“Meta-analysis: essential tool for evidence-based clinical and public health practice”
with a guest talk from Dr Nicos Middleton (Cyprus University of Technology).Abstract
In clinical and public health research, it is not uncommon to identify studies with uncertain, or even conflicting, findings with regards to the effectiveness of certain therapeutic or preventive treatments. While individual trials might be underpowered and fail to show benefit (e.g. a statistically significant difference between the experimental and control/placebo groups), when results from individual studies are combined using appropriate techniques, significant benefits of treatment may be shown. An often cited historical example is the evidence on the effectiveness of streptokinase in improving survival after myocardial infarction. A retrospective review of the published evidence suggests that there was clear evidence of its benefits (had a meta-analysis been conducted) as early as 1973 after only 8 small studies involving 2432 patients. Yet it took several years, another 25 studies and 35000 randomised patients before its wide adoption in clinical practice. Therefore, there was a period at which patients were essentially deprived of an effective therapy, and thus challenging the critical premise of healthcare practice “first, do no harm”. Meta-analysis is a simple but powerful statistical technique which systematically collates and synthesizes findings across independent trial/studies considered combinable (i.e. feasible and justifiable), with the aim to provide a summary and more precise quantitative estimate, giving due weight to the size of the different studies. It is most often used in assessing the effectiveness of healthcare interventions, but not restricted to this case since it can also be used for diagnostic, prognostic or aetiological research questions. Meta-analyses are nowadays considered an integral tool of evidence-based (clinical and public health) practice while the validity of a meta-analysis depends on the methodological quality of the systematic review on which it is based. Other than providing increased statistical power to detect small but potentially clinically significant effect, meta-analyses provide the opportunity to explore the robustness of the findings using sensitivity analyses, to quantify and investigate possible sources of heterogeneity in the observed estimates, to identify patient- or study-specific factors that may relate to the observed heterogeneity and effect size (sub-group analysis and meta-regression) as well as to assess the likely presence of publication bias (i.e. the underrepresentation of negative findings in the literature).
Biography
Dr Nicos Middleton is an Associate Professor of Health Research Methodology and Biostatistics at the Cyprus University of Technology in the Department of Nursing. His studies include a B.Sc. in Statistics and Operational Research (University College London, 1997), a MSc in Statistics (London School of Economics, 1998) and a Ph.D. in Geographical Epidemiology (Bristol University, 2004). He has also worked as a Lecturer of Medical Statistics at the Bristol University. His research interests lie in the areas of Research Methodology (analysis of geographical data and cartography, Bayesian Statistics etc) and Public Health. He was one of the leading experts in various research projects on the effect of air pollution on cardiorespiratory health with collaborators at the Harvard School of Public Health and at the Cyprus International Institute. He has a series of publications in scientific journals and conference proceedings in the field of Epidemiology and Public Health and his papers have been cited more than 300 times.
When: Wednesday, February 11th, 2015 at 16:00-17:00
Where: Room CY114 (First floor), UCLan Cyprus, Pyla CY-7080
No comments:
Post a Comment