Linear mixed-effects (LME) modeling is a widely used statistical method for analyzing repeated measures or longitudinal data. Such longitudinal studies typically aim to investigate and describe the trajectory of a desired outcome. Longitudinal data have the advantage over cross-sectional data by providing more accuracy for the model. LME models allow researchers to account for random variation among individuals and between individuals. In this project, adolescent health was chosen as a topic of research due to the many changes that occur during this crucial time period as a precursor to overall well-being in adult life. Understanding the factors that influence how adolescents' mental well-being is affected may aid in interventions to reduce the risk of a negative impact. Self-esteem, in particular, has been associated with many components of physical and mental health and is a crucial focus in adolescent health. Research in self-esteem is extensive yet, sometimes inconclusive or contradictory since past research has been cross-sectional in nature. Several factors associated with self-esteem development are considered. Participation in religious services has also been an interest in research for its impact on depression. Depression development and its predictors are evaluated using LME models. Along with this line, this project will address the research problems identified through the following specific topics (i) to investigate the impact of early adolescent anxiety disorders on self-esteem development from adolescence to young adulthood; (ii) to study the role of maternal self-esteem and family socioeconomic status on adolescent self-esteem development through young adulthood; and (iii) to explore the efficacy of religious service attendance in reducing depressive symptoms. These topics present a good introduction to the LME approach and are of significant public health importance. The present study explores varying scenarios of the statistical methods and techniques employed in the analysis of longitudinal data. This thesis provides an overview of LME models and the model selection process with applications. Although this project is motivated by adolescent health study, the basic concepts of the methods introduced have generally broader applications in other fields provided that the relevant technical specifications are met.

Background: Across North America, smoking prevalence is highest among young adults (YAs). Understanding the cessation behaviours of YAs is critical given their higher smoking rates; however, there is a paucity of prospective studies on YA cessation from a population-based sample of smokers. Objectives: This study characterizes younger and older adult smokers as well as identifies the rate of making a quit attempt (QA) and smoking cessation among a representative sample of younger and older adult smokers. Guided by the Social Cognitive Theory (SCT), this study also examines the interpersonal predictors of moving toward smoking cessation among YAs. Methods: Self-report data on 592 YA and 2777 older adult smokers were compiled from the Ontario Tobacco Survey baseline and six-month follow-up interviews. Smoking cessation behaviour was measured as an ordinal variable: made no attempt to quit, made a QA that lasted for less than 30 days, and successfully quit for 30 days or longer. Design-based analyses examined the characteristics of young and older adult smokers. Making a QA and smoking cessation during a six-month follow-up period were modeled according to the SCT constructs while accounting for the complex study design. Results: Young adult smokers were more likely to be men with lower levels of addiction but greater self-efficacy and an intention to quit than their older counterparts. While YA smokers were more likely to make a QA than their older counterparts (25% vs. 17%, respectively), they were no more likely to succeed (14% vs. 10%, respectively). Having an intention to quit smoking and having made two or more lifetime number of QAs predicted making a QA; the use of smoking cessation aids or resources and having knowledge that stop smoking medications make quitting a lot easier also contributed to making an attempt to quit. Self-efficacy, use of smoking cessation aids or resources and having someone to support one‘s QA were positive predictors of quitting whereas having high levels of addiction was a negative predictor of cessation. Conclusions: Young and older adults are distinct types of smokers with different personal and smoking characteristics. Different factors predicted making a QA and smoking cessation among YAs. Smoking cessation interventions for YAs should provide social support and skills to build and maintain self-efficacy to quit. It is critical to ensure YA smokers have effective smoking cessation aids and services that are easily accessible and appropriate for this population. Future research is needed to understand long-term smoking cessation and relapse in this vulnerable population.

国立情報学研究所で電子化したコンテンツを使用している。

国立情報学研究所で電子化したコンテンツを使用している。

Studies over the past couple of decades have depicted loneliness as a significant concern to physical health, although its meaning for overall health outcomes is still unclear. The precise impact of loneliness on life expectancy and on specific disease processes remains unknown. With regression modeling techniques, this thesis uses data from the Alameda County Health and Ways of Living Study to characterize the impact of loneliness on self-rated health, mortality, and fatalities from specific diseases. A key hypothesis is that loneliness as a health problem hinges on its persistence over time. This hypothesis is also tested by examining the reliability of the loneliness measure across the full 34 years of the survey. A second test is to examine its interplay with marital status as a mutable social circumstance. Results show that loneliness is a risk factor for poor self-rated health, non-ischemic cardiovascular diseases, cerebrovascular diseases, infections, and overall mortality. Results also show that loneliness need not be a stable problem across the life span in order to pose health risks. The reliability of the loneliness measure fades across time and levels of loneliness also vary with changes in marital status. Loneliness did not clearly mediate the impact of marital status on self-rated health, mortality, or specific causes of death.

