International Journal of Health Geographics最新文献

Background: Childrens' outdoor active play is an important part of their development. Play behaviour can be predicted by a variety of physical and social environmental features. Some of these features are difficult to measure with traditional data sources.

Methods: This study investigated the viability of a machine learning method using Google Street View images for measurement of these environmental features. Models to measure natural features, pedestrian traffic, vehicle traffic, bicycle traffic, traffic signals, and sidewalks were developed in one city and tested in another.

Results: The models performed well for features that are time invariant, but poorly for features that change over time, especially when tested outside of the context where they were initially trained.

Conclusion: This method provides a potential automated data source for the development of prediction models for a variety of physical and social environment features using publicly accessible street view images.

Background: In response to citizens' concerns about elevated cancer incidence in their locales, US CDC proposed publishing cancer incidence at sub-county scales. At these scales, confidence in patients' residential geolocation becomes a key constraint of geospatial analysis. To support monitoring cancer incidence in sub-county areas, we presented summary metrics to numerically delimit confidence in residential geolocation.

Results: We defined a concept of Residential Address Discriminant Power (RADP) as theoretically perfect within all residential addresses and its practical application, i.e., using Emergency Dispatch (ED) Address Point Candidates of Equivalent Likelihood (CEL) to quantify Residential Geolocation Discriminant Power (RGDP) to approximate RADP. Leveraging different productivity of probabilistic, deterministic, and interactive geocoding record linkage, we simultaneously detected CEL for 5,807 cancer cases reported to North Carolina Central Cancer Registry (NC CCR)- in January 2022. Batch-match probabilistic and deterministic algorithms matched 86.0% cases to their unique ED address point candidates or a CEL, 4.4% to parcel site address, and 1.4% to street centerline. Interactively geocoded cases were 8.2%. To demonstrate differences in residential geolocation confidence between enumeration areas, we calculated sRGDP for cancer cases by county and assessed the existing uncertainty within the ED data, i.e., identified duplicate addresses (as CEL) for each ED address point in the 2014 version of the NC ED data and calculated ED_sRGDP by county. Both summary RGDP (sRGDP) (0.62-1.00) and ED_sRGDP (0.36-1.00) varied across counties and were lower in rural counties (p < 0.05); sRGDP correlated with ED_sRGDP (r = 0.42, p < 0.001). The discussion covered multiple conceptual and economic issues attendant to quantifying confidence in residential geolocation and presented a set of organizing principles for future work.

Conclusions: Our methodology produces simple metrics - sRGDP - to capture confidence in residential geolocation via leveraging ED address points as CEL. Two facts demonstrate the usefulness of sRGDP as area-based summary metrics: sRGDP variability between counties and the overall lower quality of residential geolocation in rural vs. urban counties. Low sRGDP for the cancer cases within the area of interest helps manage expectations for the uncertainty in cancer incidence data. By supplementing cancer incidence data with sRGDP and ED_sRGDP, CCRs can demonstrate transparency in geocoding success, which may help win citizen trust.

Background: Precise geographical targeting is well recognised as an indispensable intervention strategy for achieving many Sustainable Development Goals (SDGs). This is more cogent for health-related goals such as the reduction of the HIV/AIDS pandemic, which exhibits substantial spatial heterogeneity at various spatial scales (including at microscale levels). Despite the dire data limitations in Low and Middle Income Countries (LMICs), it is essential to produce fine-scale estimates of health-related indicators such as HIV/AIDS. Existing small-area estimates (SAEs) incorporate limited synthesis of the spatial and socio-behavioural aspects of the HIV/AIDS pandemic and/or are not adequately grounded in international indicator frameworks for sustainable development initiatives. They are, therefore, of limited policy-relevance, not least because of their inability to provide necessary fine-scale socio-spatial disaggregation of relevant indicators.

Methods: The current study attempts to overcome these challenges through innovative utilisation of gridded demographic datasets for SAEs as well as the mapping of standard HIV/AIDS indicators in LMICs using spatial microsimulation (SMS).

Results: The result is a spatially enriched synthetic individual-level population of the study area as well as microscale estimates of four standard HIV/AIDS and sexual behaviour indicators. The analysis of these indicators follows similar studies with the added advantage of mapping fine-grained spatial patterns to facilitate precise geographical targeting of relevant interventions. In doing so, the need to explicate socio-spatial variations through proper socioeconomic disaggregation of data is reiterated.

