Rachel Murphy

Canadians have been reported to have poor dietary intake, with a high consumption of ultra-processed food (UPF), which are energy-dense foods that are typically industrially formulated and contain high levels of added sugar, sodium, and saturated fat while offering minimal nutritional value. UPF is a major contributor to chronic diseases and inversely correlated with overall diet quality. While individual-level factors influencing diet are well-documented, the role of neighborhood characteristics in shaping dietary intake remains underexplored. The retail food environment (RFE), mainly referring to where people can purchase food (e.g., supermarkets), is one such neighborhood-level factor that may affect dietary intake. This research aimed to characterize dietary intake of a contemporary cohort of British Columbians and examine the associations between features of RFE and UPF onsumption, accounting for neighbourhood level characteristics. Using data from BC Generations Project and its sub-study, with a linkage to the Canadian Food Environment Dataset (Can-FED) and Urban Environmental Health Research Consortium (CANUE), this research assessed dietary intake via a validated past year food frequency questionnaire and measured the RFE through absolute and relative densities of food outlets within 1-km radius of participants' residences. UPF intake was classified using the NOVA system and expressed as energy per 1,000 kcal/day. Findings showed that UPFs accounted for about 38% of the daily energy intake among 2,586 cohort participants, with nutrient profiles characterized by high saturated fat, sodium, and added sugars, and insufficient fiber, calcium, and vitamin D. Living in areas with higher densities of fast-food outlets was significantly associated with greater UPF intake, after adjusting for age, sex, body mass index (BMI), education, annual household income, and neighbourhood level characteristics including walkability, marginalization status, and urban sprawl. Additionally, the ecological analysis of the RFE across census subdivisions in BC revealed substantial regional variation in access to different food outlets. This research underscores that dietary intake is not solely an individual choice but is shaped by neighbourhood-level factors. The findings highlight the need for public health interventions and urban planning strategies that limit the overconcentration of fast-food outlets, aiming to support healthier dietary patterns and reduce diet-related chronic disease risk in British Columbia.

View record

Background: Ultra-processed food (UPF) consumption has been linked to adverse health outcomes, including increased cancer risks. The underlying mechanisms may involve diet-induced metabolic alterations and shifts in gut microbial communities. This study aimed to use metabolomic and gut microbiome profiling to elucidate metabolic pathways and microbial signatures associated with UPF intake and determine the relevance of pathways to carcinogenic processes.Methods: Data were obtained from 800 participants of the BC Generations Project (BCGP). Dietary intake was assessed using a past year validated food frequency questionnaire (FFQ). UPF consumption was quantified as a percentage of total energy intake. Non-fasting serum samples were assayed using mass spectrometry-based metabolomic profiling, and stool samples using whole metagenome shotgun sequencing. Metabolite data were analyzed using multivariate regression, principal component analysis, and orthogonal partial least squares discriminant analysis. Microbial diversity and composition were evaluated using alpha and beta diversity metrics, differential abundance testing, and core microbiome analysis.Results: A total of 1,080 metabolites were detected, of which 68 metabolites were significantly associated with UPF consumption after adjusting for covariates (q0.05). Among these, seven metabolites were positively associated with UPF, while 61 were negatively associated with UPF. Notably, pipecolate levels were inversely associated with UPF intake, while specific phosphatidylethanolamines and 6-deoxodolichosterone were positively associated with UPF intake. Multivariate analyses, including OPLS-DA, revealed distinct metabolic profiles for low versus high UPF consumers. Enrichment analyses indicated that many metabolite changes were concentrated in pathways frequently implicated in cancer biology such as arginine biosynthesis, TCA cycle, and nucleotide metabolism. Microbiome analyses indicated differences in core microbial taxa between these groups, with seven species unique to Quartile 1 and Quartile 4, respectively. MaAsLin2 differential abundance analyses identified 12 species that varied by UPF consumption. Of particular interest, both approaches consistently demonstrated higher abundances of butyrate-producing bacteria—Faecalibacterium prausnitzii, Roseburia hominis, and Butyribacter sp000436755—in Quartile 1 compared with Quartile 4.Conclusion: Our study suggests that high UPF consumption is associated with distinct serum metabolic profiles and differences in gut microbial composition. These findings provide mechanistic insights into the potential pathways by which UPF consumption may contribute to cancer risk.

