Meta-analysis is an important statistical tool for systematic reviews and evidence-based medicine. Extracting the observed effect sizes, assessing the magnitude of heterogeneity, choosing the suitable statistical model, and interpreting the summary effect size are four key steps in meta-analysis. It is known that each of the above steps has its own unique characteristics and may require some specific attention. As an example, the observed effect sizes from individual studies may not be reported in the same scale and hence cannot be combined directly. Another example is on selecting a model for meta-analysis from the common-effect model and the random-effects model. When a meta-analysis contains only few studies, the common-effect model and the random-effects model will often lead to misleading or unreliable results. In the first part of the thesis, we give a brief introduction on evidence-based medicine, systematic reviews and meta-analysis. We will also show their practical importance, display their relationships, and present a motivating example for conducting a meta-analysis. In Chapter 2, we first review the common effect sizes in meta-analysis for both continuous data and binary data. How to combine different categories of effect sizes is a critical issue after extracting the observed effect sizes from the clinical studies in the literature. For continuous data, researchers have recently proposed methods that transform the five number summary to the sample mean and standard deviation (Hozo et al., 2005; Wan et al., 2014; Luo et al., 2018). For binary data, the transformation from the odds ratio (OR) to the relative risk (RR) in the cohort study was proposed by Zhang and Yu (1998). To the best of our knowledge, however, there is little work in the literature that converts OR to RR in the case-control study. In view of this, we establish a new formula for this transformation to fulfill the gap. The performance of the new method will be examined through simulations and real data analysis. Our method and formulas can not only handle meta-analyses with different effect sizes, but also offer some insights for medical researchers to further understand the meaning of OR and RR in both cohort and case-control studies. In Chapter 3, we first give a brief introduction on the three available models in meta-analysis: the common-effect model, the random-effects model, and the fixed-effects model. When a meta-analysis contains only few studies, the common-effect model and the random-effects model will often lead to misleading or unreliable results. In contrast, the fixed-effects model is capable to provide a good compromise between the existing two models. In this chapter, we propose to further improve the estimation accuracy of the average effect in the fixed-effects model by assigning different weight for each study as well as fully utilizing the information in the within-study variances. Through theory and simulation, we demonstrate that the fixed-effects model can serve as the most convincing model for meta-analysis with few studies. And most importantly, with a total of three models, we expect that meta-analysis can be conducted more flexibly, more meaningfully, and more accurately. In Chapter 4, we first give a brief introduction on the heterogeneity in meta-analysis. We then review the methods for quantifying heterogeneity in three directions as follows: the tests for heterogeneity, the estimates of the between-study variance, and the measures of the impact of heterogeneity. Note that most existing methods were derived under the assumption of known within-study variances. In practice, however, a direct use of the reported within-study variance estimates may largely reduce the power of the tests and also lower the accuracy of the estimates, especially when the sample sizes in some studies are not sufficiently large. To overcome this problem, we propose a family of shrinkage estimators for the within-study variances that are able to borrow information across the studies, and derive the optimal shrinkage parameters under the Stein loss function. We then apply the new estimates of the within-study variances to some well-known methods for measuring heterogeneity. Simulation studies and real data examples show that our shrinkage estimators can dramatically reduce the estimation bias and hence improve the exiting literature. Keywords: Common-effect model, Effect size, Fixed-effects model, Heterogeneity, Meta-analysis, Odds ratio, Random-effects model, Relative risk, Risk ratio

Cluster headache is a primary headache disorder that can be highly disabling. In this thesis we look at the treatment landscape of cluster headache with a scoping review of preventive and acute therapies for cluster headache as identified in randomized controlled trials and two-arm observational studies. We subsequently compare these therapies where data are available using network meta-analysis of randomized trials, and we attempt subgroup analyses again where data are available for acute treatments of episodic and chronic cluster. We identify the ranking of treatments for acute cluster headache, and certain acute therapies that may be beneficial in episodic and chronic cluster headache. Based on our findings, we also identify future directions for cluster headache trials.

No description available.