Conclusions: In addition to creating SAEs of standard health-related indicators from disparate multivariate data, the outputs make it possible to establish more robust links (even at individual levels) with other mesoscale models, thereby enabling spatial analytics to be more responsive to evidence-based policymaking in LMICs. It is hoped that international organisations concerned with producing SDG-related indicators for LMICs move towards SAEs of such metrics using methods like SMS.

Background: Communities in the United States (US) exist on a continuum of urbanicity, which may inform how individuals interact with their food environment, and thus modify the relationship between food access and dietary behaviors.

Objective: This cross-sectional study aims to examine the modifying effect of community type in the association between the relative availability of food outlets and dietary inflammation across the US.

Methods: Using baseline data from the REasons for Geographic and Racial Differences in Stroke study (2003-2007), we calculated participants' dietary inflammation score (DIS). Higher DIS indicates greater pro-inflammatory exposure. We defined our exposures as the relative availability of supermarkets and fast-food restaurants (percentage of food outlet type out of all food stores or restaurants, respectively) using street-network buffers around the population-weighted centroid of each participant's census tract. We used 1-, 2-, 6-, and 10-mile (~ 2-, 3-, 10-, and 16 km) buffer sizes for higher density urban, lower density urban, suburban/small town, and rural community types, respectively. Using generalized estimating equations, we estimated the association between relative food outlet availability and DIS, controlling for individual and neighborhood socio-demographics and total food outlets. The percentage of supermarkets and fast-food restaurants were modeled together.

Results: Participants (n = 20,322) were distributed across all community types: higher density urban (16.7%), lower density urban (39.8%), suburban/small town (19.3%), and rural (24.2%). Across all community types, mean DIS was - 0.004 (SD = 2.5; min = - 14.2, max = 9.9). DIS was associated with relative availability of fast-food restaurants, but not supermarkets. Association between fast-food restaurants and DIS varied by community type (P for interaction = 0.02). Increases in the relative availability of fast-food restaurants were associated with higher DIS in suburban/small towns and lower density urban areas (p-values < 0.01); no significant associations were present in higher density urban or rural areas.

Conclusions: The relative availability of fast-food restaurants was associated with higher DIS among participants residing in suburban/small town and lower density urban community types, suggesting that these communities might benefit most from interventions and policies that either promote restaurant diversity or expand healthier food options.

Background: The exponential growth of location-based social media (LBSM) data has ushered in novel prospects for investigating the urban food environment in health geography research. However, previous studies have primarily relied on word dictionaries with a limited number of food words and employed common-sense categorizations to determine the healthiness of those words. To enhance the analysis of the urban food environment using LBSM data, it is crucial to develop a more comprehensive list of food-related words. Within the context, this study delves into the exploration of expanding food-related words along with their associated energy densities.

Methods: This study addresses the aforementioned research gap by introducing a novel methodology for expanding the food-related word dictionary and predicting energy densities. Seed words are generated from official and crowdsourced food composition databases, and new food words are discovered by clustering food words within the word embedding space using the Gaussian mixture model. Machine learning models are employed to predict the energy density classifications of these food words based on their feature vectors. To ensure a thorough exploration of the prediction problem, ten widely used machine learning models are evaluated.

Results: The approach successfully expands the food-related word dictionary and accurately predicts food energy density (reaching 91.62%.). Through a comparison of the newly expanded dictionary with the initial seed words and an analysis of Yelp reviews in the city of Toronto, we observe significant improvements in identifying food words and gaining a deeper understanding of the food environment.

Conclusions: This study proposes a novel method to expand food-related vocabulary and predict the food energy density based on machine learning and word embedding. This method makes a valuable contribution to building a more comprehensive list of food words that can be used in geography and public health studies by mining geotagged social media data.

Background: The aim of the study is to explore the diversity in recreational walking motives across groups with different sociodemographic characteristics, and to use a dynamic and person-centered approach to geographically assess recreational walking behavior, and preferences for place quality related to recreational walking.

