Tax compliance varies across sociodemographic categories. However, research on the relationships between compliance and age, sex, education as well as income level shows inconsistent results, both regarding the direction of the relationship and its size. The current meta-analyses target to merge findings in survey studies on age, sex, education, and income and estimate the strength of the impact on compliance by taking into account geographical regions. In four meta-analyses, comprising 459 samples (N = 614,286) from 111 countries published between 1958 and 2012, average estimated effect sizes were small, ranging from r = 0.12 for the relationship between compliance and age, r = 0.06 for sex, r = -0.02 for education to r = -0.04 for income. These effects are more pronounced in Western countries. It thus appears sociodemographic characteristics have little impact on compliance, but nevertheless should be controlled for in tax research.

The objective of this paper is to provide a robust understanding of the magnitude and direction of effect of urban green space on human health outcomes. This relationship has been studied using a variety of health and green space measures and multiple approaches to collect health data. Due to the various approaches used to study the relationship between green space and health, it is difficult to clearly understand the general relationship. In order to have a robust and clear understanding of the relationship between green space and human health, it is necessary to do a meta-analysis that considers all the approaches to assessing green space, health, and health data. Overall, the average estimated effect shows that surrounding urban green space improves human health by 1.14 fold (~u=0.13 (95% CI: (0.07-0.19)). The magnitude of effect increased when accounting for perceived health and when using greenness as the urban green space measure (~u=0.29 (95% CI: (-0.06-0.63)). In conclusion we can now objectively imply that health can be improved with increased surrounding green space. Further, that the magnitude depends on how studies assess green space and health measures, as well as how health data is collected.

Class of 2013 Abstract / Specific Aims: To evaluate three complementary and alternative medicines (CAM), tai-chi, acupuncture, and melatonin, for treating insomnia using meta-analysis assessment of randomized controlled trials. Methods: The electronic database MEDLINE (PubMed) was searched from May of 2012 to November of 2012 by using the terms “sleep initiation and maintenance disorders” AND “tai-chi” OR “melatonin” OR “acupuncture”. All of the searches ended at November of 2012. Data extraction was conducted independently by 2 investigators and any disagreements were resolved by consensus. If the 2 investigators could not agree, the study was reviewed by all 4 investigators. Main Results: Out of 500 studies that were initially retrieved, 12 studies were included; 3 for tai-chi; 4 for acupuncture; 5 for melatonin. We found that the effect of each type of intervention was significantly different than zero, p<0.01 thus all were effective in treating insomnia. From our analysis, acupuncture was the most effective (standard mean difference, SMD=-0.66; p<0.01) followed by tai-chi (SMD=-0.43; p<0.01) whereas melatonin was the least effective (SMD=-0.26; p=0.04) but difference between acupuncture and melatonin was not significant (p=0.15).       Conclusion: All three interventions were found to be effective in treating insomnia. However, due to mixed and inconsistent data of the studies, poorly designed trials, and small sample size, further large, well-controlled trials are warranted.

This thesis was motivated by data from the Emerging Risk Factors Collaboration (ERFC), a large individual participant data (IPD) meta-analysis of risk factors for coronary heart disease(CHD). Cardiovascular disease is the largest cause of death in almost all countries in the world, therefore it is important to be able to characterise the shape of risk factor–CHD relationships. Many of the risk factors for CHD considered by the ERFC are subject to substantial measurement error, and their relationship with CHD non-linear. We firstly consider issues associated with modelling the risk factor–disease relationship in a single study, before using meta-analysis to combine relationships across studies. It is well known that classical measurement error generally attenuates linear exposure–disease relationships, however its precise effect on non-linear relationships is less well understood. We investigate the effect of classical measurement error on the shape of exposure–disease relationships that are commonly encountered in epidemiological studies, and then consider methods for correcting for classical measurement error. We propose the application of a widely used correction method, regression calibration, to fractional polynomial models. We also consider the effects of non-classical error on the observed exposure–disease relationship, and the impact on our correction methods when we erroneously assume classical measurement error. Analyses performed using categorised continuous exposures are common in epidemiology. We show that MacMahon’s method for correcting for measurement error in analyses that use categorised continuous exposures, although simple, does not provide the correct shape for nonlinear exposure–disease relationships. We perform a simulation study to compare alternative methods for categorised continuous exposures. Meta-analysis is the statistical synthesis of results from a number of studies addressing similar research hypotheses. The use of IPD is the gold standard approach because it allows for consistent analysis of the exposure–disease relationship across studies. Methods have recently been proposed for combining non-linear relationships across studies. We discuss these methods, extend them to P-spline models, and consider alternative methods of combining relationships across studies. We apply the methods developed to the relationships of fasting blood glucose and lipoprotein(a) with CHD, using data from the ERFC.

