Each year, there are between 25,000 and 30,000 arrests for driving under the influence in this state, according to the Colorado Bureau of Investigation. The vast majority of these cases never make it to trial. Instead, the defendant takes whatever deal the prosecutor offers.
The reason, of course, is that their drunkenness has already been quantified.
In Colorado, the lowest blood-alcohol content that is prosecutable for adults is .05 — driving while ability impaired, or DWAI. A .08 is a DUI, which carries with it a penalty of nine months' suspended license. A .17 could mean the installation of a breathalyzer on your ignition for two years. A .2 is a mandatory jail sentence.
By a matter of .001, blood tests determine degrees of guilt, or innocence. Produced by a governmental toxicology lab, they're usually the key piece of evidence in a DUI case, and hold immense credibility for prosecutors, judges and juries — and, thus, the people charged.
"They might as well come from the mountain," says Oklahoma-based attorney Josh Lee, a nationally recognized expert on forensic science. "Because if a lab person said it, it must be gospel."
But very little, Lee says, should be taken for granted in forensic science.
In 2009, the National Academy of Sciences called for serious reform to the field, stating that "Forensic science facilities exhibit wide variability in capacity, oversight, staffing, certification, and accreditation across federal and state jurisdictions. Too often they have inadequate educational programs, and they typically lack mandatory and enforceable standards, founded on rigorous research and testing, certification requirements, and accreditation programs."
Since the issuance of that report, legislation has been proposed at the federal level to implement and standardize scientific best practices. It's gone nowhere.
Meanwhile, the list of city or county labs that have experienced problems has only grown. As Matt Clarke wrote in "Crime Labs in Crisis," published in Oct. 2010 by Prison Legal News magazine, metropolitan areas that have seen scandals yield "full or partial closures, reorganizations, investigations or firings" include Baltimore, Chicago and Detroit. Also on the list is Colorado Springs, since El Paso County's Metro Crime Lab's blood alcohol testing section was shut down after 200 tests used in drunk driving cases were discovered to be flawed in 2010.
Since the Metro lab stopped analyzing blood alcohol, local samples have been tested by the Colorado Department of Public Health and Environment toxicology lab in Denver. Of course, state labs don't always fare much better than city labs. Right now, a scandal in Massachusetts is threatening to undo thousands of already-decided cases. And Clarke's 2010 report names about a dozen embattled state-level operations.
Colorado's lab joined those distressing ranks earlier this year. One of its technicians had performed multiple faulty tests, and his results were not immediately caught by the procedural safeguards. A flurry of media detailed the supposed failures of that lone technician, who was quickly fired.
But defense attorneys who specialize in DUI cases will argue that the lab's springtime scandal was just the beginning. Add in its arguably lax standards, failure to obtain national accreditation, reliance on a controversial leader, and insistence on secrecy, and that little blood-alcohol number deciding your legal fate looks far from unimpeachable.
On April 19 of this year, the supervisor of the state laboratory, Cynthia Burbach, swore in an affidavit that her department was in the process of retesting roughly 1,700 samples handled by technician Mitchell Fox-Rivera.
In March, Burbach found that a discrepancy called into question one of the blood tests run by Fox-Rivera. According to the district attorney involved in the case, the state said the blood sample came in at .181; but when an independent lab ran its own test, it came in at .246.
Five days later, Fox-Rivera, a recent college grad with less than six months on the job, was canned.
In her affidavit, Burbach stated the bad results were due to Fox-Rivera's inability to "properly operate a standard piece of equipment," and his failure to follow the lab's standard operating procedure for gas chromatography testing (see "The vapor trail," below). She stated that additional procedures had since been instituted to prevent such mistakes in the future.
The "one bad apple" argument should sound familiar in Colorado Springs: As the Indy reported in May 2010, the internal investigation that led the Metro Crime Lab to surrender its own certification "found 167 flawed tests in 2009 and 39 more in 2007, all of them tied to a single chemist who no longer works there."
But during the state scandal, defense attorneys were troubled by more than Fox-Rivera's problems.
In her April 19 affidavit, Burbach swore that re-tests of Fox-Rivera's blood samples yielded none with a lower blood-alcohol content than first (and wrongly) reported. She'd repeat the claim in another sworn affidavit the following day.
However, Burbach actually had signed off on at least two blood-alcohol results that did come in lower than Fox-Rivera's initial findings. In one, the inaccurate .146 was revised to the lower .134. In May, the Denver Post reported that Burbach had signed off on an even earlier re-test that revised a defendant's blood alcohol content from .218, which carries a mandatory jail sentence, to .199, which does not.
Given that, Burbach's sworn statement would seem indefensible. But CDPHE told the Post that since these results fall within 10 percent of one another, they are effectively the same.
In an e-mail to the Post, Mark Salley, CDPHE's communications director, wrote: "The standard operating procedure for the state lab allows for a 10 percent variation between results when the same sample is run multiple times ... Significant variation, in this case is any variation outside of 20 percent."
The idea that the state lab could accept a difference of up to a 20 percent variance in blood-alcohol test results was news to a lot of defense attorneys. Nowhere on the litigation packet they receive in a case, they claim, does it state that their client's alleged BAC comes via testing that allows for wiggle room of up to 20 percent.
