Optimizing cost-efficiency in mean exposure assessment--cost functions reconsidered.

Area of Science:

• Epidemiology
• Biostatistics
• Occupational Health

Background:

• Reliable exposure data is crucial for medical epidemiology and intervention studies.
• Optimizing cost-efficiency in exposure assessment is essential for maximizing statistical performance within a defined budget.
• This study extends previous optimization methods to address complex, non-linear cost scenarios.

Purpose of the Study:

• To develop and apply procedures for optimally cost-efficient allocation of measurements in exposure assessment.
• To extend existing cost models for exposure assessment to include non-linear cost functions.
• To identify optimal measurement strategies under budget constraints for epidemiological research.

Main Methods:

• Statistical performance was evaluated using a three-stage hierarchical, nested measurement model assessing exposure mean precision.
• A three-stage cost model was employed, allowing for non-linear cost variations with the number of measurements (power function).
• Procedures for optimal allocation were developed and tested across 225 scenarios with varying costs, exponents, and variance components.

Main Results:

• Linear cost functions allowed for explicit mathematical optimization rules.
• Non-linear cost functions necessitated numerical methods for optimization.
• The optimal strategy often involved single measurements from as many subjects as the budget allowed, with exceptions for high recruitment costs and low between-subject variance.

Conclusions:

• Developed analysis procedures can inform exposure assessment strategies when data on variability and costs are available.
• Shortages in empirical data on costs and cost functions currently limit generalizable conclusions on optimal exposure measurement strategies.
• Non-linearities in cost functions can significantly influence optimal allocation and data set size in exposure assessment.