Flawed science behind alcohol breath tests?

We tend to think of flawed forensics happening in high-profile rape and murder cases because that's largely who DNA exonerations have let out when bad forensics falsely convicted them. But the truth is many forensic methods and technologies in use today have not been vetted to a high standard of certainty, including forensics in more common cases like alcohol breath tests used at DWI stops.

Walter Reaves out of Waco points to this recent article by Dr. Michael P. Hlastala from the Journal of Forensic Sciences titled "Paradigm Shift for the Alcohol Breath Test" (pdf) that raises questions about DWI breath test technology. Writes Hlastala:

The breath test is an indirect test, but has been considered to be a good estimate of the BAC because of the assumption that an end-exhaled breath sample accurately reflects the alveolar (or deep lung) air alcohol concentration which is thought to be in equilibrium with the blood in the pulmonary circulation. In spite of the considerable effort that has gone into the studies attempting to validate the breath test, forensic scientists and toxicologists still have only a very basic understanding of the physiological aspects of the alcohol breath test (ABT) and associated limitations.

As Reaves summarizes the research:

The "old paradigm" assumes the amount of alcohol in the breath remains constant as it goes through the lungs. It turns out that is not accurate. In fact, the amount varies - sometimes significantly. The result is that the actual blood alcohol level may be over or under-represented.

The new paradigm recognizes that alcohol is deposited in the airway surfaces during both inspiration and expiration. It also recognizes that the alcohol that comes out in the breath test comes from airway surfaces rather than the alveolar region.

Indeed, found Dr. Hlastala:

All of the alcohol exhaled at the mouth comes from the airway surface via the bronchial circulation. Very little, if any, alcohol originates from the pulmonary circulation surrounding the alveoli. The fact that alcohol comes primarily from the airways is the reason why the BrAC depends on the breathing pattern. This contributes to the very large variation in the ABT readings obtained from actual subjects.

So air breath tests are based on a fundamental misunderstanding of the relation between alcohol in the breath and the bloodstream - a false belief that the exhaled air comes from the deep lungs. The researcher suggests potential policy implications for this updated understanding of breath patterns, concluding that:

Given the variation in the breath alcohol test, it might be appropriate to consider decreasing the importance of threshold levels for penalties. After further experimentation, it might be possible to define the variation due to breathing-related variables and to allow for a magnitude of error in the breath test. Penalties could be graded with a sliding scale that allows for error in the breath test and a continuously graded scale of penalties as the BrAC increases. In any case, this new recognition of the limitations of accuracy of the ABT warrants reconsideration of the breath test protocols used.

In other words, there's a margin of error on the test that's not recognized in the law but which makes using breath tests as the threshold for penalties problematic. Reaves concludes:

In Texas, limits are important for not only determining whether someone is guilty or not, but also for determining whether certain conditions are going to be imposed - such as a interlock device. The validity and accuracy of the breath test results is therefore critical.

This article does not break new ground - problems with breath testing have long been recognized. Those problems must be explored - especially in marginal cases. Where someone's future hinges on a machine, the least we can do is make sure the machine is accurate.