Evidence types and study designs used in systematic reviews of healthcare:
An expert opinion is a personal view of a domain given by a professional in writing or verbally. It is not a reflection of experience and therefore considered a weak form of evidence.
Mechanistic studies (experimental)
These include animal trials, in vitro studies and cell cultures that are useful in prediction of effects. However, observed effects can also differ, so subsequent human trials are required before a particular effect can be said to be seen in humans. Tests on isolated cells can also produce different results to those in the body.
Case Reports & Case Series (observational)
A case report is a written record on a particular subject. Though low on the hierarchy of evidence, they can aid detection of new diseases, or side effects of treatments. A case series is similar but tracks multiple subjects. Both types of study cannot prove causation, only correlation.
Cross sectional studies, surveys (observational)
Cross sectional studies measure the prevalence of an outcome or determinants of health in a broad population at one point in time or over a short period. They are used to explore etiology, but associations are made cautiously.
Case-Control Studies (observational)
Case-control studies are retrospective, involving two groups of subjects, one with a particular condition or symptom, and one without. They then track back to determine an attribute or exposure that could have caused this. Again, these studies show correlation, but it is hard to prove causation.
Cohort Studies (observational)
A cohort study is similar to a case-control study. It involves selection of a group of people sharing a certain characteristic or treatment (e.g., exposure to a chemical), and compares them over time to a group of people who do not have this characteristic or treatment, noting any difference in outcome.
Quasi-experimental studies (experimental)
Subjects are non-randomly assigned to groups to receive either a treatment or placebo.
Randomised Controlled Trials (experimental)
Subjects are randomly assigned to a test group, which receives the treatment, or a control group, which commonly receives a placebo. In ‘blind’ trials, participants do not know which group they are in; in ‘double blind’ trials, the experimenters do not know either. Blinding trials helps remove bias.
Systematic Review (evidence summaries)
Systematic reviews draw on multiple randomised controlled trials to draw their conclusions, and also take into consideration the quality of the studies included. Reviews can help mitigate bias in individual studies and give us a more complete picture, making them the best form of evidence.
Single case studies
Single case studies are poor for practice but can generate hypotheses for later studies (e.g., interviews with small numbers of people).
Descriptive studies focus on a sample selected from a specific group or setting. They simply describe participant view or experience. They are used to provide additional information for trials (e.g., illustrative quotes).
Conceptual studies include a range of conceptual categories identified as significant in earlier research and study participants are selected to provide an understanding of each category. The focus is on developing an overall account of the views and then drawing appropriate conclusions.
Generalisable studies build on earlier studies using conceptual frameworks for sample selection and they report on issues of diversity and sample saturation.
Challenge studies (experimental)
Challenge studies are a type of trial where study subjects are deliberately exposed to the disease cause. They are conducted in veterinary medicine (e.g., proof of concept or evaluating initial safety), and they may be used to evaluate both therapeutic and preventive interventions.
Ecological studies (observational)
Ecological studies explore the association between the occurrence of disease and exposure to a known or suspected cause among a population or community.
Before and after studies
Before and after studies measure a variable of interest before and after an intervention in the same participants. They lack a comparator and therefore they are considered to be weaker in design than RCTs. They may be used to draw conclusions about the time-related relationship of the variable to the intervention.
1. Application of systematic review methodology to food and feed safety assessments to support decision making. EFSA Journal, 2010. 8(6).
2. Arksey, H. and L. O'Malley, Scoping studies: towards a methodological framework. International Journal of Social Research Methodology, 2005. 8(1): p. 19-32.
3. Lichtenstein, A.H., E.A. Yetley, and J. Lau, Application of systematic review methodology to the field of nutrition. J Nutr, 2008. 138(12): p. 2297-306.
4. Thomas J, et al., Determining the scope of the review and the questions it will address, in Cochrane Handbook for Systematic Reviews of Interventions version 6.3 T.J. Higgins JPT, Chandler J, Cumpston M, Li T, Page MJ, Welch VA Editor. 2022, Cochrane.
5. Lefebvre C, et al., Searching for and selecting studies, in Cochrane Handbook for Systematic Reviews of Interventions version 6.3 T.J. Higgins JPT, Chandler J, Cumpston M, Li T, Page MJ, Welch VA Editor. 2022, Cochrane.
6. Konno, K., et al., Ignoring non-English-language studies may bias ecological meta-analyses.
Ecology and Evolution, 2020. 10(13): p. 6373-6384.
7. Amano T, et al., Tapping into non-English-language science for the conservation of global
biodiversity. PLOS Biology, 2021. 19(10): e3001296.
8. Evans, D., Hierarchy of evidence: a framework for ranking evidence evaluating healthcare interventions. Journal of clinical nursing, 2003. 12(1): p. 77-84.
9. Daly, J., et al., A hierarchy of evidence for assessing qualitative health research. Journal of clinical epidemiology, 2007. 60(1): p. 43-49.