Thesis advisor: David T. Takeuchi / Thesis advisor: Graciela Teruel-Belismelis / Poverty and violence within cities frequently concentrate in the same places and evidence suggests these exposures have deleterious consequences on health. The 2007 homicide increase in both rich and poor Mexican municipalities and the available biomarkers in a public panel study offer a unique opportunity to test each contextual effect in isolation on an innovative health outcome. Using an ecological framework, the main hypothesis of the dissertation is that, in urban environments, exposure to higher levels of contextual violence works as a stressor that wears down the body by increasing the levels of cardiovascular risk. This effect was hypothesized to be independent from poverty but with significant interactions and with heterogeneous effects among subpopulations. Multilevel cross-sectional and longitudinal analyses were conducted treating the data as a natural experiment using the homicide rate increase as treatment. The outcomes were two indices and single biomarkers that reflect cardiovascular risk in three waves of data corresponding to the years 2002, 2006, and 2012. Results showed that three complementary statistical approaches provided evidence indicating that exposure to cumulative violence at the municipality level yielded higher cardiovascular risk when controlling for individual covariates like victimization and household expenditure. The significant threshold for homicide rates was 35 and the differences between exposed and unexposed municipalities was between 1.5% & 8.3%, while the threshold for changes in the homicide rates between 2006 and 2012 was 10, with an effect size of 7%. Poverty and violence were not correlated in Mexico during the homicide rate spike, so the effects were independent. Unexpectedly, they did not show interaction effects: affluent and violent municipalities were the most stressful contexts. These effects were higher in women, in individuals in the two lowest socioeconomic quintiles and had significant impact in cohorts younger than 40 years old. The dissertation expands the ecosocial approach by exploring independent effects that shape multiple stressful contexts. It demonstrates that violence is a public health concern in Mexico that has indirect effects in the whole population, which not only worsens the obesity epidemic, but also demands a new perspective on assessing the burden of violence on everyday life. / Thesis (PhD) — Boston College, 2017. / Submitted to: Boston College. Graduate School of Social Work. / Discipline: Social Work.

The widespread use of highly active antiretroviral therapy (HAART) has resulted in longer lifespans for HIV seropositive women in the United States, during which preventive health care is recommended. Failing to complete recommended cancer screening tests can result in cancer being diagnosed at a later stage with a poorer prognosis. The purpose of the study, based on the ecosocial theory, was to describe the sociodemographic and clinical variables of HIV seropositive women who failed to complete recommended screening tests for breast, cervical, and colorectal cancers, and determine if the presence of hypertension, obesity, diabetes, depression, or tobacco use impacted the completion of these screening tests. The electronic medical records of 142 HIV seropositive women were reviewed. Univariate analysis, bivariate analysis, and logistic regression were conducted to create a model associated with the completion of preventive health care screening tests. For breast cancer, cervical cancer, and colorectal cancer, 69%, 71.8%, and 69.7% failed to complete screening, respectively. Number of years living with HIV infection and HIV stage were associated with breast cancer screening; distance between residence and health care facility, and HIV stage were associated with cervical cancer screening; and age and marital status were associated with colorectal cancer screening. Addressing issues related to the completion of cancer screening tests over the lifespans of HIV seropositive women can result in positive social change by preventing disease and disability, which can negatively impact these women, their families, and their communities.

The widespread use of highly active antiretroviral therapy (HAART) has resulted in longer lifespans for HIV seropositive women in the United States, during which preventive health care is recommended. Failing to complete recommended cancer screening tests can result in cancer being diagnosed at a later stage with a poorer prognosis. The purpose of the study, based on the ecosocial theory, was to describe the sociodemographic and clinical variables of HIV seropositive women who failed to complete recommended screening tests for breast, cervical, and colorectal cancers, and determine if the presence of hypertension, obesity, diabetes, depression, or tobacco use impacted the completion of these screening tests. The electronic medical records of 142 HIV seropositive women were reviewed. Univariate analysis, bivariate analysis, and logistic regression were conducted to create a model associated with the completion of preventive health care screening tests. For breast cancer, cervical cancer, and colorectal cancer, 69%, 71.8%, and 69.7% failed to complete screening, respectively. Number of years living with HIV infection and HIV stage were associated with breast cancer screening; distance between residence and health care facility, and HIV stage were associated with cervical cancer screening; and age and marital status were associated with colorectal cancer screening. Addressing issues related to the completion of cancer screening tests over the lifespans of HIV seropositive women can result in positive social change by preventing disease and disability, which can negatively impact these women, their families, and their communities.

