Community Health and Epidemiology

Permanent URI for this collection

https://hdl.handle.net/10388/16193

Browse

Now showing 1 - 20 of 47

How Municipal Actors Leverage Evidence to Support Urban Planning: Perspectives from the City of Regina
(University of Toronto Press, 2024-11-14) Mahani, Akram; Lyeo, Joonsoo Sean; Fung, Agnes; Husack, Kelly; Muhajarine, Nazeem; Diener, Tania; Brown, Chelsea
Despite growing calls for evidence-informed policy-making, policy- and decision-makers in municipal governments often face barriers to the integration of evidence into their work. These barriers may be especially prominent in small-to-medium-sized cities, which are often comparatively resource-strained relative to the larger cities that are the focus of much of the existing literature. We interviewed 30 municipal actors involved in urban planning at the City of Regina—the municipal government representing the medium-sized city of Regina located in western Canada. Our study revealed that evidence was often positioned differently among policy- and decision-makers. Interviewees had different definitions of evidence, accessed various sources of evidence, placed differing levels of trust in certain forms of evidence, and experienced different barriers to the integration of evidence into their day-to-day work. This article contributes to a better understanding of the role of evidence among urban planners and has important lessons on how to bridge the gap between urban planning policy- and decision-makers and producers of evidence.
Evaluation of an Interprofessional Educational Intervention in Mental Health and Intellectual and Developmental Disability for Health and Social Service Trainees
(Springer Nature, 2024-11) Tarzi, Gabriel; Thakur, Anupam; Bobbette, Nicole; Pilatzke, Megan; Lefkowitz, Gill; Thomson, Kendra; Thatcher, Alicia; Hasan, Syeda; Fogle, Adeen; Blake, Marissa; Hines, Ann; Lunsky, Yona
Objective Adults with intellectual and developmental disabilities (IDD) experience high rates of poor mental health and challenges accessing timely and high-quality services. There is limited interprofessional training on mental health care for this population. Methods A virtual, synchronous program based on the Project Extension for Community Healthcare Outcomes (ECHO) Ontario IDD Mental Health program was developed for health and social service trainees. Participants represented 10 disciplines across 12 Canadian university or college programs. The program was taught by a team of health and social service providers together with individuals with lived experience and included didactics and case-based discussions. Program evaluation utilized a pre-, post-, and 12-week follow-up survey design with feedback surveys following each session. Results Fifty participants registered for the program; 34 (68%) completed baseline measures and attended at least two sessions. Overall, participants reported high session satisfaction (average rating of 4.47 of 5). Participants demonstrated significant improvement in self-efficacy regarding communication (p < 0.001), management of mental health needs (p < 0.001), and working across systems (p < 0.001). Participants self-reported feeling more knowledgeable about common comorbidities (p < 0.001), assessing behavioral challenges (p < 0.001), the role of interdisciplinary professionals (p < 0.001), and community resources (p < 0.001). Improvements were maintained at follow-up across outcomes. Conclusion The pilot Project ECHO for health and social service trainees in adult IDD mental health demonstrated high participant satisfaction and positive impact on trainees’ self-efficacy and knowledge. Interprofessional educational interventions can be effectively implemented using virtual technology to teach about other mental health populations requiring multisector care.
Barriers to integration of health and equity into urban design policies in Regina, Saskatchewan
(Oxford University Press, 2024-12-06) Mahani, Akram; Lyeo, Joonsoo Sean; Fung, Agnes; Husack, Kelly; Muhajarine, Nazeem; Diener, Tania; Brown, Chelsea
Although there is extensive literature on the impact of urban design on health, little is known about the barriers to integrating health into urban design policies. As cities increasingly lead efforts to improve health equity and population health, understanding the perspectives and experiences of municipal actors on health and equity is essential. To address this gap, we conducted semi-structured interviews with 30 stakeholders engaged with urban design policy- and decision-making at the City of Regina in Saskatchewan, Canada. We analysed our data using a qualitative thematic framework. Our research uncovered a lack of shared understanding of health among municipal actors. Interviewees identified several barriers to integrating health and equity in urban design policies, including inaccessibility of evidence; insufficient resourcing; fragmented governance structure; limited legal power of local governments in Canada; a deeply ingrained culture of individualism and lack of representation. Our findings underscore the importance of adopting an integrated and holistic approach for healthy and equitable urban design. As urbanization continues to bring a greater share of the world’s population into urban areas, it is crucial to understand how municipal governance can foster environments that promote residents’ well-being.
