built environment, transportation, health and education agendas.

The author(s) declare that they have no competing interests.

The study was conceived by CA, DC-Pa, and RB. Articles search and data extractions were made by CA and DC-Pz. Quality assessment by CA, DC-Pz, and RR. RB, RR and DC-Pa reviewed the article critically for important intellectual content. All authors were involved in writing the paper and had final approval of the submitted and published versions.