Background: Depression is considered the worldwide leading cause of illness and disability in young people and an urgent public health issue. Within the field of public health it is of interest to deepen the understanding of determinants of depressive symptoms (DS) that are possible to address on a political or an organizational level. Also, it is of great importance to find methods to prevent depression in adolescents. To address these issues, the present thesis had two Aims: I) To study determinants of DS in adolescents, and II) to, by means of a non-randomized pragmatic trial evaluation, investigate the effectiveness a cognitive-behavioral intervention (DISA) in a real-world setting in relation to determinants of DS in adolescent girls. This was addressed by asking a) Which determinants on individual, psychosocial and structural levels are associated with DS in adolescents? (Paper I); b) What are the directional pathways between sexual harassment (SH) and DS? (Paper II); c) Which features characterize students who were assigned to a cognitive-behavioral intervention regarding levels of DS, psychosocial aspects and socioeconomic status of the respondents as well as of schools? (Paper III); d) Does DISA have an effect on DS in girls aged 14-16? (Paper IV); and e) Are there differences between the DISA participants and non-participants in the effects of psychosocial and structural determinants on DS? (Paper IV). Method: Data was collected in January 2010, 2011, and 2012, by means of a self-administered, electronic questionnaire in school. Students aged 14-16 in all nine public and one independent high school in a municipality in northern Sweden participated in the study (~1,000-1,200 students depending on the wave). All studies had DS as the single outcome variable. Individual level determinants were self-esteem and self-efficacy. Psychosocial determinants were parental/peer/teacher support; school demands; sexual harassment; and bullying. Structural determinants were family material affluence; parental foreign background, parental employment status; disrupted family, and personal relative affluence. Logistic regression was employed for research question a) (Paper I). Structural equation (SEM) cross-lag models were modeled for research question b) (Paper II). The Mann-Whitney U statistic was employed for research question c) (Paper III). SEM was used for research questions d-e) (Paper IV). Results: Determinants on individual, psychosocial and structural levels were independently associated with depressive symptoms in both genders. Self-efficacy, low teacher support, bullying victimization, and low personal relative affluence was associated with elevated levels of DS in both genders (Paper I). In girls, low parental support, high school demands, and sexual harassment victimization (SH) were also associated with elevated levels of DS. Among boys, parental migrant background was also associated with DS. Among girls, both the targeting of girls with elevated DS, and the consequence of SH explained the relationship between DS and SH victimization over time. In boys, only the predating of DS explained the association between DS and SH (Paper II). Only girls were assigned to DISA during 2011 and DISA participants reported higher levels of DS and lower levels of self-esteem than the non-participants at pre-intervention, which indicates that DISA was used as a targeted intervention for girls with elevated symptoms. Also, DISA participants reported higher levels of SH victimization, less peer support, and lower personal relative affluence (Paper III). In contrast to the non-participants, DISA participants did not increase their mean scoring on DS at an eight months follow-up. However, SEM analyses showed that the effect of DISA participation on DS at follow-up was negligible (Paper IV). Conclusions: This study showed that SH victimization was an important determinant for DS in girls followed by personal relative affluence. Among boys, personal relative affluence and parental migrant background were the most important factors. SH victimization had mental health consequences in girls only. DISA was implemented as a targeted intervention rather than as selective or universal one, and did not have an effect on DS in this group of girls. Implications for further research and health promotion practice in the school setting are discussed. / Bakgrund: Depression anses vara den världsledande orsaken till sjukdom och funktionsnedsättning hos unga människor och en av de viktigaste folkhälsofrågorna gällande ungdomar. Inom folkhälsoområdet är det viktigt att fördjupa förståelsen av bestämningsfaktorer för depressiva symtom (DS) som är möjliga att ta itu med på en politisk eller organisatorisk nivå. Det är också av stort intresse att finna metoder för att förebygga depression hos ungdomar. För att adressera detta hade denna avhandling två Syften: I) Att studera bestämningsfaktorer för DS hos ungdomar 14-16 år, och II) Att genom en icke-randomiserad pragmatisk utvärderingsstudie undersöka effekten av en kognitiv beteendemodifierande intervention (DISA) så som den implementerats i praktiken i skolan i förhållande till bestämningsfaktorer för DS hos flickor. Följande frågeställningar formulerades: a) Vilka bestämningsfaktorer på individ, psykosocial och strukturell nivå är förknippade med DS hos ungdomar? (Paper I); b) I vilken riktning går sambandet mellan sexuella