With the growing number of older adults in the general population, there is also a concomitant rise in the number of older adults who require mental health services, making the measurement of psychological conditions in later life a priority. However, due to a lack of measures created for older adults, researchers and clinicians must often rely on measures created for younger populations. Three studies were designed to add to the field of evidence-based assessment and determine which anxiety measures possess strong evidence when used with older adults to warrant their use with this specific population. In the first study, I systematically reviewed the literature to identify the anxiety measures most commonly used with older adults. I reviewed each measure to examine its psychometric properties (e.g., internal consistency, test-retest reliability, inter-rater reliability, concurrent and discriminant validity) and the availability of age-appropriate norms in order to evaluate whether the instruments are appropriate for use with older adults. In the second study, I conducted a reliability generalization meta-analysis to estimate the mean reliability of each commonly used anxiety measure identified in the first study. Finally, in the third study, I examined whether the anxiety measures commonly used with an older population can be consistently and accurately categorized as evidence-based. The literature review and the reliability generalization study both revealed that most of the most commonly used measures lacked sufficient evidence to warrant their use with older adults. However, three measures (Beck Anxiety Inventory, Penn State Worry Questionnaire, and Geriatric Mental Status Examination) showed psychometric properties sufficient to justify the use of these instruments when assessing anxiety in older adults. In addition, two measures developed specifically for older adults (Worry Scale and Geriatric Anxiety Inventory) were also found to be appropriate for use with older adults. This suggests that based on their overall level of reliability and previous psychometric evidence, both researchers and clinicians assessing anxiety in a geriatric population should consider these measures as likely to be the best currently available.

Osteoporosis is a growing cause of morbidity and mortality in aging populations worldwide, especially in postmenopausal women. Bisphosphonates are widely prescribed for fracture prevention in osteoporosis. Meta-analyses have been performed for alendronate, risedronate, and etidronate, examining their effectiveness versus placebo in fracture prevention. Total withdrawals and adverse event withdrawals were examined as safety outcomes. Systematic reviews were performed for two other bisphosphonates, ibandronate and zoledronic acid and the results combined with previously obtained data for the other bisphosphonates. Indirect treatment comparisons of the drugs against each other and versus placebo were performed using Bayesian and frequentist methods. Both types of analyses yielded almost identical results: zoledronic acid and alendronate were the most effective bisphosphonates for preventing vertebral fractures. No differences were found regarding withdrawals. Subgroup analyses found that fracture prevention was generally more effective with longer therapy (greater than or longer than 3 years).

The development of competency in clinical supervision is becoming an increasingly important element of training and practice in professional psychology. To assist students in developing supervisory competencies, knowledge of relevant research and effective training methods is required. Three studies were designed to add to the field of supervisor development research. In the first study, I conducted a systematic review to examine the extent to which theoretical models of supervisor development have been used in empirical studies of supervisor development within professional psychology. This study revealed that studies rarely employ theory, and when they do, the Supervisor Complexity Model is the most often used theory, likely because it is the only one with an associated measure, the Psychotherapy Supervisor Development Scale (PSDS; Watkins, Schneider, Haynes, & Nieberding, 1995). Building on the findings of this systematic review, in the second study I conducted a reliability and validity generalization on the PSDS. Reliability and validity generalizations are meta-analytic methods that allow reliability and validity data from a measure to be summarized across studies. Results indicate the PSDS has a history of excellent reliability, and higher scores are found with increased supervisory training and experience, evidence of the scale’s validity. The third study is a meta-analysis of the supervisor development literature, in order to assess the effect sizes of training and experience on supervisory development and self-efficacy. Results indicate that training and experience have significant positive effect sizes, though they do not differ from one another. Based on results from these studies I provide recommendations about: the samples with which the PSDS tool is appropriately used, and how researchers might increase reliability within their own future studies, and evidence based recommendations for competency based education efforts.

