[Blog editor’s note: Joshua Goldberg of Pittsburgh, PA writes into us to share information about Breath Alcohol Content machines.]

Additional  Breath Testing Device Accuracy Limitations

The main three issues with breath testing are:  (1) an assumed partition ratio; (2) an assumed breath temperature; and (3) measuring devices not specific to ethyl alcohol.  Breath testing faces additional issues including:

1)   Breath volume/ Breathing Patterns:

The rate of breathing affects the concentration of ethanol in the breath. Hyperventilation causes a lower breath result (up to 55%) while hypoventilation (breath holding) increases the breath result (up to 14% higher than actual BAC) (Ohlsson et. al. 1990). Both breathing disorders can be caused by disease, trauma, and drugs and should be considered as a source of potential error in breath testing.

Reference: Stefan Rose and Kenneth Furton, Variables Affecting the Accuracy and Precision of Breath Alcohol Instruments Including the Intoxilyzer 5000 4 (manuscript appended to this brief as Appendix 1).

2)  Hematocrit- hematocrit represents the fraction of whole blood composed of red cells and is correlated with the aqueous content of blood.  The higher the hematocrit, the lower the concentration of water in blood, and vice versa.  The average hematocrit for normal, healthy males is 47%, with a range of 40-54% for females, the average is 42%, and the range is 36-47%.  Given that the Breathalyzer uses only one partition ratio, Smith and Payne, et. al. have predicted that the normal variation in hematocrit can produce errors in breath test results in the 10 to 14% range.

Reference: Dominick Labianca, The Chemical Basis of the Breathalyzer – A Critical Analysis, Journal of Chemical Education, Vol. 67, Number 3, 259-61 (March 1990)(Analysis conducted using an older breath test device that does not employ infrared spectrometry).

3)  Analytical Errors

All measurements are subject to three types of experimental errors which need to be evaluated.  The first errors are systematic errors which result from instrumental, method or personal errors and result in a bias of values positive or negative relative to the true value.  Detection of this type of error is by the analysis of standard samples such as a calibration standard at the legal limit for the jurisdiction (i.e. 0.08g/210L) and correcting for any bias.

Reference: Stefan Rose and Kenneth Furton, supra. at 7-8.

The result of any measurement is only an approximation of the particular quantity being measured . . . A “result is complete” only when an instrument’s bias has been determined and the result corrected for that bias.  The reporting of BAC results is no exception. Where bias exists in a breath test machine, any results it yields must be corrected for that bias before they are reported.

Reference: Ted Vosk, Chaos Reigning, The Champion, May/ June 2008 57 (citations omitted).

Therefore, two basic assumptions, and a myriad of accuracy limitations, mar breath testing’s scientific pedigree that the court cannot overlook.