Further, they ask, where does CDPHE get this number?
Vince Todd, a former attorney who now works as a legal consultant out of his Lakewood office, submitted a Colorado Open Records Act request seeking "Any professional standards that justify stating, without qualification, that an 'actual' test value was not lower that [sic] a comparative value if the result were within 20%."
The state responded by pointing to a guideline from the Society of Forensic Toxicologists that states that "A maximum deviation of ±20% of the mean is recommended." But that guideline is for all kinds of toxicology tests, from testing for the tranquilizer clonazepam to LSD.
As explained by Dr. Robert Lantz, with Fort Collins' Rocky Mountain Instrumental Laboratories, testing for LSD occurs at levels of a trillionth of a gram.
"So we are talking about extremely low levels," Lantz says. "Twenty percent is reasonable if you are looking at a drug level down at the low-end of what you can detect. But it's not reasonable for alcohol, because the alcohol levels are huge. And it is an easy assay. So it's really not reasonable to shoot for 20 percent."
If his laboratory comes back with a test with a more than 2 percent variance, Lantz says, they reject it.
Fox-Rivera, who couldn't be contacted for this article, challenged his termination. In his claim before the State Personnel Board, he stated that as his supervisor, Burbach was responsible for examining his work. The Board eventually ruled that Fox-Rivera was terminated rightfully, as he was a probationary employee. (According to his attorney, Andrew Brake, Fox-Rivera will be filing a civil suit.)
Burden of proof
Burbach, who has worked for CDPHE since 1984, has been the state's witness in hundreds of DUI cases. She's called to explain in detail the results that her lab provides, what those results mean, and how they were obtained. So she's left a significant trail of testimony.
As we detailed in an Aug. 29 article ("Questioning the witness," News), Gary Pirosko, a Denver-based DUI attorney, employed an intern to wade through these court transcripts, looking for inconsistencies and damaging statements. He has included their findings in a motion that he claims to have filed 40 to 50 times in the past couple years. Among those findings:
Burbach has claimed to have taught at the University of Colorado at Colorado Springs, even though the university has no record of having paid her (which the Indy verified with the school's public information office). Burbach appears to have lifted wholesale, and without attribution, language from a scientific journal, then presented it as her own in a sworn affidavit.
Also, Pirosko explains in his motion, Burbach's estimates of how many urine-sample tests she has personally conducted have increased dramatically from one telling to another. One month, she testified 150,000; later that month, it was 300,000; a year later, it was more than a million. (Pirosko figures that to process a million samples, she'd have had to worked for 3,000 years.)
But most concerning for Pirosko, and the number of attorneys who are familiar with his research, was her claim that she had dual bachelor's degrees in biology and chemistry — even though she has in other testimony admitted that she only has a biology degree. To date, Burbach and CDPHE have refused to release her school records, and her alma mater, New Mexico State University, has declined to provide that information to the Indy.
According to the department, Burbach is unavailable for interviews.
"The problem is openness," says Lee, the national expert. "That's a huge problem with Cynthia Burbach. How can you be a scientist without everything being open? There's no 'trust me' in science. All science is open except forensic science."
This lack of transparency, say defense attorneys, is endemic with the lab, and often allowed by the courts.
Take, for example, the motion to compel discovery — a litigation tactic to acquire information from your opponent — that Denver-based attorney Rhidian Orr filed in a Jefferson County case this past September. The motion requested 51 items from the lab, such as "a drawn to scale floor plan of the entire laboratory facility," "a description of the laboratory's HVAC (heating/ventilation/air conditioning)," "a copy of the laboratory's procurement and receipt records for gloves used by analysts," "a copy of the laboratory's contamination control policies and procedures," and so on. Orr argues this information is critical to mount a defense against the lab's results.
All but four items he requested were denied by the judge in that trial.
However, as of this writing, nationally known Certified Quality Auditor Janine Arvizu was preparing to testify in another case of Orr's in Arapahoe County that each of his requested materials is necessary to evaluate the working environment and quality controls of the lab; the training of the staff; and the procedures and protocols followed by the technicians.
Another transparency issue they cite has to do with the information generated by the lab's gas-chromatography test. Called a gas chromatogram, it looks something like an EKG reading, a graph with peaks that rise off a baseline.
Orr provided the Indy with a number of chromatograms he finds disturbing. One is from the run of a "negative" sample, used to gauge the machinery's accuracy.
For this sample, the lab is only testing the presence of the internal standard, n-Propanol. However, in this readout there are six foreign peaks, including one at .001 for ethanol, the pure alcohol that shows up in an intoxicated person's blood.
According to CDPHE, ethanol in "this particular sample presents no problem and would be considered negative or non-detected. Any blood alcohol result less than 0.010 is reported as negative or non-detected."
Lantz says .001 is below his lab's cut-off as well. But viewing a copy of the test, Lee calls it worrisome.
"That is absolute proof of contamination," he says. "How did you get ethanol in a standard that has no ethanol in it?