Health and equity considerations in policy and practice related to sustainable transportation interventions in four Canadian cities
(Elsevier, 2025-03) Verity, Wendy; Stephens, Zoe Poirier; Lin, Helena; Ottoni, Callista A.; Bourgeois, Eve; Kestens, Yan; Fuller, Daniel; Manaugh, Kevin; Winters, Meghan
Cities around the world are making major investments in sustainable transportation infrastructure in response to diverse societal challenges. These projects have the potential to advance healthy cities, but it is unknown to what extent municipalities incorporate health and equity goals into sustainable transportation policy and practice. We provide an empirical analysis of how health and equity were understood, operationalized, and measured in the planning, design, and implementation of four sustainable transportation projects in four Canadian cities, Victoria, Vancouver, Saskatoon, and Montreal, as part of the INTerventions, Equity, Research, and Action in Cities Team (INTERACT) project. We reviewed nine policy documents that guided the sustainable transportation interventions and conducted content analysis to understand how health and equity were referred to within public-facing policy documents. We also interviewed 23 city staff, councillors, or members of advocacy groups who were involved in the planning, design, and implementation of these interventions and conducted framework analysis on the transcripts. We developed themes from integrated findings of our policy scan and interviews with practitioners. We found that health was broadly used in the policy framing, and that safety and modal shift towards active modes were key health outcomes in both policy and practice. Other health outcomes were described as co-benefits but were minimally operationalized and had limited influence on the prioritization of municipal investments. Equity priorities in policies and practice centered on design and implementation features geared towards equity-deserving groups and accessibility for all ages and abilities. Participants emphasized the importance of policy and political will for advancing equity goals and navigating trade-offs between prioritizing high-use or equity-focused routes. Participants also described challenges with measuring health (due to institutional mandates) and equity (due to lack of disaggregated data). These findings highlight the opportunities for clearer direction in policy and practice in relation to health and equity goals in sustainable transportation interventions.
Perspective on strategies for matching across age and sex in physiology research: “recreationally active” is not good enough
(Frontiers Media SA, 2025-01-06) Schweitzer, Allyson M.; Fuller, Daniel; Fliss, Matthew D.; Mitchell, Cameron J.
Cardiorespiratory fitness (CRF) and muscular fitness are powerful confounders in age and sex-related comparisons. This paper provides a perspective on the benefits and limitations of matching participants by physical activity behaviour, objectively measured fitness and normative fitness percentiles. Data presented herein are a subset of a larger study, and highlight that matching by physical activity, does not necessarily match on other metrics like physical fitness, especially when age-related comparisons are being made. Our data showed that young and older adults matched by physical activity behaviours showed the expected higher CRF and muscular fitness in male and younger participants, but older adults had higher CRF percentiles. This suggests that matching by physical activity behaviour may select older adults with relatively higher CRF. Researchers must choose their matching method carefully to ensure the appropriate aspects of fitness have been matched between groups. For clarity, they should also report when certain aspects of fitness have not been accounted for and give an explanation as to why.
