Breath alcohol measurement variability associated with different instrumentation and protocols

Area of Science:

• Forensic Science
• Analytical Chemistry
• Toxicology

Background:

• Breath alcohol measurement variability arises from instrumental, procedural, and biological factors.
• Accurate standard deviation (S.D.) estimates are essential for calculating measurement uncertainty and confidence intervals.
• Existing methods for S.D. estimation can be concentration-dependent.

Purpose of the Study:

• To derive predictive models for standard deviation (S.D.) as a function of alcohol concentration.
• To analyze duplicate breath test data from multiple jurisdictions and instruments.
• To provide a method for estimating S.D. for breath alcohol measurements.

Main Methods:

• Analysis of duplicate breath alcohol test data from four US state jurisdictions.
• Utilized data from specific instruments: Drager 7110, Intoxilyzer 5000, BAC Datamaster, and Intoximeter EC/IR.
• Developed regression models to predict S.D. based on measured alcohol concentration.

Main Results:

• All analyzed jurisdiction/instrument combinations showed forensically acceptable variation.
• A general linear increase in standard deviation with increasing alcohol concentration was observed.
• The findings are consistent with a multiplicative error model for breath alcohol testing.

Conclusions:

• The derived models provide reliable estimates of standard deviation for breath alcohol measurements.
• These models can be used by jurisdictions not performing duplicate analyses to estimate S.D.
• The study enhances the accuracy and reliability of breath alcohol testing uncertainty calculations.