"Think about it this way," he continues. "I'm testing 100 samples, and I'm going to put your sample in there. And the guy before you is drunker than Cooter Brown. And they used the same needle, the same line, the same column that tests his [to test] yours. How do you know that you are not being affected by his? Easiest way to figure it out is to put a blank between them. It will have nothing in it. When they run it, it should have no ethanol in it."
But lawyers like Orr aren't empowered to determine why these extra peaks show up. The lab does not release the raw data produced by the testing; the chromatograms that lawyers receive are simply cumulative snapshots of the test results, with no further explanations given.
"I can take a sample that we have pulled out of our lab a year ago, and interpret it one way in the software, and get a .077, which is below the legal limit for the DUI," says Lee. "I can test it another way, and get it above a .08. I can get it as high as a .15 or .16. And if I gave you just the chromatograms, you would never be able to tell the difference."
Local defense attorney Timothy Bussey notes that DUI evidence is allowed in more easily than evidence in other criminal cases. And to understand its value, you must see it from two perspectives.
"One," he says, "you need to look at it from the perspective of the juror, who knows nothing about the system, but who has always been told that a person above .08 is guilty. And then you take a look at it from the perspective of a judge."
State law directs judges to allow test results generated by the toxicology lab, even though these tests might not even meet the minimum quality standards set by CDPHE itself. "Judges have been told by the Legislature that they don't have to have complete compliance," Bussey says, "only substantial compliance."
Indeed, the DUI statute states that strict compliance to the rules "shall not be a prerequisite to the admissibility of the test results."
"It's the government's burden to get this right," Bussey argues. "It's not my client's burden to prove that it's wrong."
In March of this year, Burbach testified in court that one of her responsibilities as the lab director was to "work on our national accreditation."
In 2010, the state lab, which also oversees urine and pill testing, had failed in its application for certification from the American Board of Forensic Toxicology, one of a few national accreditation programs available for toxicology laboratories. The lab was then, and is still, certified solely by CDPHE's own Laboratory Certification Program.
Todd filed a CORA request in June of this year seeking "any and all correspondence with any professional forensic association relating to any request for any accreditation or certification of the Toxicology Laboratory ..."
The response came back: "The department has no responsive documents."
Todd, incredulous, threatened to sue. The state responded that it had, in fact, once possessed documents relating to certification, but that a computer virus had wiped them out.
Todd continued to push, and the state soon provided the application Burbach had filed with ABFT in 2010. But Todd wanted a copy of the denial, which would lay out the lab's problems. He finally received it, after more back-and-forth.
In a detailed inspection summary dated Oct. 1, 2010, the inspectors reported, among other complaints, that the lab was critically under-funded; that key personnel were stretched too thin in their duties; and that procedures and protocols weren't fully developed. They also said those procedures and protocols were not "deemed to be fully acceptable for forensic analytical toxicology" and that technicians were not adhering to procedures in the Standard Operating Procedure.
ABFT recommended that corrective actions be taken, stating that "because of the extent and nature of corrective action required, and the fact that many deficiencies documented were due to non-compliance with the laboratory's existing SOP, that a second on-site compliance inspection will be required before accreditation of the laboratory can be considered. This inspection would likely involve two inspectors for two days and would be conducted at the expense of the laboratory."
Todd followed up on this revelation to find out whether, in the years following, the lab had instituted any written corrective actions. He also asked for any indication that Burbach had sent notification up the chain of command, including to the state's lab certification program, about the application and rejection. He sought documentation that would indicate whether the lab had requested the funds for a re-inspection.
The answer came back: "no responsive documents."
Which leads us back to Fox-Rivera.
Bussey filed a CORA in April of this year seeking any documents relating to the state's investigation into any inaccurate blood results since 2010. His request was denied, as the documents he sought contained "deliberative process privileged information." He also specifically requested documentation regarding the training protocols for any blood-analysis technicians who had handled inaccurate blood tests, only to be denied due to employer-employee protections.
When the investigation of Fox-Rivera's work was finally completed at the end of September, the attorneys received a copy. In it, Burbach, who oversaw the investigation, stated that 1,397 samples had been retested, and only five were found "were significantly outside the scientific range" of 20 percent. (That number has since been revised to four.)
The investigation laid the ultimate blame again on Fox-Rivera, stating that he had failed to properly use a pipette to dispense the blood into test vials, and failed to properly document results that had exceeded the scientific range. Burbach wrapped up the investigation by stating that the Toxicology Unit would be instituting changes to protocol to ensure that such mistakes would not occur in the future.
Following up, Todd submitted a CORA for any documentation that related to any official policy changes that were enacted in response to this investigation. The response: "no responsive documents."
"When you have a lab that has made mistakes," says Bussey, "you need to understand what caused those mistakes, determine if it's a systemic problem, a problem with their procedures, a problem with their oversight, a technical problem, and make it available to all of those involved in the system. When you don't make it a transparent process, you are precluding a person who has been charged with a DUI from having all of the relevant and necessary information to determine if their blood alcohol testing was done correctly.
"The lab should be an open book. It shouldn't be a Wizard of Oz scenario where we're told to trust the man behind the curtain. The science should stand on its own."