Urban vegetation and well-being: A cross-sectional study in Montreal, Canada
(Wiley, 2025-01-13) Sousa-Silva, Rita; Kestens, Yan; Stephens, Zoé Poirier; Thierry, Benoit; Schoenig, Daniel; Fuller, Daniel; Winters, Meghan; Smargiassi, Audrey
1. As urbanisation continues to accelerate, urban green spaces are increasingly recognised as key elements for enhancing people's health and well-being. However, most research has used vegetation metrics that may not capture the specific associations between different types of vegetation and different mental health outcomes. 2. In this study, we investigate the cross-sectional associations between residential vegetation exposure and individual well-being in Montreal, Canada, using different vegetation and well-being measures: The proportion of grass cover, tree cover, and average NDVI value within buffers of various radii (100–1000 m) were linked to each participant's residence (n = 1072, aged 18 years or older), while well-being was assessed using subjective happiness, emotional well-being, and personal well-being scales. The associations were analysed using generalised additive regression models. 3. Our findings show that more vegetation was linked to enhanced well-being, although the effect sizes were relatively small. Irrespective of the buffer distance, the positive associations for grass and NDVI were more pronounced than those for trees, though these associations varied across the different well-being outcome measures. We also observed that increasing tree coverage has a stronger positive effect on the well-being of individuals who are dissatisfied with the current number of street trees. 4. Synthesis and applications. Everyday exposure to nearby nature is associated with better self-reported mental health, suggesting urban greening policies should focus on including more vegetation within built spaces, from individual street trees to small and large parks. Our study also highlights the importance of distinguishing between different types of vegetation (e.g. grass vs. trees) when studying the effects of vegetation on well-being or other health-related outcomes. Likewise, using different measures of well-being may provide a more nuanced and comprehensive understanding of how vegetation impacts people's well-being.
The promises and perils of a free rural inter-city transportation scheme: A mixed-methods study from Northern Saskatchewan
(Springer Nature, 2025-02-05) Alhassan, Jacob Albin Korem; Fuller, Daniel; Woytowich, Ron
Objective Transportation is a critical health determinant, yet the last decade has witnessed rapid disinvestment across Canada (particularly in rural contexts) with negative health consequences. We sought to explore and describe the benefits and challenges faced in operating the first community-driven free-transportation scheme in Saskatchewan that emerged in response to widespread unavailability of public transportation due to budget cuts (austerity). Methods We conducted a mixed-methods community-based participatory research study involving 22 interviews with bus riders and service administrators. We also performed descriptive statistics and chi-squared analyses on bus rider data (data on 1185 trips routinely collected between July 2023 and December 2023) to explore sociodemographic characteristics and trip purposes of bus riders. Results All trips were completed by 616 community members using the free bus service between July 2023 and December 2023. Community members took an average of 5 trips (median = 2.0) with a maximum of 22 trips being taken by one community member (1.9% of all trips). Most trips were by women (53%), and older adults mostly used the free bus for medical purposes (22% of riders were older adults and 34% of these used the bus for medical reasons). Qualitatively, the bus service has increased access to care and promotes social participation and autonomy, especially for older adults. The service however faces some challenges, including funding disruptions and difficulty recruiting and retaining drivers. Conclusion Free inter-community transportation (i.e. transportation across cities and municipalities) promotes health equity and access. In contexts without access to public transportation, governments could support community-driven initiatives through increased funding.
Unpacking the Mood States of Children and Youth in Saskatchewan, Canada, in the Context of the COVID-19 Pandemic: Insights from the “See Us, Hear Us 2.0” Study
(Children, 2025-01-10) Shakurun, Nahin; Hinz, Tamara; Adeyinka, Daniel Adedayo; Muhajarine, Nazeem
Background/Objectives: The COVID-19 pandemic created a growing need for insights into the mental health of children and youth and their use of coping mechanisms during this period. We assessed mood symptoms and related factors among children and youth in Saskatchewan. We examined if coping abilities mediated the relationship between risk factors and mood states. Methods: “See Us, Hear Us 2.0”, a cross-sectional study of 563 child–parent dyads, provided the data. The dependent variable, current mood state, was measured by the CoRonavIruS health Impact Survey (CRISIS) scale. Independent variables included sociodemographics, behaviors, household conditions, and coping ability. Multiple linear regression and mediation analyses were conducted, ensuring sample representativeness with sampling weights and addressing missing data through multiple imputations. Results: The participants reported mood symptoms (“moderate” to “extreme”) ranging from 23% to 38% on the CRISIS scale. Factors such as older children, hybrid learning, disrupted activities, and increased screen time worsened moods. The ethnic minority groups (BIPOC) living in mid-sized cities/towns experienced more negative moods compared to Whites residing in cities. Coping ability mediated the relationship between extracurricular activities and mood states. Conclusions: Our results underscore the importance of tailored interventions, recognizing the diverse needs of specific age groups, gender identities, and ethnicities and addressing the adverse effects of the pandemic- related disruptions on the mental health and well-being of school children in Saskatchewan. Our study also suggests prioritizing the diverse needs of children and youth during the planning and implementation of mental health services in the province.
