A novel framework for classification of selection processes in epidemiological research

Area of Science:

• Epidemiology
• Observational Research
• Biostatistics

Background:

• Selection and selection bias definitions vary across disciplines, leading to confusion.
• Current definitions conflate selection mechanisms with their consequences.
• Assessing selection's impact on study validity requires case-by-case evaluation based on research context.

Purpose of the Study:

• To develop a general framework for classifying selection processes in epidemiological research.
• To clarify terminology and underlying mechanisms of selection.
• To enhance the understanding of selection's role in epidemiological study validity.

Main Methods:

• Systematic review of original articles in epidemiological and observational research literature.
• Identification and analysis of examples of selection processes, terminology, and mechanisms.
• Classification of selection processes based on defined dimensions.

Main Results:

• Selection processes were classified along three dimensions: level (population vs. study-specific), mechanism type (in exposure vs. population composition), and timing (entry, follow-up, post-outcome).
• This classification provides a structured approach to understanding diverse selection scenarios.
• The framework highlights the multifaceted nature of selection in research.

Conclusions:

• Improved understanding of selection processes is crucial for enhancing the validity of epidemiological research.
• A clear framework for classifying selection aids in identifying and mitigating potential biases.
• This work contributes to more rigorous and reliable observational studies.