Osteoporotic fractures are a major and increasing clinical and public health concern internationally. Identification of individuals at high risk for fragility fractures may enable us to target preventive interventions more effectively. In this thesis, I aimed to evaluate novel risk factors for osteoporosis and develop a fracture risk assessment model among the middle-aged and older people. I used data from the European Prospective Investigation into Cancer (EPIC)-Norfolk study, which is a large population-based prospective study started in 1993. About 25,000 men and women were assessed at baseline and about 15,000 of them returned for a second examination 4 years later. All participants are followed up to the present for clinical events including fractures. My work is in two parts. For the first part, I examined the risk of fracture associated with some novel or less well studied risk factors. These risk factors included change in height over time, respiratory function, physical activity and body fat mass. We found that men and women with annual height loss &gt;0.5 cm are at increased risk of hip and any fracture (relative risk=1.9 (95% CI 1.3-2.7) per cm/year height loss). One litre lower forced expiratory volume in 1 second (FEV1) was associated with a 2-fold risk of hip fracture in men and women. We also observed a non-linear association, independent of body mass index, between increasing body fat mass and lower fracture risk in women but not in men. I performed a systematic review and meta-analysis of studies evaluating the association between physical activity and hip fractures. Using a new validated questionnaire in EPIC-Norfolk, we observed varying relationships between physical activity in different domains of life and fracture risk in men and women. For the second part of the thesis, I developed a biostatistical model to calculate 10-year risk of developing a fracture among EPIC-Norfolk study participants. This model incorporates clinical and radiological assessments known to be associated with fractures and can be extended to other risk factors assessed in other prospective cohorts. This helps clinicians to achieve a better estimate of the prospective risk of fracture in their patients. I applied this model to compare the predictive value of two different clinical assessment methods for osteoporosis, namely dual-energy X-ray absorptiometry (DXA) and quantitative ultrasound (QUS). We found that that the predictive power of QUS is comparable to, and independent of, predictive power of DXA. In summary, my studies have added to our knowledge about some novel and easy-to-use risk factors of osteoporosis and proposed a practical method to merge and utilise data from different risk factors for estimation of fracture risk in individuals.

Antiplatelet therapy with clopidogrel is widely used in patients with coronary artery disease, but the recent development of the new P2Y12 receptor antagonists prasugrel and ticagrelor will increase treatment options. An overview of systematic reviews was performed to summarize available evidence on clopidogrel. Current data on prasugrel and ticagrelor were identified by a systematic review and used for an indirect treatment comparison (ITC) of the drugs against each other and versus placebo in the absence of head-to-head clinical trials. Adjusted indirect comparison according to Bucher, Bayesian methods for mixed treatment comparisons using Winbugs, and generalized linear mixed models using SAS were employed for ITC, yielding almost identical results: prasugrel was favored regarding stent thrombosis and ticagrelor regarding major bleeding. However, substantial differences in trial design were identified, demanding caution when interpreting these results. On the basis of the obtained results, a randomized controlled trial was designed within the gap of current evidence.

The study of ethnicity in bullying research has yielded inconsistent findings regarding the involvement and prevalence rates among ethnic majority and minority groups. On one hand, individual studies using ethnic group membership as a demographic variable indicate that ethnic minority groups are at times more or less likely to experience or perpetrate bullying compared to White students. On the other hand, contextual factors such as ethnic diversity have yielded more consistent findings showing that increased ethnic diversity is associated with lower bullying victimization among ethnic minority students. The role of ethnicity in bullying and peer victimization was examined in this dissertation by investigating both individual and contextual variables. Studies 1 and 2 consisted of two meta-analyses that systematically addressed comparisons between ethnic majority (i.e., White) and minority students (i.e., Black, Asian, Hispanic) on bullying perpetration and peer victimization. Results indicated small and non-significant overall effect sizes; however, methodological moderators suggested that ethnic groups differ on bullying and peer victimization across countries, measurements, and age groups. Study 3 examined school ethnic composition and bullying involvement in a population-based, ethnically diverse Canadian sample. Results indicated that ethnic minority students experienced less bullying victimization in schools with a higher proportion of ethnic minority peers. School ethnic composition was not associated with bullying victimization for White students or bullying perpetration across both ethnic groups. Taken together, these studies suggest that ethnicity as a demographic variable is not sufficient to account for differences in bullying involvement and that contextual variables are more adequate at explaining patterns of bullying across ethnic groups within the larger school and societal contexts.