ALF–Score—A novel approach to build a predictive network–based walkability scoring system
(PLoS One, 2022) S. Alfosool, Ali M.; Chen, Yuanzhu; Fuller, Daniel
Walkability is a term that describes various aspects of the built and social environment and has been associated with physical activity and public health. Walkability is subjective and although multiple definitions of walkability exist, there is no single agreed upon definition. Road networks are integral parts of mobility and should be an important part of walkability. However, using the road structure as nodes is not widely discussed in existing methods. Most walkability measures only provide area–based scores with low spatial resolution, have a one–size–fits–all approach, and do not consider individuals opinion. Active Living Feature Score (ALF–Score) is a network–based walkability measure that incorporates road network structures as a core component. It also utilizes user opinion to build a high–confidence ground–truth that is used in our machine learning pipeline to generate models capable of estimating walkability. We found combination of network features with road embedding and points of interest features creates a complimentary feature set enabling us to train our models with an accuracy of over 87% while maintaining a conversion consistency of over 98%. Our proposed approach outperforms existing measures by introducing a novel method to estimate walkability scores that are representative of users opinion with a high spatial resolution, for any point on the road.
Predicting lying, sitting and walking at different intensities using smartphone accelerometers at three different wear locations: hands, pant pockets, backpack
(BMJ Open Sport & Exercise Medicine, 2022-04) Khataeipour, Seyed Javad; Anaraki, Javad Rahimipour; Bozorgi, Arastoo; Rayner, Machal; Basset, Fabien A; Fuller, Daniel
Objective This study uses machine learning (ML) to develop methods for estimating activity type/intensity using smartphones, to evaluate the accuracy of these models for classifying activity, and to evaluate differences in accuracy between three different wear locations. Method Forty-eight participants were recruited to complete a series of activities while carrying Samsung phones in three different locations: backpack, right hand and right pocket. They were asked to sit, lie down, walk and run three Metabolic Equivalent Task (METs), five METs and at seven METs. Raw accelerometer data were collected. We used the R, activity counts package, to calculate activity counts and generated new features based on the raw accelerometer data. We evaluated and compared several ML algorithms; Random Forest (RF), Support Vector Machine, Naïve Bayes, Decision Tree, Linear Discriminant Analysis and k-Nearest Neighbours using the caret package (V.6.0–86). Using the combination of the raw accelerometer data and the computed features leads to high model accuracy. Results Using raw accelerometer data, RF models achieved an accuracy of 92.90% for the right pocket location, 89% for the right hand location and 90.8% for the backpack location. Using activity counts, RF models achieved an accuracy of 51.4% for the right pocket location, 48.5% for the right hand location and 52.1% for the backpack location. Conclusion Our results suggest that using smartphones to measure physical activity is accurate for estimating activity type/intensity and ML methods, such as RF with feature engineering techniques can accurately classify physical activity intensity levels in laboratory settings.
