Manana Maridashvili. MD. PhD. Professor Evidence Based Medicine Topic 11 Advanced topics. Fixed and random effects models. Hypothesis testing. Examples of critical appraisal Advanced topics in EBM Evidence-based medicine (EBM) is a clinical approach that involves the integration of the best available research evidence with clinical expertise and patient values to make decisions about patient care. Here are a few advanced topics in EBM: 1) Meta-analysis: A statistical technique for combining the results of multiple studies to provide a more precise estimate of the overall effect of an intervention. 2) Systematic review: A systematic and transparent approach to reviewing the available research evidence on a specific topic. 1 Manana Maridashvili. Advanced topics. 3) Grading of recommendations, assessment, development, and evaluation (GRADE): A method for evaluating the quality of evidence and the strength of recommendations in clinical practice guidelines. 4) Clinical prediction rules: A tool for estimating the probability of a particular outcome based on the presence or absence of certain clinical features. 5) Decision analysis: A structured approach to comparing the potential benefits, harms, and costs of different options for management of a clinical problem. 6) Clinical practice guidelines: Evidence-based recommendations for the management of specific health conditions, based on a systematic review of the available evidence. 7) Risk stratification: The process of identifying individuals who are at high risk for a particular health outcome, so that they can receive targeted interventions to prevent or mitigate that outcome. 8) Evidence-based public health: The use of research evidence to inform the development, implementation, and evaluation of public health programs and policies. 2 Manana Maridashvili. Advanced topics. Fixed and random effects models. In the context of evidence-based medicine (EBM), fixed and random effects models can be used to synthesize the results of multiple studies on a particular topic or question. These models can help to identify the overall magnitude and direction of an effect, as well as the degree of heterogeneity or variability among the studies. Fixed and random effects models are used in meta-analysis, a statistical method for synthesizing the results of multiple studies on a particular topic or question. Meta-analysis involves calculating a weighted average of the effect sizes estimated in each study, with the weights typically being proportional to the inverse of the variance of the effect size estimate. A fixed effects model assumes that the differences between the studies are due to the fixed characteristics of those studies, rather than to random variation. This means that the estimates of the effect sizes for the studies are fixed and do not vary across different samples or studies. On the other hand, a random effects model allows for the estimates of the effect sizes for the studies to vary across different samples or studies. This is because the model assumes that the differences between the studies are due to both fixed characteristics and random variation. The choice between a fixed and random effects model depends on the research question and the characteristics of the data. In general, a fixed effects model is more appropriate when the studies being compared are similar and the goal is to estimate the average effect size across those studies, while a random effects model is more appropriate when the studies are more diverse and the goal is to account for that diversity in the estimates of the effect sizes. 3 Manana Maridashvili. Advanced topics. Hypothesis testing In the context of evidence-based medicine (EBM), hypothesis testing is a statistical method for evaluating the strength of the evidence supporting a particular hypothesis or claim. A hypothesis is a statement about a relationship between two or more variables. For example, a hypothesis might be that a particular medical treatment is effective in reducing the symptoms of a particular condition. In order to test a hypothesis, researchers design and conduct a study to collect data on the variables of interest. The study is typically designed to be representative of the population of interest and to minimize the impact of confounding variables (factors that could affect the relationship between the variables being studied). Once the data have been collected, the researchers use statistical tests to determine whether the observed relationship between the variables is statistically significant, or whether it could have occurred by chance. If the relationship is statistically significant, this means that it is unlikely to have occurred by chance, and provides evidence in support of the hypothesis. If the relationship is not statistically significant, this means that it is possible that the observed relationship occurred by chance, and provides less evidence in support of the hypothesis. It is important to note that statistical significance does not necessarily mean that the hypothesis is true or that the observed relationship is biologically or clinically meaningful. It simply means that the observed relationship is unlikely to have occurred by chance. 4 Manana Maridashvili. Advanced topics. Examples of critical appraisal. Critical appraisal is the process of evaluating the quality and relevance of research evidence in order to determine its validity, reliability, and clinical significance. Here are a few examples of questions that might be asked during a critical appraisal of a research study: ü Is the research question well-defined and relevant to the clinical question being addressed? ü Is the study design appropriate for answering the research question? ü Was the study population representative of the population of interest, and were the inclusion and exclusion criteria clearly defined? ü Was the sample size sufficient to provide an accurate estimate of the effect size, and was the sample size calculated a priori? ü Was the study properly randomized, and were the treatment groups similar at baseline? ü Were the outcome measures relevant and reliable, and were they measured in a consistent manner across all participants? ü Were the data analyzed appropriately, and were any statistical tests used appropriate for the study design and data? ü Were the results of the study clearly and accurately reported, and were any limitations of the study adequately discussed? ü Is the study generalizable to other populations or settings, or are the results specific to the study population and setting? ü Is the study an important contribution to the existing evidence on the topic, and does it have the potential to change clinical practice or policy? 5 Manana Maridashvili. Advanced topics.