The aim of research is to faithfully observe and describe, predict, determine causation, explain the hitherto unexplained, and conjure further questions. Medical research is concerned with the application of the scientific method to investigating both our population at risk, and our patients. However, it is often impossible to study entire populations. Therefore, many studies are undertaken to analyse representative cohorts before extrapolating the data to the population of interest. Every study is susceptible to bias and error. It is important to determine whether the evidence from an individual or groups of studies is strong or weak. In other words, are these data sets truly representative of the entire population of interest, or are the data distorted due to potential confounding factors? Furthermore, have the investigators presented a measured interpretation of their results in the context of the potential errors within their experimental system? It is these questions that have led to the concept of research evidence, or the hierarchical system in which the strength of the argument within some studies is simply better, or more persuasive, than others. When planning a research study, design is the first consideration. The highest possible level of evidence is a systematic review or meta-analysis of randomized controlled trials (RCTs), or an individual RCT. These are considered the ‘gold standard’ of clinical research. The design of RCTs allows exclusion of confounding factors and bias as much as possible. These studies work very well for certain interventions, such as drug trials. However, where the control and sample groups cannot be blinded, (e.g. ‘sham acupuncture’ or ‘sham manipulation’ as the control), RCTs may be less appropriate. Meta-analyses are considered as type 1 evidence. The more data is pooled, the more valid the results. However, the data may be less relevant to individual patients. Therefore, although potentially the most powerful type of evidence, meta-analyses can have some important limitations. A meta-analysis takes a number of trials from the literature, e.g. 10 trials of 100 patients each.
Oxford Scholarship Online requires a subscription or purchase to access the full text of books within the service. Public users can however freely search the site and view the abstracts and keywords for each book and chapter.
If you think you should have access to this title, please contact your librarian.