Study protocol for CELLAR (COVID-related Eating Limitations and Latent dietary effects in the Atlantic Region): population-based observational study to monitor dietary intakes and purchasing during COVID-19 in four Atlantic Canadian provinces
(BMJ Open, 2022) Mah, Catherine L; Foster, Karen; Jago, Emily; Hajizadeh, Mohammad; Luongo, Gabriella; Taylor, Nathan; Fuller, Daniel; Yi, Yanqing; Esan, Olukorede T; Lukic, Ryan; Clarke, Maria; Wranik, Wiesława Dominika; Brimblecombe, Julie Kay; Peeters, Anna
Introduction Poor diet is a leading preventable risk for the global burden of non-communicable disease. Robust measurement is needed to determine the effect of COVID-19 on dietary intakes and consumer purchasing, given the widespread changes to consumer food environments and economic precarity. The research objectives are as follows: (1) describe dietary intakes of foods, beverages and nutrients of concern during the COVID-19 pandemic; (2) quantify change in diet during COVID-19 as compared with prepandemic, previously captured in the provincial samples of the population-representative 2015 Canadian Community Health Survey-Nutrition and (3) examine how household purchasing practices predict dietary intakes during COVID-19. Methods and analyses Observational study of diet, using a population-based stratified probability sampling strategy allocated via dual-frame (landline and cellphone) calls to random-digit dialled numbers, followed by age-sex group quotas. The base population comprises the four provinces of the Atlantic region of Canada, jurisdictions with an excess burden of pre-existing dietary risk, compared with the rest of Canada. Our aim is n=1000 to obtain reliable estimates at a regional level to describe intakes and compare with prepandemic baseline. Data collection entails 12 weeks participation: (1) enrolment with sociodemographics (key dietary risk predictors such as age, sex, gender, pre-COVID-19 income, employment, household composition, receipt of economic relief, rural residence); (2) two 24hour diet recalls using the online ASA-24 Canada 2018 tool; and (3) online uploads of household food purchase receipts over the 12 weeks enrolled. Participation incentives will be offered.
Why machine learning (ML) has failed physical activity research and how we can improve
(BMJ Open Sport & Exercise Medicine, 2022) Fuller, Daniel; Ferber, Reed; Stanley, Kevin
Measuring physical activity is a critical issue for our understanding of the health benefits of human movement. Machine learning (ML), using accelerometer data, has become a common way to measure physical activity. ML has failed physical activity measurement research in four important ways. First, as a field, physical activity researchers have not adopted and used principles from computer science. Benchmark datasets are common in computer science and allow the direct comparison of different ML approaches. Access to and development of benchmark datasets are critical components in advancing ML for physical activity. Second, the priority of methods development focused on ML has created blind spots in physical activity measurement. Methods, other than cut-point approaches, may be sufficient or superior to ML but these are not prioritised in our research. Third, while ML methods are common in published papers, their integration with software is rare. Physical activity researchers must continue developing and integrating ML methods into software to be fully adopted by applied researchers in the discipline. Finally, training continues to limit the uptake of ML in applied physical activity research. We must improve the development, integration and use of software that allows for ML methods’ broad training and application in the field.
Using combined Global Position System and accelerometer data points to examine how built environments and gentrification are associated with physical activity in four Canadian cities
(International Journal of Behavioral Nutrition and Physical Activity, 2022) Firth, Caislin; Kestens, Yan; Winters, Meghan; Stanley, Kevin; Bell, Scott; Thierry, Benoit; Phillips, Kole; Poirier-Stephens, Zoe; Fuller, Daniel
Background Built and social environments are associated with physical activity. Global Positioning Systems (GPS) and accelerometer data can capture how people move through their environments and provide promising tools to better understand associations between environmental characteristics and physical activity. The purpose of this study is to examine the associations between GPS-derived exposure to built environment and gentrification characteristics and accelerometer-measured physical activity in a sample of adults across four cities. Methods We used wave 1 data from the Interventions, Research, and Action in Cities Team, a cohort of adults living in the Canadian cities of Victoria, Vancouver, Saskatoon, and Montreal. A subsample of participants wore a SenseDoc device for 10 days during May 2017–January 2019 to record GPS and accelerometry data. Two physical activity outcomes were derived from SenseDoc data: time spent in light, moderate, and vigorous physical activity; and time spent in moderate or vigorous physical activity. Using corresponding GPS coordinates, we summarized physical activity outcomes by dissemination area—a Canadian census geography that represents areas where 400 to 700 people live- and joined to built (active living space, proximity to amenities, and urban compactness) and gentrification measures. We examined the associations between environmental measures and physical activity outcomes using multi-level negative binomial regression models that were stratified by city and adjusted for covariates (weekday/weekend), home dissemination area, precipitation, temperature) and participant-level characteristics obtained from a survey (age, gender, income, race). Results We found that adults spent more time being physically active near their homes, and in environments that were more walkable and near parks and less time in urban compact areas, regardless of where participants lived. Our analysis also highlighted how proximity to different amenities was linked to physical activity across different cities. Conclusions Our study provides insights into how built environment and gentrification characteristics are associated with the amount of time adults spend being physically active in four Canadian cities. These findings enhance our understanding of the influence that environments have on physical activity over time and space, and can support policies to increase physical activity.
