The support of medical decisions comes from several sources. These include individual physician experience, pathophysiological constructs, pivotal clinical trials, qualitative reviews of the literature, and, increasingly, meta-analyses. Historically, the first of these four sources of knowledge largely informed medical and dental decision makers. Meta-analysis came on the scene around the 1970s and has received much attention. What is meta-analysis? It is the process of combining the quantitative results of separate (but similar) studies by means of formal statistical methods. Statistically, the purpose is to increase the precision with which the treatment effect of an intervention can be estimated. Stated in another way, one can say that meta-analysis combines the results of several studies with the purpose of addressing a set of related research hypotheses. The underlying studies can come in the form of published literature, raw data from individual clinical studies, or summary statistics in reports or abstracts. More broadly, a meta-analysis arises from a systematic review. There are three major components to a systematic review and meta-analysis. The systematic review starts with the formulation of the research question and hypotheses. Clinical or substantive insight about the particular domain of research often identifies not only the unmet investigative needs, but helps prepare for the systematic review by defining the necessary initial parameters. These include the hypotheses, endpoints, important covariates, and exposures or treatments of interest. Like any basic or clinical research endeavor, a prospectively defined and clear study plan enhances the expected utility and applicability of the final results for ultimately influencing practice or policy. After this foundational preparation, the second component, a systematic review, commences. The systematic review proceeds with an explicit and reproducible protocol to locate and evaluate the available data. The collection, abstraction, and compilation of the data follow a more rigorous and prospectively defined objective process. The definitions, structure, and methodologies of the underlying studies must be critically appraised. Hence, both “the content” and “the infrastructure” of the underlying data are analyzed, evaluated, and systematically recorded. Unlike an informal review of the literature, this systematic disciplined approach is intended to reduce the potential for subjectivity or bias in the subsequent findings. Typically, a literature search of an online database is the starting point for gathering the data. The most common sources are MEDLINE (United States Library of Overview, Strengths, and Limitations of Systematic Reviews and Meta-Analyses
[1]
N. Laird,et al.
Meta-analysis in clinical trials.
,
1986,
Controlled clinical trials.
[2]
G. Glass.
Primary, Secondary, and Meta-Analysis of Research1
,
1976
.
[3]
G. Howard,et al.
Meta-analysis of data from the six primary prevention trials of cardiovascular events using aspirin.
,
2006,
The American journal of cardiology.
[4]
Al Bartolucci,et al.
The significance of clinical trials and the role of meta‐analyses
,
1999,
Journal of surgical oncology.
[5]
Shekelle Pg,et al.
Meta-regression Approaches: What, Why, When, and How?
,
2004
.
[6]
Paula R Williamson,et al.
Aggregate data meta‐analysis with time‐to‐event outcomes
,
2002,
Statistics in medicine.
[7]
S. Thompson,et al.
Quantifying heterogeneity in a meta‐analysis
,
2002,
Statistics in medicine.
[8]
Richard D Riley,et al.
Meta‐analysis of continuous outcomes combining individual patient data and aggregate data
,
2008,
Statistics in medicine.
[9]
R. Peto,et al.
Effects of adjuvant tamoxifen and of cytotoxic therapy on mortality in early breast cancer. An overview of 61 randomized trials among 28,896 women.
,
1989,
The New England journal of medicine.
[10]
J. Carlin.
Meta-analysis for 2 x 2 tables: a Bayesian approach.
,
1992,
Statistics in medicine.
[11]
T C Chalmers,et al.
A method for assessing the quality of a randomized control trial.
,
1981,
Controlled clinical trials.
[12]
L. Hedges,et al.
Statistical Methods for Meta-Analysis
,
1987
.
[13]
R. Peto.
Five years of tamoxifen--or more?
,
1996,
Journal of the National Cancer Institute.
[14]
M. Pittler.
Systematic Reviews in Health Care: Meta‐analysis in Context
,
2010
.
[15]
L. Simpson.
Report on Certain Enteric Fever Inoculation Statistics
,
1904,
British medical journal.
[16]
F. Benes,et al.
Meta-analysis of postmortem studies of Alzheimer's disease-like neuropathology in schizophrenia.
,
1997,
The American journal of psychiatry.
[17]
G. Smith,et al.
Bias in meta-analysis detected by a simple, graphical test
,
1997,
BMJ.
[18]
S. Thompson,et al.
How should meta‐regression analyses be undertaken and interpreted?
,
2002,
Statistics in medicine.
[19]
S Greenland,et al.
Random-effects meta-analyses are not always conservative.
,
1999,
American journal of epidemiology.