View record

Background: Multiple factors shape dietary intake. One potential factor is the availability of various food outlets within a locality, which constitutes the food environment. Previous studies in Canada have primarily been descriptive, focused on characterizing food outlets in a given regional food environment. There has been limited exploration of the potential association between food outlets with individual dietary intake. This study aims to characterize neighbourhood food environments of participants living across British Columbia (BC), and to investigate how the food environment is associated with consumption of fruits and vegetables (FV).Methods: This cross-sectional study used data from the BC Generations Project linked to geospatial measures of the food environment (Can-FED) and neighbourhood environment (CANUE). Five food environment measures were categorized by density, with class 0 representing no presence and class 4 the highest density. Descriptive analysis examined the distribution of residents across classification levels, while cross-tabulation explored relationships between measures. Multivariable mixed-effect models assessed associations, using linear models for daily FV consumption and logistic models for the odds of meeting FV recommendations (≥5 servings/day).Results: Around 50% of BC residents lived in neighbourhoods without direct access to grocery stores, fast food outlets, or convenience stores within walking distance. No significant differences in FV consumption were observed across density levels. Logistic models showed inverse associations between neighbourhoods with the highest density of fast food restaurants and FV consumption (OR=0.88, 95% CI: 0.80–0.94). Class 2 density of grocery stores (OR=0.86, 95% CI: 0.75–0.97) and convenience stores (OR=0.92, 95% CI: 0.84–0.99) were also associated with lower FV consumption. These associations remained significant after adjusting for neighbourhood characteristics, except for convenience stores.Conclusion: Participants living in neighbourhoods with higher densities of fast food restaurants were less likely to consume ≥5 servings of FV per day. Other retail food environments showed no consistent associations with FV consumption.

View record

Background/Objectives: Demographic, health history, and lifestyle factors have been associated with prognosis of colorectal cancer (CRC), but the mechanisms underlying these associations remain poorly understood. Our primary objective was to explore the association of these factors with expression of two biomarkers in CRC, SPARC and PD-L1, for which lower and higher levels of expression, respectively, have been associated with poorer CRC prognosis. Methods: Participants were drawn from the British Columbia Generations Project (BCGP), who at time of recruitment, provided data on various demographic, health history, and lifestyle factors. Formalin-fixed paraffin-embedded blocks were obtained for 49 incident BCGP CRC cases and used to create tissue microarrays. Slides created from the microarrays were stained with SPARC and PD-L1 antibodies and analyzed to calculate H-scores as measures of expression in both epithelial and non-epithelial tissues. Linear regression analyses were used to evaluate associations between the various factors and ln-transformed H-scores.Results: In CRC tumour epithelium, smoking was associated with a 0.53-fold lower level of SPARC expression (p=0.05); a similar though non-significant, association was observed in non-epithelial tumour tissue. Higher income was associated with a 1.33-fold greater level of SPARC expression in tumor non-epithelial tissue (p=0.04); a similar effect was not observed in tumour epithelium. Higher stage was associated with a 0.74-fold lower level of non-epithelial tumour SPARC expression (p=0.04). A similar non-significant effect was observed in tumour epithelium. PD-L1 expression in tumour epithelium was 2.84-fold greater among females (p=0.005). A similar non-significant effect was observed in tumour non-epithelial tissue. History of CRC screening was statistically significantly associated with 2-fold greater levels of PD-L1 expression in tumour epithelial and non-epithelial tissues (p=0.04). Conclusion: Associations of SPARC expression with smoking, income, and stage, but not PD-L1 expression with biological sex and history of CRC screening, were consistent with the previously established associations of these factors with CRC prognosis. Larger scale studies with prognostic data are needed, but our results suggest that differences in expression of SPARC may contribute to the previously observed impacts of factors on CRC prognosis. However, additional work to understand the role of PD-L1 in CRC prognosis is needed.

