Quantile Regression in Epidemiology: Capturing Heterogeneity Beyond the Mean.

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Ordinary linear regression is standard in epidemiology but assumes normality, homoscedasticity, and linearity.
• These assumptions are often unmet in biomedical data, potentially masking outcome distribution heterogeneity.
• Mean-based estimates may not fully represent complex relationships in epidemiological studies.

Purpose of the Study:

• To demonstrate the advantages of quantile regression over ordinary regression for analyzing epidemiological data.
• To illustrate how quantile regression captures covariate effects across the entire outcome distribution.
• To highlight the interpretive benefits of distribution-sensitive modeling in health research.

Main Methods:

• Utilized secondary data from a cross-sectional study of 1415 healthy Greek adults (aged 25-82 years).
• Applied both ordinary and quantile regression to model Body Mass Index (BMI) using predictors: sex, age, physical activity, dieting status, and daily energy intake.
• Estimated associations at the 25th, 50th, 75th, and 90th BMI quantiles.

Main Results:

• Ordinary regression showed positive associations of BMI with age and energy intake, and negative with physical activity.
• Quantile regression revealed these associations varied across the BMI distribution.
• The inverse association with physical activity strengthened at higher BMI quantiles, and gender effects reversed at the upper tail.

Conclusions:

• Quantile regression provides a distribution-sensitive alternative to ordinary regression in epidemiological research.
• It offers deeper insights into how covariates affect outcomes at different distribution points, unlike mean-based models.
• Quantile regression serves as a robust analytical tool and educational framework for understanding complex health data heterogeneity.