Plasma lipid, lipoprotein and apolipoprotein levels are considered as important and well-established intermediate quantitative phenotypes of Cardiovascular Disease (CVD) risk. Both the mean values and the phenotypic variance vary over the human lifespan. However, it is not known whether there is a genetic basis for this age variability. For example, might different genes act, or different gene interactions occur, as a person ages? If so, how might this be influenced by both environment and phenotype? An understanding of traits at different ages will not only provide insight into the genetic components involved in CHD development, but may also identify additional genetic factors that predispose an individual or population to premature (and later-onset) CHD. By identifying genetic factors that account for variation in important intermediate traits (i.e. lipid levels), we hope to gain a better understanding of disease mechanisms and thus a better chance of developing clinical strategies for preventing or possibly treating abnormal lipid levels and, by association, CHD. The aim of this thesis was to better identify and explain the genetic basis of CHD by focusing on the use of lipid traits as intermediate quantitative phenotypes of CHD. First, phenotypic analyses using structural equation modeling were performed to estimate the relative importance of genetic and environmental factors, and also to investigate whether these traits are influenced by the same gene(s) across time or whether they are age-specific genetic effects. Then, genome-wide linkage analysis was performed to localize cardiovascular susceptibility loci. Finally, a small genome-wide association scan (GWAS) was performed on a subset of the data to identify the relevant variants, in particular those showing associations across time. Phenotypes and marker data were collected in two Australian samples: an adolescent and adult twin pair samples. The adult sample consisted of 1453 twin pairs (968 monozygotic and 485 dizygotic), measured for lipid traits. 415 adult twins provided blood on two to five occasions. The adolescent dataset consisted of 965 twin families (397 monozygotic and 568 dizygotic) measured longitudinally at ages twelve, fourteen and sixteen, and their siblings tested once for the same lipid variables. Results from both the adult and adolescent cohorts indicated that there is more than one genetic factor influencing total cholesterol, HDL, LDL and triglycerides over time (i.e. from different measurement occasions). Common environmental factors did not contribute to variances (except for HDL in adolescents). There were no sex differences in the heritabilities of these intermediate phenotypes. Non-shared environmental factors did not have significant long-term effects. Overall, these two cohorts confirmed that genetic variation contributes substantially to variation in these traits, and suggested there are changes in the genes affecting plasma lipid concentration at different periods of life. Thus, there are age-dependent gene effects influencing HDL, LDL, total cholesterol, or triglycerides at different ages. In the adult genome data, there were 485 adult dizygotic twin pairs typed on average 595 markers, at an average inter-marker distance of 5.0 cM. The genome-wide linkage analysis revealed evidence for linkage in the 7p13 region for triglycerides. Possible candidate genes included NPC1L1 and GSBS. Other regions of “suggestive” linkage identified were chromosome 4p13 (at 62 cM) and Xq26.2-28 (81 cM). Adolescent twins and their siblings from 760 families were typed for linkage using 16,781 markers spaced across the genome at an average distance of 6.25 cM. The adolescent data revealed evidence for linkage to region 6p24.3 for triglycerides (–log10p = 6.81; equivalent LOD = 6.13; p = 0.00000016) and to region 2q31.1 for HDL (–log10p = 3.22, equivalent LOD = 2.27; p = 0.00061). No obvious candidate gene is known in this 6p region. Possible candidate genes in the 2q region include LRP2 and ABCB11. A significant region of linkage was also found on 2q35 for LDL (–log10p = 5.59; equivalent LOD score = 4.53). Other interesting regions of linkage included chromosomes 1q32.1, 4p15.1, 5q13.2, 11p14.3 and 18q11.2. Thus, regions were identified by linkage analyses that are likely to harbour genetic risk factors for cardiovascular disease in the analysed Australian population: chromosomes 7p13 (in adults), 6p24 (adolescents), 2q31.1 and 2q35 (in adolescents). Other regions included 1q32.1, 4p15.1, 5q13.2, 11p14.3 and 18q11.2 in adolescents and chromosome 4p13 and Xq26.2–28 in adults. Genome-wide association results for adolescents showed significant evidence of association between total cholesterol at age 14 (p = 8.24x10-7) and rs10503840 on 8p21.1. Such association has not previously been reported. Evidence of differential association across time was also found between HDL and variant rs10492859, located in the intron of the CDH13 gene, consistent with earlier studies on larger datasets. Significant association (p = 2.25x10-6) was also found between rs10507266 on 12q24.21 in an intron of THRAP2, a gene involved in early development of heart and brain, with triglycerides at age 12. Evidence of association was also found between HDL across time and variant rs10492859 on 16q23. Several other “suggestive” potential loci associated with lipid traits at one time point as well as across time were also found. In conclusion, the work described in this thesis establishes the importance of age-specific genetic effects on plasma lipids and lipoproteins, and identifies several regions of highly significant genetic linkage with these phenotypes in either adolescence or adulthood. It is clear that, as well as cross-sectional studies to identify genes affecting CVD risk factors, longitudinal genetic linkage and association studies are needed to assess relative contributions to risk across the lifespan.