Recruiting Participants for Population Health Intervention Research: Effectiveness and Costs of Recruitment Methods for a Cohort Study
(Journal of Medical Internet Research, 2021) Wasfi, Rania; Poirier Stephens, Zoe; Sones, Meridith; Laberee, Karen; Pugh, Caitlin; Fuller, Daniel; Winters, Meghan; Kestens, Yan
Background: Public health research studies often rely on population-based participation and draw on various recruitment methods to establish samples. Increasingly, researchers are turning to web-based recruitment tools. However, few studies detail traditional and web-based recruitment efforts in terms of costs and potential biases. Objective: This study aims to report on and evaluate the cost-effectiveness, time effectiveness, and sociodemographic representation of diverse recruitment methods used to enroll participants in 3 cities of the Interventions, Research, and Action in Cities Team (INTERACT) study, a cohort study conducted in Canadian cities. Methods: Over 2017 and 2018 in Vancouver, Saskatoon, and Montreal, the INTERACT study used the following recruitment methods: mailed letters, social media (including sponsored Facebook advertisements), news media, partner communications, snowball recruitment, in-person recruitment, and posters. Participation in the study involved answering web-based questionnaires (at minimum), activating a smartphone app to share sensor data, and wearing a device for mobility and physical activity monitoring. We describe sociodemographic characteristics by the recruitment method and analyze performance indicators, including cost, completion rate, and time effectiveness. Effectiveness included calculating cost per completer (ie, a participant who completed at least one questionnaire), the completion rate of a health questionnaire, and the delay between completion of eligibility and health questionnaires. Cost included producing materials (ie, printing costs), transmitting recruitment messages (ie, mailing list rental, postage, and sponsored Facebook posts charges), and staff time. In Montreal, the largest INTERACT sample, we modeled the number of daily recruits through generalized linear models accounting for the distributed lagged effects of recruitment campaigns. Results: Overall, 1791 participants were recruited from 3 cities and completed at least one questionnaire: 318 in Vancouver, 315 in Saskatoon, and 1158 in Montreal. In all cities, most participants chose to participate fully (questionnaires, apps, and devices). The costs associated with a completed participant varied across recruitment methods and by city. Facebook advertisements generated the most recruits (n=687), at a cost of CAD $15.04 (US $11.57; including staff time) per completer. Mailed letters were the costliest, at CAD $108.30 (US $83.3) per completer but served to reach older participants. All methods resulted in a gender imbalance, with women participating more, specifically with social media. Partner newsletters resulted in the participation of younger adults and were cost-efficient (CAD $5.16 [US $3.97] per completer). A generalized linear model for daily Montreal recruitment identified 2-day lag effects on most recruitment methods, except for the snowball campaign (4 days), letters (15 days), and reminder cards (5 days). Conclusions: This study presents comprehensive data on the costs, effectiveness, and bias of population recruitment in a cohort study in 3 Canadian cities. More comprehensive documentation and reporting of recruitment efforts across studies are needed to improve our capacity to conduct inclusive intervention research.