View record

Background: Despite evidence that substance use is associated with adverse health outcomes, the use of legal substances such as alcohol, cigarettes, cannabis, and e-cigarettes is common in Canada. Reducing and preventing the use of these substances has implications for the primary prevention of many chronic health conditions. The current study aimed to identify individual- and neighbourhood-level risk factors for the use of alcohol, cigarettes, cannabis, and e-cigarettes in the Canadian Partnership for Tomorrow’s Health cohort. Methods: Using a cross-sectional research design, individual-level factors: stress, depression, and sleep, and neighbourhood-level factors: deprivation, gentrification, household security, labour force participation, immigration and visible minority proportion, and population density, were studied as risk factors for substance use. Regression models were built to understand the association between substance use and potential risk factors while adjusting for covariates.Results: High levels of neighbourhood material deprivation, social deprivation, and household insecurity were associated with higher odds of using cigarettes, cannabis, and e-cigarettes. Odds of using these substances was also higher for participants who lived in gentrified neighbourhoods. Lower odds of cigarette, cannabis, and e-cigarette use were found for participants who lived in neighbourhoods with a high proportion of residents who are recent immigrants and/or visible minorities. Among individual factors, participants that experienced frequent stress and depression had higher odds of using cigarettes, cannabis, and e-cigarettes. Associations between alcohol consumption and individual- and neighbourhood-level risk factors were generally modest.Conclusions: The associations between substance use and neighbourhood environment are largely novel findings in Canada. Further, these findings suggest experiencing negative emotions is associated with substance use among Canadian adults and that strategies aimed at reducing stress and depression could have implications for decreasing the prevalence of substance use behaviours. Evidence from the current study suggests interventions aimed at reducing or preventing substance use should be multidimensional, encompassing strategies directed at both individuals and the larger environments to which people belong (e.g., neighbourhoods). Findings from the current study may help inform intervention planning, resource distribution, and the development of health and social policies with the aim of reducing and preventing substance use.

View record

Background: Mental health disorders are highly prevalent in the Canadian population and has been associated with cancer risk; however, previous findings in literature are inconsistent (1–4). This study aims to elucidate the relationship between mental health disorders and cancer risk, as well as explore the potential mediating effects of lifestyle behaviours.Methods: A cohort study was conducted with 34,571 participants aged 40-69 years from the province of Quebec. Depression was conceptualized from the PHQ-9, antidepressant use, and either a positive screen from PHQ-9 scores, antidepressant use, or self-report of physician diagnosis. Anxiety was defined using the GAD-7, and co-morbid depression and anxiety was assessed using the PHQ-9 and the GAD-7. Cox proportional hazards regression models were used to investigate the association between mental health exposures and risk of prostate, lung, and all cancers combined. Mediating effects of health behaviours were assessed using Baron and Kenny mediation criteria., then a Quasi-Bayesian/Monte Carlo approximation was used to obtain confidence intervals. Results: For risk of all cancers combined, there was a modest positive association with all mental health exposures, however none reached significance with full adjustment. No relationships reached significance for prostate cancer. There were positive associations between mental health disorders and lung cancer risk, but only anxiety and lung cancer in women was significant with full adjustment (HR = 1.67, 95% CI: 1.01-2.76). Women had consistently higher risk estimates than men for all cancer and lung cancer risk for the majority of exposures. Smoking status mediated the relationship between depression (PHQ-9) and lung cancer, anxiety and lung cancer, and co-morbidity and lung cancer by 27%, 18%, and 26%, respectively in women. In men, smoking status mediated 17% of the relationship between depression (PHQ-9, antidepressant use, or self-report of physician diagnosis) on all cancers. Conclusions: Positive associations were observed between mental health disorders and overall and lung cancer risk, however few relationships reached significance. Risk estimates were generally higher in women than in men, suggesting a differential risk. Smoking status mediated a significant proportion of the relationships between mental health disorders and cancer risk.

