Quantitative bias analysis methods for summary-level epidemiologic data in the peer-reviewed literature: a systematic...

Area of Science:

• Epidemiology
• Biostatistics
• Health Research Methods

Background:

• Quantitative bias analysis (QBA) methods are crucial for evaluating systematic errors in observational studies.
• Assessing the impact of biases on study results is essential for reliable scientific conclusions.
• A comprehensive understanding of available QBA methods for summary-level data is needed.

Purpose of the Study:

• To systematically review and characterize quantitative bias analysis (QBA) methods for summary-level data.
• To identify the range of biases addressed by existing QBA methods.
• To summarize the applicability and features of published QBA methods.

Main Methods:

• Systematic literature search across major databases (MEDLINE, Embase, Scopus, Web of Science).
• Inclusion of English-language articles describing QBA methods for summary-level data.
• Extraction and recording of key characteristics: study designs, biases addressed, bias parameters, and software availability.

Main Results:

• Identified 57 QBA methods from 53 articles, primarily for observational studies (93%).
• Most methods addressed unmeasured confounding (51%), followed by misclassification (33%) and selection bias (11%).
• A majority (67%) generated bias-adjusted estimates, while 39% provided accessible code or tools.

Conclusions:

• A diverse set of 57 QBA methods for summary-level epidemiologic data exists in the literature.
• This systematic review provides a valuable resource for researchers seeking appropriate QBA methods.
• Future research can leverage this summary to select and apply suitable bias analysis techniques.