Indiana University-Purdue University Indianapolis (IUPUI) / Purpose: The purpose was to comprehensively assess the impact of PD-L1 expression on the efficacy of immune checkpoint inhibitors on Overall Survival (OS) and Progression-Free Survival (PFS). Methods: A systematic literature search and review was conducted through June 2019. I searched all eligible randomized controlled trials comparing PD-1/PD-L1 monotherapy to an active comparator in adult patients with advanced cancer across multiple tumor types. The Cochrane risk-of-bias tool was used to assess trial quality. A random-effects model was used for the meta-analysis. Heterogeneity was assessed using Cochran Q statistic and I2 test. Publication bias was assessed by visual inspection of a funnel plot and Begg’s test. Results: I identified and included 23 trials involving 14,434 participants. When stratifying PD-L1 positive (+) and negative (-) patients using varying thresholds of expression, a significant group difference was observed at PD-L1 >1% ( p=0.04; PD-L1(+): HR, 0.72; 95% CI, 0.65-0.79; PD-L1(-): HR,0.83; 95% CI, 0.75-0.91), at PD-L1 >10% (p=0.02; PD-L1(+): HR,0.50; 95% CI, 0.38-0.62; PD-L1 (-): HR, 0.74; 95% CI, 0.57-0.90) and at PD-L1>50% (p=0.01; PD-L1(+): HR,0.59; 95% CI, 0.51-0.68; PD-L1(-): HR, 0.93; 95% CI, 0.71-1.15). Across tumor types, both PD-L1(+) and PD-L1(-) patients treated with an immunotherapy had improved OS compared with patients receiving standard care therapies. A PFS benefit was observed and favored patients treated with a PD-1/PD-L1 inhibitor versus standard of care. However, there was significant heterogeneity and the benefit on PFS was not statistically significant between PD-L1(+) and PD-L1(-) groups using varying cut-off levels of PD-L1 expression. No differences between sub-groups of interest including median follow-up time, type of inhibitor, and line of therapy for either PD-L1(+) or PD-L1(-) patients at 1% cut-off were identified. Conclusion: This study supports the use of PD-L1 as a predictive biomarker of improved response to immunotherapies. As thresholds increase and specifically above the 10% PD-L1 expression threshold, patients who were positive for PD-L1 appeared to have better OS compared to those who were negative for PD-L1. Further investigation is needed to assess the clinical usefulness of PD-L1 at various expression levels with improved technologies that have the potential to enhance assay accuracy and precision.

This thesis consists of two related studies presented as three separate manuscripts (all three have been published in peer-reviewed journals) and a study protocol that has been submitted for peer-reviewed funding. The over-arching theme of this thesis was to better characterize the efficacy of different intravenous fluids used for the resuscitation of intensive care unit (ICU) patients with severe sepsis or septic shock. We performed an extensive search including multiple databases which found 20 randomized controlled trials (RCTs) that examined the effects of different intravenous fluids used in septic patients and met our a priori inclusion and exclusion criteria. In the first manuscript, we described in detail the composition of the 19 unique fluid products that were used in the various studies. This description included the fluid type, trade name, osmolality, tonicity, electrolyte content, molecular composition, pH, and manufacturer. We reviewed manufacturer’s websites, product monographs, and emailed industry representatives or study authors for more information regarding the fluids as required. The results of this study and systematic review led us to the second and third manuscripts which reported on a Bayesian network meta-analysis (NMA) of all fluid type comparisons. Despite multiple well-done RCTs, comparative data regarding the clinical effect of different resuscitative fluids when used for sepsis was incomplete. Most RCTs used 0.9% saline (normal saline) as control fluid and very few studies compared colloids directly. The advantage of using an NMA model in this setting was the ability to include indirect data into the overall point estimates. Data was abstracted from the 14 studies which focused on adult ICU patients and analyzed examining the outcomes of mortality (manuscript #2) and the use of renal replacement therapy (RRT) (manuscript #3). Certainty of evidence was evaluated for both outcomes using the GRADE approach. Results of the analysis clearly document the harm of starch-based fluids when used in septic patients. Albumin containing fluids and crystalloids (such as normal saline and Ringer’s Lactate) are better options. Lower chloride solutions, such as Ringer’s Lactate, showed a signal towards decreased mortality and a decreased use of renal replacement therapy when compared to higher chloride fluids, such as normal saline, however this was based on indirect data, not statistically significant, and warrants direct comparison trials. The final component of this thesis is a pilot study protocol for a study assessing the feasibility of a larger RCT examining the effect of low chloride versus high chloride fluids for resuscitation in patients with sepsis and septic shock. This protocol has been submitted as part of a peer-reviewed grant with the hopes of addressing this clinically important and timely question. / Thesis / Master of Science (MSc) / This thesis examines the ideal intravenous fluid to be given to patients with severe infection causing low blood pressure. A review of the current literature is presented with a protocol for future work.

No description available.

No description available.

Yes / We conduct a meta-analysis of the literature of financial development and economic growth. We cover a large number of empirical studies and estimations that have been published in journal articles. We measure the degree of heterogeneity and identify the causes of the observed differentiation. Among the most significant factors behind this heterogeneity is the choice of financial-variable proxies, the kind of data used as well as whether a study takes into account the issue of endogeneity. Our results suggest that the empirical literature on the finance–growth nexus is not free from publication bias. Also, a genuine positive effect exists between financial development and economic growth.

Publication bias (PB) makes the resources for meta-analysis (M-A) unreliable in the sense of completion and accuracy, so to investigate, identify and correct PB is a very important issue in M-A. The current study proposed an empirical comparison in both detection and correcting PB, using a Monte Carlo study. Conditions to be manipulated include the number of primary studies, number of missing studies and true effect size. RANNOR in SAS will be used to generate normally distributed random variables and, for each condition, 10,000 M-As will be simulated. Type I error rates are to be calculated for the conditions with no PB and powers were estimated for the conditions with PB and adequate type I error control. Finally, a demonstration of how M-A can and should be used as a part of program evaluations was given. / text

No description available.