Methods: A total of 1838 adult respondents (age 15-90 years), who engage in recreational walking, participated in the map-based survey. We used the online platform Maptionnaire to collect georeferenced information on the respondents' home location, other start locations for walking trips, and point of interest on their trips. Distance between home location and other start locations as well as point of interest were computed using a Geographic Information System (GIS). Additional information on recreational walking behavior and motives were collected using the traditional questionnaire function in Maptionnaire.

Results: The most prevalent motives for walking were mental well-being and physical health, together with enjoyment and experiences related to walking. Having a tertiary education was positively associated with mental well-being motive, experiences, and taking the dog and the children outside. Income was also positively associated with experiences and walking the dog together with enjoyment of walking and spending time with others. Using the map-based approach, we found that recreational walking often starts at a location away from home and is not limited to the nearest neighborhood. A total of 4598 points of interest were mapped, and the most frequently reported place qualities were greenery, water, wildlife, good views, and tranquility.

Conclusion: We used a dynamic and person-centered approach and thereby giving the respondents the opportunity to point to relevant locations for their walking behavior independently of their residential neighborhood. Recreational walking often starts away from home or is not limit to the nearest neighborhod. The median distance from home to the mapped points of interests was between 1.0 and 1.6 km for home-based trips and between 9.4 and 30.6 km for trips with other start locations. The most popular place quality related to the mapped points were greenery, water, wildlife, good views, and tranquility.

Background: Cardiovascular disease (CVD) is increasing in Sub-Saharan Africa (SSA). Overweight/obesity and tobacco use are modifiable CVD risk factors, however literature about the spatiotemporal dynamics of these risk factors in the region at subnational or local scales is lacking. We describe the spatiotemporal trends of overweight/obesity and tobacco use at subnational levels over a 13-year period (2003 to 2016) in five East African nations.

Methods: Cross-sectional, nationally representative Demographic and Health Surveys (DHS) were used to explore the subnational spatiotemporal patterns of overweight/obesity and tobacco use in Burundi, Kenya, Rwanda, Tanzania, and Uganda, five East African Community (EAC) nations with unique cultural landscapes influencing CVD risk factors. Adaptive kernel density estimation and logistic regression were used to determine the spatial distribution and change over time of CVD risk factors on a subnational and subpopulation (rural/urban) scale.

Results: Subnational analysis shows that regional and national level analysis masks important trends in CVD risk factor prevalence. Overweight/obesity and tobacco use trends were not similar: overweight/obesity prevalence increased across most nations included in the study and the inverse was true for tobacco use prevalence. Urban populations in each nation were more likely to be overweight/obese than rural populations, but the magnitude of difference varied widely between nations. Spatial analysis revealed that although the prevalence of overweight/obesity increased over time in both urban and rural populations, the rate of change differed between urban and rural areas. Rural populations were more likely to use tobacco than urban populations, though the likelihood of use varied substantially between nations. Additionally, spatial analysis showed that tobacco use was not evenly distributed across the landscape: tobacco use increased in and around major cities and urban centers but declined in rural areas.

Conclusions: We highlight the importance of de-homogenizing CVD risk factor research in SSA. Studies of national or regional prevalence trends mask important information about subpopulation and place-specific behavior and drivers of risk factor prevalence. Spatially explicit studies should be considered as a vital tool to understand local drivers of health, disease, and associated risk factor trends, especially in highly diverse yet low-resourced, marginalized, and often homogenized regions.

Background: The city of Port-au-Prince, Haiti, is experiencing an epidemic of firearm injuries which has resulted in high burdens of morbidity and mortality. Despite this, little scientific literature exists on the topic. Geospatial research could inform stakeholders and aid in the response to the current firearm injury epidemic. However, traditional small-area geospatial methods are difficult to implement in Port-au-Prince, as the area has limited mapping penetration. Objectives of this study were to evaluate the feasibility of geospatial analysis in Port-au-Prince, to seek to understand specific limitations to geospatial research in this context, and to explore the geospatial epidemiology of firearm injuries in patients presenting to the largest public hospital in Port-au-Prince.

