The hierarchy of scientific evidence is a ranking system which provides a detailed outline of the quality and reliability of different types of research studies. You can use the hierarchy of scientific evidence to help determine the reliability of a study. Higher quality research will have a higher rank than lower quality research. The hierarchy looks at three things: novelty, importance, and rigor to determine the quality of research.

Our current knowledgebase is the foundation on which we stand. It can come from text books or from a qualified individual in their field. Expert opinion is subjective and one of the lesser reliable forms of evidence. It’s based on a person’s experience and observations. They also rely on anecdotes and personal testimonies, which are not necessarily representative or applicable to the whole population. Less reliable doesn’t mean it should be dismissed. It’s a primary stepping stone towards scientific evidence.

There are two primary classifications of studied. Observational studies are studies that do not involve any intervention or experiments. These include case reports, case series, cross-sectional studies, and cohort studies. Experimental studies introduce a variable such a treatment or supplement and study the effects. Experimental studies are generally considered more reliable than observational because they control for extraneous variables. Further on the hierarchy are a synthesis of multiple studies. Systematic reviews and meta-analyses are detailed reviews of the information collected from various sources.

Case Report/Case Series

Case reports follow an individual patient and detail the diagnosis, treatment, and follow-up. There are inherent limitations to these types of studies due to the lack of statistical sampling. Case series examine a collection of similar cases and draw conclusions from the outcomes. Case reports provide initial insight and are a stepping stone into recognition of novel disease presentations and potential treatment options.

Case Control

Case Control studies determine the risk factors for disease by comparing previous exposure and looking to see if it results in a particular outcome. Cases that were already confirmed and controls are reviewed for exposure of a risk factor to determine a link. They are retrospective studies that are used to estimate the odds. Retrospective studies have issues with data quality due to recall bias.  They can be used to establish an association and are good for studying rare conditions.

The hierarchy is separated into different levels with each level more specific in design. For example, at Level 1 you might find randomized controlled trials, while at Level 5 you might find systematic reviews with meta-analysis and other statistical analysis techniques.

Cohort Studies

Cohort studies follows participants with similar characteristics over a period of time. Events are measured chronologically and therefore can be used to elucidate cause and effect. You can use cohort studies to associate the risk with an outcome. The link between smoking and lung cancer was identified with a cohort study.  

Randomized Control Trials

Probably the most well-known, randomized control trials compare outcomes of experimental and control groups. Clinical research uses a placebo or current standard of care as a comparison with a proposed treatment. They are easier to blind and have clearly identified populations. Blinding is essential to limit the effects from psychological bias. Once subjects are assessed for eligibility, they can be randomized to the allocated intervention. The final analysis comes after all data is collected and unblinded for the statisticians.

Systematic Review

Systematic reviews use a system to include, combine, and evaluate studies. You have to determine the question and the inclusion/exclusion criteria and must include any studies that fit the pre-determined criteria. The system aims to reduces bias. Anyone repeating the process will end up with the same literature. Researchers compile primary studies and summarize them into one paper.

Meta-Analysis

Meta-analysis is a subset of systematic review. It’s a statistical process to synthesize the the data of primary studies and pool them together. By pooling primary studies together, you increase the sample size. Systematic reviews and meta-analyses are often confused with one another because of their similarities. Synthesis studies are used to increase the precision or resolve discrepancies.