Comparison of web versus interview participants in a case-control study

Area of Science:

• Epidemiology
• Biostatistics
• Public Health Research

Background:

• Internet-based studies offer convenience but may introduce self-selection bias.
• Understanding participant recruitment methods is crucial for valid epidemiological research.
• Case-control studies rely on accurate participant selection to minimize bias.

Purpose of the Study:

• To assess potential selection bias from differential self-selection in an internet-based case-control study.
• To compare sociodemographic and behavioral factors between internet and telephone recruitment groups.
• To evaluate the impact of recruitment method on identifying risk factors, such as maternal smoking in clubfoot.

Main Methods:

• Comparison of sociodemographic and behavioral factors between cases and controls recruited via the internet (Web group) versus telephone (Phone group).
• Utilized data from an existing clubfoot study to analyze recruitment method differences.
• Statistical analysis to determine odds ratios for risk factors in each recruitment group.

Main Results:

• Mothers in the Web group were more likely to be non-Hispanic White, non-smokers, more educated, wealthier, and used more medications/vitamins during pregnancy.
• These demographic and behavioral differences were more pronounced in case mothers, indicating potential selection bias.
• Maternal smoking, a known clubfoot risk factor, was confirmed in the Phone group (OR=1.61) but not observed in the Web group (OR=1.15).

Conclusions:

• Solely internet-based enrollment in case-control studies carries a significant risk of selection bias.
• Differential self-selection can obscure or alter the identification of established risk factors.
• Researchers should carefully consider recruitment strategies to ensure representativeness and minimize bias in online epidemiological studies.