Video taken during controlled alcohol testing. Subject with well-documented true blood alcohol concentration between 116 and 120 mgs/100mls blows 303 mgs/100 mls as a result of mouth alcohol bias. Mouth alcohol detection algorithm fails to flag phenomenon as "Invalid Sample". Instrument is an Intoxilyzer 5000C.

The mouth alcohol detection system is looking for a pattern in the blow, whereby the BAC concentration rises quickly and then falls. However, breathing patterns during a subject test on an approved instrument aren't always the same. There are occasions when the presentation of some alcohol in the blood combined with some alcohol in the mouth or stomach, with a long gentle blow is a presentation not contemplated by the design of the instrument. The instrument accepts the sample as a valid sample, notwithstanding the presence of mouth alcohol.

Mouth alcohol detection systems are far from foolproof. Protocols followed by legal systems and operators must therefore require:

1. 15 to 20 minute observation/deprivation period before each subject test to watch for alcohol or food placed in the mouth, burping, belching, vomit
2, two tests with good agreement 15 minutes apart
3. never use a screening device (PBT or ASD) to attempt to obtain a quantitative result (ASDs and PBTs usually don't have mouth alcohol detection systems)