Results: To overcome limited mapping penetration, multiple data sources were combined. Boundaries of informally developed neighborhoods were estimated from the crowd-sourced platform OpenStreetMap using Thiessen polygons. Population counts were obtained from previously published satellite-derived estimates and aggregated to the neighborhood level. Cases of firearm injuries presenting to the largest public hospital in Port-au-Prince from November 22nd, 2019, through December 31st, 2020, were geocoded and aggregated to the neighborhood level. Cluster analysis was performed using Global Moran's I testing, local Moran's I testing, and the SaTScan software. Results demonstrated significant geospatial autocorrelation in the risk of firearm injury within the city. Cluster analysis identified areas of the city with the highest burden of firearm injuries.

Conclusions: By utilizing novel methodology in neighborhood estimation and combining multiple data sources, geospatial research was able to be conducted in Port-au-Prince. Geospatial clusters of firearm injuries were identified, and neighborhood level relative-risk estimates were obtained. While access to neighborhoods experiencing the largest burden of firearm injuries remains restricted, these geospatial methods could continue to inform stakeholder response to the growing burden of firearm injuries in Port-au-Prince.

Background: Some studies have established associations between the prevalence of new-onset asthma and asthma exacerbation and socioeconomic and environmental determinants. However, research remains limited concerning the shape of these associations, the importance of the risk factors, and how these factors vary geographically.

Objective: We aimed (1) to examine ecological associations between asthma prevalence and multiple socio-physical determinants in the United States; and (2) to assess geographic variations in their relative importance.

Methods: Our study design is cross sectional based on county-level data for 2020 across the United States. We obtained self-reported asthma prevalence data of adults aged 18 years or older for each county. We applied conventional and geographically weighted random forest (GWRF) to investigate the associations between asthma prevalence and socioeconomic (e.g., poverty) and environmental determinants (e.g., air pollution and green space). To enhance the interpretability of the GWRF, we (1) assessed the shape of the associations through partial dependence plots, (2) ranked the determinants according to their global importance scores, and (3) mapped the local variable importance spatially.

Results: Of the 3059 counties, the average asthma prevalence was 9.9 (standard deviation ± 0.99). The GWRF outperformed the conventional random forest. We found an indication, for example, that temperature was inversely associated with asthma prevalence, while poverty showed positive associations. The partial dependence plots showed that these associations had a non-linear shape. Ranking the socio-physical environmental factors concerning their global importance showed that smoking prevalence and depression prevalence were most relevant, while green space and limited language were of minor relevance. The local variable importance measures showed striking geographical differences.

Conclusion: Our findings strengthen the evidence that socio-physical environments play a role in explaining asthma prevalence, but their relevance seems to vary geographically. The results are vital for implementing future asthma prevention programs that should be tailor-made for specific areas.

Background: Seasonal floods pose a commonly-recognised barrier to women's access to maternal services, resulting in increased morbidity and mortality. Despite their importance, previous GIS models of healthcare access have not adequately accounted for floods. This study developed new methodologies for incorporating flood depths, velocities, and extents produced with a flood model into network- and raster-based health access models. The methodologies were applied to the Barotse Floodplain to assess flood impact on women's walking access to maternal services and vehicular emergency referrals for a monthly basis between October 2017 and October 2018.

Methods: Information on health facilities were acquired from the Ministry of Health. Population density data on women of reproductive age were obtained from the High Resolution Settlement Layer. Roads were a fusion of OpenStreetMap and data manually delineated from satellite imagery. Monthly information on floodwater depth and velocity were obtained from a flood model for 13-months. Referral driving times between delivery sites and EmOC were calculated with network analysis. Walking times to the nearest maternal services were calculated using a cost-distance algorithm.

Results: The changing distribution of floodwaters impacted the ability of women to reach maternal services. At the peak of the dry season (October 2017), 55%, 19%, and 24% of women had walking access within 2-hrs to their nearest delivery site, EmOC location, and maternity waiting shelter (MWS) respectively. By the flood peak, this dropped to 29%, 14%, and 16%. Complete inaccessibility became stark with 65%, 76%, and 74% unable to access any delivery site, EmOC, and MWS respectively. The percentage of women that could be referred by vehicle to EmOC from a delivery site within an hour also declined from 65% in October 2017 to 23% in March 2018.

Conclusions: Flooding greatly impacted health access, with impacts varying monthly as the floodwave progressed. Additional validation and application to other regions is still needed, however our first results suggest the use of a hydrodynamic model permits a more detailed representation of floodwater impact and there is great potential for generating predictive models which will be necessary to consider climate change impacts on future health access.