(This blog was spurred on by my conversation today with a Texas DPS lab analyst in Austin that agreed to talk to me after my dissatisfaction with the noncompliance of 2 subpoena duces tecums. Specifically, "I need the mathematical analysis worksheet to verify the numbers.." Response: "We don't have that." Then: "I need the calibration curve report so that I can do my own mathematical verification..." Response: "We use single point calibration." The most concerning statement was that they just have "the machine do the numbers." To the analyst's credit he does not write the regulations, he is responsive and direct in telling me what the lab can and can't provide but my job is not to make nice nice, my job is to get the data, analyze the data and render a professional opinion to my client as to its validity. Clearly it is time for all us Americans to demand better in our labs. It is time to educate the judges so they don't tolerate this and require proper scientific protocols. It is time for the legislature to understand that nothing is more important than one's liberty (NO not health care, social security, taxes...) and for them to regulate toxicology lab work via proper scientific procedures that are well established and the best and most cutting edge in the scientific community. The "ASCLD" lab certification that government labs have (because most can't qualify for any other) is a scientific joke. For mere insight: http://gritsforbreakfast.blogspot.com/2012/03/is-ascldlab-crime-lab-accreditation.html It is truly time that we bring truth to "government science."
Now back to the drunk driving blood test: gas chromatography blood tests are only as good as the math behind the numbers generated by each chromatogram. The math is only as good as the formulas/calibrations used. In the scientific peer review community it is well established that you cannot calculate a slope from a single point. In order to calibrate you must verify at more than one point. At minimal you use 3-5 points (the more the better). Texas DPS, probably the largest crime lab in the US (except maybe to California) uses a single point calibration. They do not mathematically verify any of the machine's calculations in DWI blood testing. What does this mean ? I quote the Institute for Chromatography "The single point calibration line is source of serious quantity errors." There is no way to accurately mathematically quantify the % of error using a single point. My fellow Texans, this is no different than craps.... the number is just a random one. Time for the bullshit to end. As a defense lawyer I am crushed with emotion thinking about all the people who have been convicted trusting the Texas DPS labs.
Some great basic textbook reading on GC calibration: http://www.interchromforum.com/html/sampling_calibration.html
Now back to the drunk driving blood test: gas chromatography blood tests are only as good as the math behind the numbers generated by each chromatogram. The math is only as good as the formulas/calibrations used. In the scientific peer review community it is well established that you cannot calculate a slope from a single point. In order to calibrate you must verify at more than one point. At minimal you use 3-5 points (the more the better). Texas DPS, probably the largest crime lab in the US (except maybe to California) uses a single point calibration. They do not mathematically verify any of the machine's calculations in DWI blood testing. What does this mean ? I quote the Institute for Chromatography "The single point calibration line is source of serious quantity errors." There is no way to accurately mathematically quantify the % of error using a single point. My fellow Texans, this is no different than craps.... the number is just a random one. Time for the bullshit to end. As a defense lawyer I am crushed with emotion thinking about all the people who have been convicted trusting the Texas DPS labs.
Some great basic textbook reading on GC calibration: http://www.interchromforum.com/html/sampling_calibration.html
Comments