I heard an interesting story on NPR this morning while driving to the office from court that demonstrates why there should always be reasonable doubt about breath test results from the Intoxilyzer 8000, which is used in all Kansas DUI breath tests, and many in Missouri: Over one hundred years ago scientist forged a kilogram in London, meticulously weighed it, and sent it to Paris where it was placed in a box and stored. It is the standard against which all other weights in the world are measured. It has only been taken out and looked at by scientists 3 times in the last one hundred years. Unfortunately, “The Kilogram”, as it is known, has lost some weight inexplicably and it is a big problem for science. It has only lost about 50 billionths of a kilogram, but for science, which is concerned with precise measurements, that is a big problem. So, there is an effort underway to replace “The Kilogram” so that measurements can be precise.
The Intoxilyer 8000 is supposed to measure the grams of alcohol per 210 liters of breath. 210 liters is roughly the size of a 55 gallon drum barrell. Of course, no one could ever blow 210 liters of breath into a machine. The sample chamber of the Intoxilyzer 8000 is only 29 cc. The machine has to multiply 29 cc enough to make it as large as a 55 gallon drum, or 105 2-liter bottles. If you do the math, it has to multiply 29 cc a total of 7,421 times to equal 210 liters. To get to a .08, the amount of alcohol in a 29cc chamber would be literally microscopic! That means any error, to even the slightest degree, in the measurement of 1 cc will cause the machine to multiply that error times 7,142.
The breath test result can be wrong if the measurement is not precise, down to the millionth of a gram. Alcohol is more concentrated in the blood than in the breath. Therefore, some conversion factor is needed to convert breath concentrations to blood concentrations. In breath testing that conversion factor is known as the “partition ratio.” The partition ratio generally used is that of 2100:1 – i.e. the amount of alcohol contained in 2100 milliliters of breath is the same amount of alcohol that is found in 1 milliliter of blood. This partition ratio of 2100:1 is based upon several assumptions. The primary assumptions that affect breath testing are:
1. That the person being tested is in the elimination phase and is no longer absorbing alcohol into the bloodstream;
2. That the test subject’s expired breath has a temperature of 34° C or 93.2° F and that this temperature corresponds to the same core temperature in all individuals;
3. That the test subject’s blood particulate (hematocrit) levels are “normal;”
4. That all of the alcohol that passes into the test chamber comes from the deep lung air of the test subject; and
5. The only substance that is absorbing light in the test chamber is ethanol.
Inaccuracies in these assumptions create inaccuracies in the estimation of blood alcohol from breath alcohol by the Intoxilyzer 8000.
Variations from these assumptions create variations in the partition ratio. Measured variations in the post-absorptive partition ratio encompassing two standard deviations of 95% of the population run from a low of 1797:1 to a high of 2763:1. When that range is extended to cover 99.7% of the population or three standard deviations, the range is 1555:1 to 3005:1. Individuals with a partition ratio of 1797:1 who are tested at .080% and .090% on an Intoxilyzer 8000 have actual blood-alcohol concentrations of .068 % and 077%, respectively. Individuals with a partition ratio of 1555:1 who are tested at .080%, .090% and .100% on an Intoxilyzer 8000 have actual blood-alcohol concentrations of .059%, .067% and .074, respectively. Moreover, as noted above, these ranges are for post-absorptive individuals, i.e. individuals in the elimination phase. Variations in the absorptive phase are even greater.
So, the Intoxilyzer 8000 makes a host of assumptions about the person blowing into the machine, including his or her partition ratio, breath and body temperature, and hematocrit level. Any deviation from these assumptions and the error is magnified thousands of times by the machine. Of course, assuming that everyone is exactly the same makes as much sense as assuming that everyone in America has a 30 inch waist.
Real scientists are deeply concerned about the 50 billionths of a kilogram which “The Kilogram” has lost. All other weight measurements in the world hinge on this. Accurate science requires accurate measurements. Kansas DUI cases, and driving privileges, hinge on the Intoxilyzer 8000 being accurate. Yet, the “science” of Kansas DUI does not require accurate measurement. Good science would also require two breath tests, even Santa Claus checks his list twice. But we don’t do a confirmation breath test in Kansas. That is bad science. Anyone that says “close is good enough for government work” is not a scientist, but merely a spokesmodel for Intoxilyzer and the prosecution.
The story is interesting and you can find it here: http://www.npr.org/templates/story/story.php?storyId=112003322