trakasserier (ST) och DS? (Paper II); c) Vad karaktäriserar elever som gått en kognitiv beteendemodifierande metod (DISA) avseende nivåer av DS, psykosociala aspekter och respondenternas och skolornas socioekonomiska status? (Paper III); d) Har DISA en effekt på DS hos flickor i åldern 14-16? (Paper IV); och e) Fanns det skillnader mellan DISA-deltagare och icke-deltagare avseende effekterna av psykosociala och strukturella faktorer på DS? (Paper IV). Metod: Data samlades in i januari 2010, 2011 och 2012 med hjälp av en självadministrerad, elektronisk enkät i skolan. Elever i åldrarna 14-16 i samtliga nio kommunala skolor och i en friskola i en kommun i norra Sverige deltog i studien (~1000-1200 elever beroende på datainsamlingsår). Samtliga studier hade DS som enda hälsoutfall. Individfaktorer som antogs ha samband med DS var självkänsla och upplevd förmåga att hantera problem i livet. Psykosociala faktorer var föräldra-, kamrat-, och lärarstöd; krav i skolan; sexuella trakasserier och mobbning. Strukturella faktorer var materiellt välstånd i familjen; föräldrar med utländsk bakgrund, föräldrars arbetslöshet; splittrad biologisk familj och personligt relativt välstånd. Logistisk regression användes för forskningsfråga a) (Paper I). Strukturell ekvationsmodellering (SEM) med en cross-lag design användes för forskningsfråga b) (Paper II). Mann-Whitney U test användes för forskningsfråga c) (Paper III). SEM användes för forskningsfrågorna d-e) (Paper IV). Resultat: Bestämningsfaktorer på individ-, psykosocial och strukturell nivå var associerade med DS hos båda könen. Upplevd förmåga att hantera problem i livet, lågt lärarstöd, utsatthet för mobbning, och lågt personligt relativt välstånd var associerat med förhöjda nivåer av DS för både pojkar och flickor (Paper I). Bland flickor var dessutom lågt föräldrastöd, höga krav i skolan, och utsatthet för sexuella trakasserier (ST) också associerade med förhöjda nivåer av DS och hos pojkar var även föräldrars invandrarbakgrund associerat med DS. Bland flickor förklarades sambandet mellan DS och ST både av att flickor med förhöjda DS blev utsatta för ST, och av att ST ledde till senare DS. Hos pojkar var det endast att pojkar med DS blev utsatta för ST som förklarade sambandet (Paper II). Endast flickor deltog i DISA och de rapporterade högre nivåer av DS och lägre självkänsla än de som inte deltog, vilket tyder på att DISA användes som en riktad intervention för flickor med förhöjda symtom. DISA-deltagare rapporterade även högre nivåer av utsatthet för ST, lägre kamratstöd och lägre personligt relativt välstånd (Paper III). DISA-deltagarnas genomsnittliga nivå av DS hade inte försämrats vid åttamånadersuppföljningen vilket icke-deltagarnas nivåer hade. SEM-analysen visade dock att effekten av DISA-deltagande på DS vid uppföljning var försumbar (Paper IV). Slutsatser: Denna studie visade att utsatthet för SH var en viktig faktor för DS hos flickor, följt av lågt personligt relativt välstånd. Hos pojkar var lågt personligt relativt välstånd den viktigaste bestämningsfaktorn följt av föräldrarnas invandrarbakgrund. Utsatthet för ST hade konsekvenser för psykisk ohälsa för flickor men inte för pojkar. DISA genomfördes som en riktad intervention snarare än som en selektiv eller universell intervention och hade inte en effekt på DS i denna grupp av flickor. Implikationer för fortsatt forskning samt för hälsofrämjande arbete i skolan diskuteras. / <p>Vid tidpunkten för disputationen var följande delarbeten opublicerade: delarbete 4 manuskript.</p><p>At the time of the doctoral defence the following papers were unpublished: paper 4 manuscript.</p>

Type 2 diabetes has reached epidemic proportions in adults and youths. Persistent organic pollutants and endocrine-disrupting chemicals (EDCs), such as pesticides, dioxins, and organochlorines, are omnipresent and persist in the environment with potential for human exposure via contaminated air, waterways, soil, and human food supply. EDCs have been correlated with diabetes incidence and risks. Residential proximity to hazardous waste sites (HWS) has been correlated with increased hospital admission rates for diabetes. The study used a sample population (N = 1,724), ages ≥ 12 years from the 2005-2012 Continuous NHANES and HWS data from the National Priorities List of Superfund Sites. The ecosocial theory of disease distribution, and geocoordinates provided theoretical support. Mann-Whitney U test and binary logistic regression analysis were used to investigate the relationship between residence ≤ 1 mile compared with residence >1 mile from a HWS in NHANES surveyed counties of NJ, PA, NY, and CA on the outcome abnormal A1c ≥ 5.7% while controlling the effect of the moderators: abnormal body mass index (BMI), age, sex, and race/ethnicity on the relationship. Participants with a BMI ≥ 28.95 kg/m2 were 1.8 times and persons ≥ 58 years of age were 2.1 times more likely to have an A1c ≥ 5.7%. Also, non-Hispanic Whites residing >1 mile of a HWS had 82.1 % reduced risk of abnormal A1c compared with the same group residing ≤ 1 mile of a HWS. The results forge opportunities for future studies to consider border distance between residence and HWS. In addition, the results may promote positive social change through diabetes risks education, environmental health education, and practices and raise dialogue about social justice and the geographic distribution of hazardous waste sites.