Wave 1 results of the INTerventions, Research, and Action in Cities Team (INTERACT) cohort study: Examining spatio-temporal measures for urban environments and health
(Health & Place, 2021) Fuller, Daniel; Bell, Scott; Firth, Caislin; Nazeem, Muhajarine; Nelson, Trisalyn; Stanley, Kevin; Sones, Meridith; Smith, Jefferson; Thierry, Benoit; Laberee, Karen; Stephens, Zoe Poirier; Phillips, Kole; Kestens, Yan; Winters, Meghan
Built environment interventions have the potential to improve population health and reduce health inequities. The objective of this paper is to present the first wave of the INTErventions, Research, and Action in Cities Team (INTERACT) cohort studies in Victoria, Vancouver, Saskatoon, and Montreal, Canada. We examine how our cohorts compared to Canadian census data and present summary data for our outcomes of interest (physical activity, well-being, and social connectedness). We also compare location data and activity spaces from survey data, research-grade GPS and accelerometer devices, and a smartphone app, and compile measures of proximity to select built environment interventions.
Exploring Orientation Invariant Heuristic Features with Variant Window Length of 1D-CNN-LSTM in Human Activity Recognition
(Biosensors, 2022-07) Barua, Arnab; Fuller, Daniel; Musa, Sumayyah; Jiang, Xianta
Many studies have explored divergent deep neural networks in human activity recognition (HAR) using a single accelerometer sensor. Multiple types of deep neural networks, such as convolutional neural networks (CNN), long short-term memory (LSTM), or their hybridization (CNN-LSTM), have been implemented. However, the sensor orientation problem poses challenges in HAR, and the length of windows as inputs for the deep neural networks has mostly been adopted arbitrarily. This paper explores the effect of window lengths with orientation invariant heuristic features on the performance of 1D-CNN-LSTM in recognizing six human activities; sitting, lying, walking and running at three different speeds using data from an accelerometer sensor encapsulated into a smartphone. Forty-two participants performed the six mentioned activities by keeping smartphones in their pants pockets with arbitrary orientation. We conducted an inter-participant evaluation using 1D-CNN-LSTM architecture. We found that the average accuracy of the classifier was saturated to 80 ± 8.07% for window lengths greater than 65 using only four selected simple orientation invariant heuristic features. In addition, precision, recall and F1-measure in recognizing stationary activities such as sitting and lying decreased with increment of window length, whereas we encountered an increment in recognizing the non-stationary activities.
ALF-Score++, a novel approach to transfer knowledge and predict network-based walkability scores across cities
(Scientific Reports, 2022-08) Alfosool, Ali M. S.; Chen, Yuanzhu; Fuller, Daniel
Walkability is an important measure with strong ties to our health. However, there are existing gaps in the literature. Our previous work proposed new approaches to address existing limitations. This paper explores new ways of applying transferability using transfer-learning. Road networks, POIs, and road-related characteristics grow/change over time. Moreover, calculating walkability for all locations in all cities is very time-consuming. Transferability enables reuse of already-learned knowledge for continued learning, reduce training time, resource consumption, training labels and improve prediction accuracy. We propose ALF-Score++, that reuses trained models to generate transferable models capable of predicting walkability score for cities not seen in the process. We trained transfer-learned models for St. John’s NL and Montréal QC and used them to predict walkability scores for Kingston ON and Vancouver BC. MAE error of 13.87 units (ranging 0–100) was achieved for transfer-learning using MLP and 4.56 units for direct-training (random forest) on personalized clusters.
