Estimating mixture effects and cumulative spatial risk over time simultaneously using a Bayesian index low-rank...

Area of Science:

• Environmental Health
• Biostatistics
• Epidemiology

Background:

• The exposome concept highlights lifelong environmental exposure risks to health.
• Existing data collection on exposome components is increasing.
• Novel statistical methods are needed to analyze multidimensional exposure risks simultaneously.

Purpose of the Study:

• To introduce a Bayesian index low-rank kriging (LRK) multiple membership model (MMM).
• To simultaneously estimate health effects of multiple exposure groups and their relative importance.
• To assess cumulative spatial risk over time using residential histories.

Main Methods:

• Developed a Bayesian index LRK-MMM integrating multiple residential locations weighted by duration.
• Employed LRK for computational efficiency in analyzing complex exposure data.
• Validated the model through a simulation study for accuracy and power in identifying spatial risk regions.

Main Results:

• The Bayesian index LRK-MMM accurately estimated health effects of single and multiple exposure groups.
• The model demonstrated high power in detecting regions of elevated spatial risk from unmeasured exposures.
• A significant positive association was found between pesticide exposure index and non-Hodgkin lymphoma (NHL) risk in Iowa.
• An area of significantly elevated spatial risk for NHL was identified in Los Angeles.

Conclusions:

• The Bayesian index LRK-MMM advances the practical application of exposome research for environmental risk analysis.
• The model successfully identified specific environmental risk factors and spatial risk clusters for NHL.
• This approach offers a robust framework for integrating complex exposure data in public health research.