Testosterone products comprise a multi-billion dollar industry; however, Health Canada and the United States Food and Drug Administration have issued statements of concern about their safety, and their benefits are unclear. In the first phase of this study, we performed a systematic review of the current evidence for the benefits of testosterone replacement therapy. In the second phase, we analyzed the evidence from randomized controlled trials for the benefits of testosterone replacement therapy (depression, quality of life, erectile function, libido, total testosterone level). In the third phase, we analyzed data for harms (cardiovascular death, myocardial infarction, stroke, newly diagnosed prostate cancer, coronary artery disease or diabetes, serious adverse events, withdrawals due to adverse events, erythrocytosis) from randomized controlled trials and non-randomized studies. In general, most included studies had a short duration and were poorly reported. Additional high-quality research into the long-term safety of testosterone products is needed.

As the HIV-positive population ages, managing non-AIDS-related comorbidities such as cardiovascular disease (CVD) complicates HIV care. Effectively treating risk factors for CVD will help reduce its burden in the HIV-infected population. However, the evidence base for the efficacy of statins as lipid-lowering therapies in HIV-infected patients has yet to be synthesized. Most trials do not compare statins directly to each other. In the absence of head-to-head evidence, the relative treatment effects of different statins can be indirectly obtained through a network meta-analysis (NMA). This NMA aims to evaluate the use of statins for treating dyslipidemia in HIV-infected individuals. Bayesian methods were used for obtaining treatment effect estimates and probabilistic rankings of treatments. Among lipid-lowering therapies, statins were most effective in treating dyslipidemia. All statins were found to offer the same treatment benefits. To our knowledge, this is the first NMA on this topic. It provides clinicians, health economists, and policy decision-makers with precise and reliable estimates for making definitive recommendations for the use of statins in dyslipidemic HIV-positive patients.

Class of 2008 Abstract / Objectives: To assess the effect of treatment with placebo on weight loss in obese patients. Methods: The placebo groups from nine orlistat and three sibutramine weight loss studies, as identified from the Padwal meta-analysis, were analyzed for: weight loss in kilograms; change in BMI; type of nutrition and exercise interventions; and bias. The means and standard deviations for weight loss were entered into the Stata meta-analysis program to obtain a forest plot to determine an overall estimate of weight loss. The Potential for Methodological Bias Assessment Tool (PMBAT) was used to create a bias score for each study. Results: Overall, subjects in the placebo group lost 3.8 kg (p<0.001). Placebo subjects in the orlistat trials lost significantly more weight (4.3 kg) than did subjects in the sibutramine trials (1.1 kg, p=0.014). The subjects in the sibutramine studies did not lose a significant amount of weight (p=0.397). There were distinct differences between the sibutramine and orlistat studies with regard to nutrition and exercise interventions. The mean score on the PMBAT was 16.7 ± 3.6; and no study scored higher than 25 out of 45 total points. Conclusions: Overall subjects in the placebo group in the combined orlistat and sibutramine studies lost a significant amount of weight. Differences in study design, bias, and the Hawthorne effect may have contributed to this weight loss.