View record

Background: The Food Explorers (FE) program was designed and developed for use with kindergarten and grade 1 children. Offered within the classroom setting, the program aims to encourage children to identify and experience a wide variety of foods. The objective of this study was to examine the impact of the program on children’s willingness to try new and familiar foods. Methods: A quasi- experimental pre-post design was employed to evaluate the impact of the program on children’s familiarity and willingness to try novel and familiar test foods. Children in the FE group participated in food -related classroom activities led by teachers, over 5 months and the control group were taught the standard curriculum. Assessments of familiarity and willingness to try foods was completed at baseline and after 5 months in both groups. Parent-perceived food neophobia was also assessed using a validated scale at both times. Open-ended questions were used to assess parent and teacher experience with the program. Results: Analyses were conducted with 194 children (FE group, n=102 and control group, n=92). Knowledge of familiar foods increased over time in both groups and a significant group by time interaction was observed with 2 test foods (out of 10) at follow-up, indicating increased knowledge in the FE group, compared to the control group. Willingness to try the test foods, measured as food preference scores significantly increased for five novel foods in the FE group compared to the control group. No significant difference was observed in parent-perceived food neophobia scores between groups. Parents indicated positive experiences with respect to their children’s willingness to try new foods at home. Positive experiences were also noted by the teachers.Conclusion: While there is some evidence to suggest that the FE program may positively impact food knowledge and willingness to try new foods in kindergarten and grade 1 children, limitations of the study and implementation of the FE program warrant consideration when considering the impact of the program. The findings of the study provide valuable information on barriers for implementation of the FE program in classrooms and will inform future evaluations of the program.

View record

Background: Cancer survivors may be motivated to change their diet for weight management and improved survival. However, dietary behaviours and obesity are influenced by a range of individual-level and environmental factors rather than simply individual choices. This study aims to understand the diet quality of cancer survivors and its association with obesity and neighbourhood environment in comparison to non-cancer controls.Methods: A cross-sectional study was conducted with 19,973 participants aged 35 to 69 years from Atlantic Canada. A healthy eating index (HEI) was used to evaluate diet quality using dietary intake collected from a food frequency questionnaire. Obesity was assessed using anthropometric and bio-impedance measures. Neighbourhood environment data were derived from the Canadian dissemination-area-level census data. Multivariable multi-level models were used to investigate HEI and its association with obesity in cancer survivors compared to non-cancer controls. The associations between neighbourhood deprivation, population density and HEI were also explored. Results: Cancer survivors (n = 1930) had a higher mean HEI than non-cancer controls (mean difference: 0.45, 95% CI [0.07, 0.84]). Body mass index, waist circumference, waist-to-hip ratio, body fat percentage, and fat-free mass index were not associated with HEI, while trunk fat percentage had a weak positive association with HEI (0.45, 95% CI [0.08, 0.83]). The diet-obesity relationship did not significantly differ by cancer status. Mean HEI was lower in the most compared to the least socially deprived neighbourhoods (-0.56, 95% CI [-0.88, -0.25]), and in the most compared to the least dense neighbourhoods (-0.39, 95% CI [-0.77, -0.01]), but was not associated with material deprivation. Associations between diet quality, material deprivation and population density significantly differed by urbanicity. In rural areas, diet quality decreased with increasing material deprivation and decreasing population density, while the reverse was true for urban areas. Conclusion: Cancer survivors had a slightly better diet quality than non-cancer controls, but both groups are in need of dietary improvement. Diet was not associated with obesity measures in the cancer and non-cancer groups but was associated with neighbourhood deprivation and population density with evidence of urban-rural differences, suggesting the complexity of dietary behaviour and the need for multi-level interventions.

View record