BACKGROUND AND OBJECTIVES: Standard meta-analysis compares two treatments whereas network meta-analysis compares multiple treatments. In light of the increasing number of treatments available, we have seen a shift from using standard meta-analysis towards using network meta-analyses to support regulatory and reimbursement decisions. This thesis (composed of nine separate papers) applied network meta-analysis to a series of real world problems which simultaneously addressed many of these methodological issues while also supporting healthcare decision making. METHODS: In the first chapter, background information on network meta-analysis and outline the rationale for each of the subsequent chapters is provided. In Chapter 2, an applied example is presented where we use FDA Public Summary documents to populate a network meta-analysis of antithrombotic therapies for atrial fibrillation. In Chapter 3, the advantages and disadvantages of rapid network meta-analysis using ClinicalTrials.gov are investigated by re-doing Chapter 3 but only using data available in ClinicalTrials.gov and examining concordance. In Chapter 4, the application of network meta-analysis when events are rare are investigated using an illustrative example investigating the risk of serious infection in biologics. In Chapter 5, the application of network meta-analysis to a complex network – triptans for migraine – are presented where there are a large number of studies and treatments. In Chapter 6 a network meta-analysis which assesses both a benefit and harm and integrates findings using a benefit harm methodology is presented. In Chapter 7, methods to incorporate non-randomized studies into network meta-analysis are investigated using data on riskof myocardial infarction derived from the Mini-Sentinel distributed data network. FINDINGS AND CONCLUSION: Pragmatic research around methodological areas in network meta-analysis were conducted to address real world problems for decision makers. It is my hope that this thesis and the approaches used in this thesis for the application of network meta-analysis will be disseminated to enhance research capacity in conducting network meta-analysis in Canada.

Objective Postpartum hemorrhage (PPH) is a major cause of maternal mortality world-wide; misoprostol is a relatively cheap, easily administered, and an efficient medication to be given after the delivery of the baby to prevent PPH, thus posing it as a first choice in resource-poor communities. The aim of this study is to answer questions regarding the most appropriate dose (400 µg versus 600 µg), effect of labour settings (community or clinical), and management of labour on misoprostol effectiveness and safety in preventing PPH. Methods We developed a search strategy and conducted a search within five key databases. Two reviewers screened the articles for predefined inclusion/exclusion criteria, quality, and performed data extraction. Discrepancy was dealt with by reaching consensus. In article 1, we only included randomized controlled trials, we performed a random-effects Bayesian network meta-analysis comparing 400 µg to 600 µg misoprostol over five outcomes of interest: blood loss ≥500 ml, blood loss ≥1000 ml, using additional uterotonics, shivering, and pyrexia. In article 2, we included any experimental trial, we performed a random effects model meta-analysis, pooling the incidence of PPH from each misoprostol arm. Subsequently, a meta-regression model was performed on identified potential effect-modifiers, including clinical settings and labour management. Results Of 444 identified records, 46 trials met the inclusion/exclusion criteria in article 1, and 56 trials in article 2. The odds ratio (OR) of misoprostol 400 µg vs. 600 µg for bleeding ≥ 500 ml is 0.86 [95% Credible Intervals: 0.46 − 1.54], for bleeding ≥ 1000 ml the OR is 0.83 [95% CrI 0.54 – 1.26], for additional uterotonics is 0.75 [95% CrI 0.40 – 1.40], for pyrexia and shivering an OR of 0.57 [95% CrI 0.15 – 2.18] and 0.63 [95% CrI 0.29 – 1.31] respectively. The overall incidence of PPH was 6.62 per 100 pregnancies (95%CI 4.71 per 100 – 8.53 per 100). Labour settings and other aspects of active management of labour had no statistically significant effect on the incidence of post-partum hemorrhage. Conclusion We found no difference between the administration of misoprostol 400 µg or 600 µg for the prevention of PPH and side effects of misoprostol, as well as no effect of labour settings and management of labour on misoprostol effectiveness.

Intensive Behavioural Intervention (IBI) is one of the most widely used treatments for children with an autism spectrum disorder (ASD). While IBI has been recognized as the treatment of choice for very young children with an ASD, its sensible use among school age children is a matter of dispute. The aim of this thesis was to determine the clinical effectiveness of IBI, as compared with no treatment or treatment-as-usual, for the management of cognitive functioning and adaptive skills in preschool and school age children with an ASD, as well as to examine predictors of treatment response. Peer-reviewed, English language publications were identified using MEDLINE, EMBASE, PsychINFO, CINAHL, and ERIC from 1995 to September 1, 2014. Grey literature and reference lists of published papers were also searched for relevant records. Retrieved citations were screened by two independent reviewers, and data extraction was performed by a single reviewer with verification by a second reviewer. The methodological quality and procedural fidelity of included studies was assessed by one reviewer, and a subset of included studies were pooled in a random-effects meta-analysis using the standardized mean difference (SMD) effect size. A total of 24 unique studies were selected for inclusion in this review, comprising a total of 1,816 participants. Findings revealed that IBI improves full-scale IQ (SMD ES = 0.66, 95% CI 0.46 to 0.85, p<0.00001; 13 studies) and adaptive skills (SMD ES = 0.57, 95% CI 0.33 to 0.82, p<0.00001; 12 studies) in preschool and school age children with an ASD, with seemingly higher clinical benefits in children aged under 4 years at intake. Better outcomes with IBI are predicted by children’s relatively younger age, increased cognitive and adaptive ability, as well as a milder severity of symptoms at treatment entry. Results warrant careful interpretation in light of several methodological limitations and inadequate monitoring of procedural fidelity.