Despite the large volume of research in the area of loyalty-repurchase-satisfaction, the findings on the relationship between these variables are conflicting. It seems that links between loyalty, repurchase, and satisfaction are not fully understood yet. The study provided the theoretical background on loyalty, repurchase, satisfaction, and their interrelationships. The Theory of Planned Behavior and the Expectation Confirmation Theory were discussed. The loyalty-repurchase differentiation with five contributing factors was addressed. From the literature review nine hypotheses were proposed. The study was conducted in three phases. The purpose of the first phase was to statistically identify the magnitude and the direction of relationships: loyalty-satisfaction, repurchase-satisfaction, and loyalty-repurchase by providing a quantitative review. A Hunter and Schmidt (1990) meta-analysis technique was employed. The results demonstrated that those three constructs have positive, strong relationships. However, the relationships were moderated by a number of factors, including the geographic region of the collected sample, the category (products versus service), and the business setting (B2B versus B2C). The purpose of the second phase was to statistically identify relationships between the research constructs, such as loyalty dimensions (commitment, trust, involvement, and word of mouth), repurchase/repurchase intent, and satisfaction, by conducting a field study with customers of apparel products. A structural equation modeling technique was employed to test the proposed hypotheses. Eight hypotheses were supported. The results indicated that although positive relationships between the research constructs exist, not all relationships are significantly strong. The purpose of the third phase was to compare the meta-analysis results, obtained from the large number of independent empirical studies, with the field study results, obtained from surveying consumers of apparel products. The overall findings of this research indicated that while meta-analysis and the field study results agree on positive relationships between the research constructs, the differences lay within the strength of the investigated relationships. The possible explanation was provided based on the theoretical foundation from the literature review. The study contributes to the growing knowledge of the relationships between loyalty, repurchase, and satisfaction by assessing the current state of the empirical research on those three variables. This research addresses the existing gap in the literature, and attempts to resolve the existing mixed views on the studied concepts. The mixed results of meta-analysis and the field study in terms of the strength of the investigated relationships indicate the need to expand this area of research further.

The purpose of the meta-analysis was to address the varied and somewhat stratified study results within the area of singing ability and instruction by statistically summarizing the data of related studies. An analysis yielded a small overall mean effect size for instruction across 34 studies, 433 unique effects, and 5,497 participants ranging in age from 5- to 11-years old (g = 0.43). The largest overall study effect size across categorical variables included the effects of same and different discrimination techniques on mean score gains. The largest overall effect size across categorical moderator variables included research design: Pretest-posttest 1 group design. Overall mean effects by primary moderator variable ranged from trivial to moderate. Feedback yielded the largest effect regarding teaching condition, 8-year-old children yielded the largest effect regarding age, girls yielded the largest effect regarding gender, the Boardman assessment measure yielded the largest effect regarding measurement instrument, and song accuracy yielded the largest effect regarding measured task. Conclusions address implications for teaching, research pedagogy, and research practice within the field of music education.

Philosophiae Doctor - PhD / Overweight/obesity is a risk factor for non-communicable diseases such as cardiovascular diseases, diabetes, and some cancers. Obesity in childhood is known to predict later obesity in adolescence and adulthood. Understanding the factors associated with overweight/obesity among children may present an opportunity for timely and appropriate interventions in the African setting.

Abstract Aim: It has been suggested that chronic hyperinsulinemia from insulin resistance is involved in the etiology of endometrial cancer (EC). We performed a systematic review and meta-analysis to assess whether insulin resistance is associated with the risk of EC. Methods: We searched PubMed-Medline, Embase, Scopus, and Web of Science for articles published from database inception through 30th September 2014. We included all observational studies evaluating components defining insulin resistance in women with and without EC. Quality of the included studies was assessed by NewcastleeOttawa scale. Randomeffects models and inverse variance method were used to meta-analyze the association between insulin resistance components and EC. Results: Twenty-five studies satisfied our inclusion criteria. Fasting insulin levels (13 studies, n Z 4088) were higher in women with EC (mean difference [MD] 33.94 pmol/L, 95% confi- dence interval [CI] 15.04e52.85, p Z 0.0004). No differences were seen in postmenopausal versus pre- and postmenopausal subgroup analysis. Similarly, non-fasting/fasting C-peptide levels (five studies, n Z 1938) were also higher in women with EC (MD 0.14 nmol/L, 95% CI 0.08e0.21, p < 0.00001). Homeostatic model assessment - insulin resistance (HOMA-IR) values (six studies, n Z 1859) in EC patients were significantly higher than in women without EC (MD 1.13, 95% CI 0.20e2.06, p Z 0.02). There was moderate-to-high heterogeneity among the included studies. Conclusion: Currently available epidemiologic evidence is suggestive of significantly higher risk of EC in women with high fasting insulin, non-fasting/fasting C-peptide and HOMAIR values. / Revisión por pares