Background There is a body of literature suggesting that postnatal depression – and poor maternal mental health generally – may lead to impaired child mental health, child behaviour and poorer cognitive outcomes. Some studies have also provided data showing significant associations between maternal mental illness and impaired child outcomes. While the available data is suggestive, it is largely derived from cross-sectional studies which do not have a capacity to assess the temporal sequences involved. Little is known about the impact of the severity, timing or chronicity of maternal mental health problems over the child’s life course on child mental health and behaviour. After examining patterns of maternal mental health impairment over the period of child rearing, the impact of maternal mental health impairment on adolescent behaviour problems at fourteen years of age is assessed. Specifically, the role of maternal anxiety, depression and co-occurring anxiety and depression were considered. This study used data from a prospective birth cohort study, with repeated measures of maternal mental health, to determine whether impaired maternal mental health in pregnancy, early childhood or early adolescence impact on the adolescent child’s mental health and behaviour. Material and Methods The Mater University Study of Pregnancy (MUSP) provided the data base for this study. Women were recruited from a large metropolitan hospital in Brisbane, Australia at their first antenatal visit – generally around the middle of the second trimester of pregnancy (on average at eighteen weeks gestation). Additional data was obtained 3-5 days after the birth of the child, six months after the birth of the child, when the child reached five years of age, and again when the study child reached fourteen years of age. During these phases of data collection, detailed information was obtained on the mother’s demographic characteristics, health information and also relating to a range of social factors. Data relating to the child was also derived using maternal reports throughout the study, and also child self-report at the fourteen-year follow up. Maternal anxiety and depression was measured at each phase of the study using the Delusions-Symptoms-States Inventory (DSSI) developed by Bedford & Foulds (1978). Child self-reported mental health and behaviour was assessed using the Child Behaviour Check List.(CBCL) (Achenbach & Edelbrock 1990). Maternal age and family income were included in analyses as potentially confounding factors, and maternal marital change and marital satisfaction were included as potential mediators. The full birth cohort of the MUSP study constituted 7223 mothers and their babies. The subset used in the current study includes 4297 mother and child pairs. This constitutes approximately sixty percent of the original birth cohort. Findings Maternal mental health impairment was analysed in terms of severity of impairment, timing of mental health impairment (during sensitive periods of child development – the antenatal/postnatal period and during later childhood/adolescence), and recurrence of episodes of mental health impairment. Rates of maternal mental health impairment were found to increase in occurrence and severity over the course of child rearing, with recurrence of episodes being a common manifestation of all forms of impairment. It is significant to note the mental health of women in the sample appears to decline as their children grow and develop into adolescents. The increase in maternal mental health impairment experienced as children progress through childhood into adolescence might be indicative of the pressures associated with raising children. This is a notion not previously explored in the literature. Contrary to existing literature, the analyses conducted in this study found that maternal mental health impairment occurring during the antenatal/postnatal period of child development had no independent negative impact on adolescent behavioural outcomes. The longitudinal nature of the data allowed for an assessment of these sensitive periods, whilst controlling for later occurring mental health impairment – a potential confounder not accounted for in many previous studies. Proximal and recurrent exposure to maternal mental health impairment was associated with a significantly increased risk of adolescent mental health and behavioural problems in children at fourteen years of age. Conclusion There is a need to further explore the nature and course of maternal mental health impairment during child rearing. The findings presented here point to the need for support for mothers to improve their mental health during later stages of child rearing – particularly in the period during which their children are adolescents. This period is significant both in terms of increased rates of maternal mental health impairment, and also the negative influences of exposure to maternal mental health impairment in terms of impaired adolescent mental health and behavioural outcomes.

Thesis (Ph. D.)--University of Washington, 2007. / Vita. Includes bibliographical references (p. 177-191).

Thesis (Ph. D.)--University of Washington, 2008. / Vita. Includes bibliographical references (p. 116-123).

Thesis (M.S.)--University of Michigan, 1961.