The Inadequacy of Mouth Alcohol Detection Systems in Breath Alcohol Testing
So-called "slope detectors" in breath alcohol analyzers don't always catch mouth alcohol.
Undetected mouth alcohol is a problem in breath alcohol testing.
Most modern breath alcohol analyzers are programmed to look for signs of a breath sample contaminated with alcohol that doesn’t come from the lungs.
However, the algorithm to detect mouth alcohol contamination isn’t perfect.
In my experience, the slope detector doesn’t always catch mouth alcohol.
I know this because I’ve done testing in the lab when training officers about the DMT.
During one of the exercises, I would have the officers swish a small amount of Vodka in their mouths and then blow into the DMT.
And the DMT wouldn’t always flag for it. This was in the case where the subject had just swished with alcohol.
If the DMT doesn’t always catch it in the above situation, how likely not to see it when the subject has underlying alcohol already in their system?
Occasionally, the officers dosed to 0.08 g/210L would swish with one of the leftover shots of Vodka. In those situations, the instrument was even more unlikely to flag for mouth alcohol. This is because there was a mixing effect between the recently swished alcohol and the alcohol coming from the lungs.
What the manufacturer says
When teaching about mouth alcohol, John Fusco, the president of NPAS, one of the major manufacturers of breath alcohol analyzers, said,
If you do it right, you can actually take a person who is 0.05, 0.06, 0.07, and pull that reading up to 0.12, 0.20… The DataMaster will pick up 85-90% of the mouth alcohols. It does a pretty good job compared to everybody else. Some of the instruments that are out there are going to do as little as 50%.
Mouth alcohol is a bigger problem than most realize
In the past year, I’ve had several cases where the slope (red line below) of the breath alcohol concentration (BrAC) showed mouth alcohol, but the DMT didn’t detect it.
Take a look:
In situations like this, a scientist can evaluate the slope of the BrAC to determine if undetected mouth alcohol may have contaminated the breath.
It’s one of the reasons that being able to look at the expirogram post hoc and analyze the data is essential. However, not every state provides this data. And the states that do are reluctant to give it out.
In the above case, the unusual spike in BrAC was followed by an extremely negative slope, which went negative by 0.005 for 3/4 of a second.
But the slope monitoring system didn’t flag it. Why?
The slope monitoring system requires the slope to go negative by 0.003 over 3/4 of a second.
But for the DMT to flag for mouth alcohol, the slope must decrease by greater than or equal to -0.001 in three consecutive quarter-second averages.
In the above case, the slope decreased by -0.0002, -0.003, and -0.002.
The above case was so extreme in the second quarter-second average that it didn’t cause a flag in the DMT.
This is a stark example of why we can’t rely solely on machines and algorithms. As scientists, we must examine all the data and ensure it makes sense.
The scientific literature shows that slope monitoring doesn’t always work
I’m not the only one who’s found that slope monitoring doesn’t always work.
I’ve been to scientific conferences where there were presentations about cases when instruments didn’t flag for mouth alcohol when they should have.
In Sweden, they even do an 11% deduction for undetected mouth alcohol.
The papers below also mention the pitfalls of slope detection systems.
References
Wells D, Farrar J. BREATH-ALCOHOL ANALYSIS OF A SUBJECT WITH GASTRIC REGURGITATION. California Defender.
Staubus A. False high breath alcohol readings due to gastric reflux. Proc Amer Acad Forens Sci.
Hlastala MP. The Slope Detector Does Not Always Detect the Presence of Mouth Alcohol. The Champion 2006; 3: 57–61.
Labianca DA. Non-foolproof nature of slope detection technology in the Dräger Alcotest 9510. Forensic Toxicol 2018; 36: 222–224.
Fransson, Andersson L, Jones AW. Evidenzer – A New Breath-Alcohol Analyzer For Evidential Testing.
Gullberg RG. The inadequacy of instrumental ‘mouth alcohol’ detection systems in forensic breath measurement. Crime Scene 2001; 27: 9.
Simpson D, Kerby JA, E, Kerby Scott. Effects of Mouth Alcohol on Breath Alcohol Results. International Journal of Drug Testing; 3: 1–14.
Jones AW. Gastric reflux, regurgitation and potential impact of mouth-alcohol on results of breath-alcohol testing. DWI Journal Law & Science 2007; 22: 1–8.
Jones AW. Reflections on the GERD defense. DWI Journal of Law & Science 2005; 20: 3–8.