Same pattern, different result: Algorithmic inconsistency in breath alcohol testing—case reports

Breath alcohol analyzers employ algorithmic safeguards to identify samples that may be falsely elevated due to mouth alcohol contamination. One such safeguard in the DataMaster DMT (also known as the Intox DMT) is commonly referred to as the 95% rule. Under this criterion, if the final quarter-second average breath alcohol concentration (BrAC) measured during sample delivery is less than 95% of a previous high value recorded during the same sample test, the instrument is expected to generate an Invalid status message. Four cases are presented in which the stated criterion was met, yet the instrument failed to generate an Invalid status message in three of the four cases. The absence of the Invalid status message in the three cases is not explained by any documented alternative criteria or exceptions in the manufacturer's materials or operator training resources. These findings suggest that the 95% rule is either inconsistently applied or subject to additional parameters that govern its implementation. An independent review of the instrument's source code would help clarify the basis for these discrepancies. • Invalid status message did not align with stated 95% criterion. • 3 of 4 cases meeting threshold did not produce an Invalid message. • Manufacturer documentation does not explain observed inconsistency. • Findings support need for source code review and transparency. • Inconsistent message generation raises reliability concerns.