<p>Over the last few decades, household incomes have been rising steadily but the biggest share of that growth has been in the top income quintile. This unequitable growth is exacerbated by spatial factors, such as a household’s home location and the distribution of opportunities and resources around it. This thesis examines the effects of the distribution of places of employment and healthy food retailers on socioeconomic outcomes for households in the Phoenix Metropolitan Statistical Area (MSA). The main research objectives are to: i) calculate accessibility to employment and healthy food for public transit and automobile and identify low-access areas for the two modes; ii) determine the underlying reasons and recommend mitigation measures for the access gaps; and iii) investigate the associations between the calculated accessibility measures and socioeconomic outcomes for households. This is achieved by calculating healthy food accessibility, general job accessibility, industry job accessibility (i.e., access to jobs by the different North American Industry Classification System (NAICS) different industry sectors) and industry job mismatch (i.e., the difference in the general job accessibility and the industry job accessibility). We employ a modified Enhanced 2-step Floating Catchment Area (E2SFCA) method with continuously decaying weights using census data at the census block level, employment data from the Longitudinal Employer-Household Dynamics (LEHD) program in various North American Industry Classification System (NAICS) sectors, and locations of grocery stores from the Safegraph places dataset. Using spatial regression models, we estimate the associations between these calculated measures and socioeconomic outcomes. The analysis shows that there is a lack of transit access to healthy food and employment in the periphery of the MSA and that greater access is clustered around interstates and primary roads. The access by transit is worse than access by automobiles for most of the block groups. The spatial regression results show that renters find block groups with better access more attractive, and homeowners trade off that accessibility for better living conditions and other factors, such as school districts. People using the Supplemental Nutrition Assistance Program (SNAP) and people not high school educated have worse industry job accessibility; jobs that they are employed and qualified in are not accessible to these vulnerable populations. City planners and policy makers can use these findings to implement policies that can address the access gaps found and impact communities. </p>

L’objectif de la thèse était de développer un indicateur mesurant l’accessibilité spatiale des premiers soins (nommé Index of Spatial Accessibility : ISA) pour les femmes enceintes, à partir de sources de données collectant des informations en routine. Ce travail a pour vocation de mettre en évidence les limites des indicateurs existants tout en apportant des améliorations. Une attention particulière a consisté à étudier l’impact de l’effet des contours administratifs (ou effet de bord), la limite pointée dans de nombreuses études portant sur l’accès aux soins. L’indicateur d’accessibilité aux professionnels de santé que nous avons développé a pour objectif de mettre en évidence les disparités spatiales à une échelle géographique fine afin d’identifier les zones géographiques dans lesquelles il faudrait intervenir en priorité. Ce travail s’est concentré tout d’abord sur les professionnels de santé intervenant dans le suivi de la grossesse : médecins généralistes, sages-femmes et gynécologues. Les résultats mettent en évidence qu’en combinant la disponibilité avec la proximité des soins, les besoins en matière de santé et la mobilité, l’ISA permet de fournir une meilleure mesure d’accessibilité. L’ISA a été construit afin de pouvoir interroger l’accès aux soins pour d’autres pathologies ou d’autres populations. L’analyse de l’impact de l’effet de bord met en évidence que : 1) la moyenne et l'écart-type sont légèrement inférieurs avec effet de bord que sans, quel que soit le type des professionnels de santé ; 2) La variation d’ISA est plus marquée pour les sages-femmes et les gynécologues, et pour les zones rurales. Nous avons également menée une étude pilote sur le recours aux soins des femmes enceinte, à partir des données SNIIRAM afin d’étudier la relation entre le recours aux soins et l’indicateur ISA. / This paper developed an improved indicator: the Index of Spatial Accessibility (ISA) to measure geographical healthcare accessibility at the census blocks level, and seeks to assess the effect of edge on the accuracy of defining healthcare provider access by comparing healthcare provider accessibility accounting or not for the edge effect, in a real-world application. The indicator of accessibility to health professionals developed aims to highlight spatial disparities measured at a fine geographical scale and to identify area where actions are needed in priority. This work focused first of all on the health professionals involved in the follow-up of the pregnancy: general practitioners, midwives and gynecologists. The main finding is that by combining availability with proximity to services, health needs and mobility, and by calculating at the smallest feasible geographical scale, ISA provides a better measure of accessibility. ISA was conceived so that we could question the access to care for other pathologies and other populations. When we compare the variation of ISA with and without edge effect, we found that (1) mean and standard deviation are slightly below when offer and demand outside are taken in to account, whichever health professionals considered; 2) the variation of ISA is higher for midwives and gynecologists, and for rural areas. In addition, we also conducted a pilot study on the health use of pregnant women, using SNIIRAM data to examine the relationship between use of care and the ISA indicator.