A theory-based model of cumulative activity
(Scientific Reports, 2022-09) Phillips, Kole; Stanley, Kevin; Fuller, Daniel
Energy expenditure can be used to examine the health of individuals and the impact of environmental factors on physical activity. One of the more common ways to quantify energy expenditure is to process accelerometer data into some unit of measurement for this expenditure, such as Actigraph activity counts, and bin those measures into physical activity levels. However, accepted thresholds can vary between demographics, and some units of energy measurements do not currently have agreed upon thresholds. We present an approach which computes unique thresholds for each individual, using piecewise exponential functions to model the characteristics of their overall physical activity patterns corresponding to well established sedentary, light, moderate and vigorous activity levels from the literature. Models are fit using existing piecewise fitting techniques and software. Most participants’ activity intensity profile is exceptionally well modeled as piecewise exponential decay. Using this model, we find emergent groupings of participant behavior and categorize individuals into non-vigorous, consistent, moderately active, or extremely active activity intensity profiles. In the supplemental materials, we demonstrate that the parameters of the model correlate with demographics of age, household size, and level of education, inform behavior change under COVID lockdown, and are reasonably robust to signal frequency.
Changes in physical activity after joining a bikeshare program: A cohort of new bikeshare users
(International Journal of Behavioral Nutrition and Physical Activity, 2022-10) Auchincloss, Amy; Michael, Yvonne L.; Niamatullah, Saima; Li, Siyu; Melly, Steven J; Pharis, Meagan L; Fuller, Daniel
Background There are hundreds of bikeshare programs worldwide, yet few health-related evaluations have been conducted. We enrolled a cohort of new bikeshare members in Philadelphia (Pennsylvania, USA) to assess whether within-person moderate and vigorous physical activity (MVPA) increased with higher use of the program and whether effects differed for vulnerable sub-groups. Methods During 2015–2018, 1031 new members completed baseline and one-year follow-up online surveys regarding their personal characteristics and past 7-day MVPA minutes per week (minutes per week with- and without walking). Participants were linked to their bikeshare trips to objectively assess program use. Negative binomial (for continuous outcomes) and multinomial (for categorical outcomes) regression adjusted for person characteristics (socio-demographics, health), weather, biking-infrastructure, and baseline biking. Results Participant median age was 30, 25% were of Black or Latino race/ethnicity, and 30% were socioeconomically disadvantaged. By follow-up, personal bike ownership increased and 75% used bikeshare, although most used it infrequently. Per 10 day change in past year (PY) bikeshare use, non-walking MVPA min/wk increased 3% (roughly 6 min/wk, P < 0.014). More days of bikeshare was also associated with change from inactive to more active (odds ratio for ≥ 15 days in PY vs. no bikeshare use 1.80, CI 1.05–3.09, P < 0.03). Results were consistent across vulnerable sub-groups. In general, impacts on MVPA were similar when exposure was personal bike or bikeshare. Conclusions Bikeshare facilitated increases in cycling, slightly increased non-walking MVPA, and showed potential for activating inactive adults; however, for larger program impact, members will need to use it more frequently.
Predicting walking-to-work using street-level imagery and deep learning in seven Canadian cities
(Scientific Reports, 2022-11) Doiron, Dany; Setton, Eleanor M; Brook, Jeffrey R; Kestens, Yan; McCormack, Gavin R; Winters, Meghan; Shooshtari, Mahdi; Azami, Sajjad; Fuller, Daniel
New ‘big data’ streams such as street-level imagery are offering unprecedented possibilities for developing health-relevant data on the urban environment. Urban environmental features derived from street-level imagery have been used to assess pedestrian-friendly neighbourhood design and to predict active commuting, but few such studies have been conducted in Canada. Using 1.15 million Google Street View (GSV) images in seven Canadian cities, we applied image segmentation and object detection computer vision methods to extract data on persons, bicycles, buildings, sidewalks, open sky (without trees or buildings), and vegetation at postal codes. The associations between urban features and walk-to-work rates obtained from the Canadian Census were assessed. We also assessed how GSV-derived urban features perform in predicting walk-to-work rates relative to more widely used walkability measures. Results showed that features derived from street-level images are better able to predict the percent of people walking to work as their primary mode of transportation compared to data derived from traditional walkability metrics. Given the increasing coverage of street-level imagery around the world, there is considerable potential for machine learning and computer vision to help researchers study patterns of active transportation and other health-related behaviours and exposures.

Browse

Recent Submissions