Three days earlier, Colorado Springs police had tried to put "closure" on a screwup that rocked their crime lab and cast doubt on the blood tests used to prove drunken driving.
Reading stiffly from a news release Monday, April 19, police spokesman Sgt. Steve Noblitt told about a dozen reporters and photographers that an internal investigation found 167 flawed tests in 2009 and 39 more in 2007, all of them tied to a single chemist who no longer works there.
"It is very important to recognize that these errors were only identified because of the checks and balances of the Metro Crime Lab's quality assurance program," Noblitt said reassuringly.
Next, it was District Attorney Dan May's turn to speak. Of the 206 bad tests, he said, only nine had pushed defendants over a legal threshold: .05 for driving impaired, .08 for driving under the influence and .20 for driving so drunk that jail time is required. In those cases, he explained, charges were generally dismissed and court costs refunded.
He also praised the lab for figuring out the problem: "We have every confidence in the Metro Crime Lab," he doted.
TV and daily newspaper stories echoed that confidence in the hours after that news conference. Then the subject was dropped.
But in the lab later that week, the testing problem is far from forgotten. Senior chemist Bobby Striebel frankly says he can't explain how a colleague of seven years made mistakes that boosted the alcohol content of some blood samples by more than 40 percent.
"The error was very difficult to identify," he says, with no apparent pattern or equipment failure to explain.
The lingering uncertainty is uncomfortable. Tim Bussey, a Colorado Springs defense attorney who specializes in DUI cases, puts it bluntly: "If they never really identified the problem, how do they fix it?"
Under Colorado law, blood alcohol levels are as precisely determined as mileposts on a highway. If you're under 21, you only need a blood alcohol level of .02 percent — a single drink for most people — to lose your license and get slapped with a fine, an alcohol class and hours of community service. (Underage drinking can also result in a minor in possession charge, with its own penalties.)
Commercial drivers who test at the .04 level while working lose their licenses for a year.
For most adults, the crucial levels are .05 (driving while ability impaired, or DWAI), .08 (DUI), and .20 (mandatory jail time). Also, a test result of .17 is enough for the state Department of Revenue to label you a "persistent drunk driver," earning you an ignition interlock device — a breathalyzer hooked to your ignition — for at least two years after your license is reinstated.
The choices available to anyone accused of drunken driving in Colorado are simple: You can have your blood tested, opt for a breathalyzer or reject a test altogether, which carries a stiff mandatory one-year suspension of your driver's license.
Local attorney Geoff Heim estimates that 60 percent of people choose the blood test, though he, like many defense attorneys, says the breath test gives him more avenues to question the results in court. (That, at least, was his advice before the lab mistakes became public in December; now he's not sure which is more advantageous.)
The Metro Crime Lab conducted about 2,000 tests a year from 2006 to 2009 for the Colorado Springs Police Department, the El Paso County Sheriff's Office, Manitou Springs police and about a half-dozen other law enforcement agencies.
Those who choose the blood-test route actually have two tubes of blood drawn. One tube is used to prove a crime was committed; the other is left sealed for defendants who want a retest.
This normally happens only with defendants who have private attorneys; Striebel estimates that about 25 percent of the second samples are tested. But getting a second sample tested has paid off big for some people affected by the lab's screw-up.
In January 2009, Heim's colleague Michael Moran had a client whose first blood test came back at a whopping .222, a ticket to jail. Moran's retest came back at .173, and prosecutors were willing to deal. The client pleaded guilty to DWAI, agreeing to alcohol classes and public service.
No one told the lab there was a problem with that case. According to lab director Ian Fitch, it wasn't until November that alarm bells started ringing, when one of the lab's three chemists was found to have failed an Oct. 20 proficiency test.
Such tests are required for a lab to be certified in Colorado. The state runs one lab in Denver, and three police departments (including CSPD) hold the certification, as well as a few private labs.
One of those private labs is Rocky Mountain Instrumental Laboratories in Fort Collins, and that's where Colorado Springs attorney Sarah Christensen sent a sample in November after the state's test for her client came back "significantly over .20."
The retest came in more than 30 percent lower than the state's test from Sept. 24, dropping her client below even the .17 level.
Stunned, Christensen called the crime lab's Fitch. At the time, he and Striebel were trying to gauge the extent of the problems, and he says Christensen's complaint — the first from an attorney — "made us realize the scope was broader than we thought."
On Dec. 11, police went public, announcing that "approximately" 82 tests had been inflated. The number climbed to 167 in January, then 206 by April 19.
The errors police reported generally came in batches of about 20 tests, each from a separate day of testing. That some of them were off by 20 percent or more stunned Rocky Mountain co-director Patricia Sulik.
"Until this year, you usually only see small differences," says the doctor of chemistry.
With the results from three bad batches in January 2009 having shown up in court in March and April, Sulik wonders why no one raised an alarm earlier last year: "The problem would have been a lot smaller."
Senior deputy district attorney Frederick Stein has no simple answer. He says the dozen or so prosecutors who handle DUI cases focus on the particular cases in front of them. And most, he says, greet test results from private labs with a "healthy skepticism," seeing a profit motive for private labs in challenging the state's results.
Where it happened
Though the forensic specialties offered in the Metro Crime Lab sound impressive — DNA analysis, blood-stain pattern analysis and fire debris analysis, to name a few — the lab in the basement of the Police Operations Center feels a world apart from CSI. Fully staffed, the lab has a director, a firearms expert, two DNA analysts, six crime scene technicians and three chemists, each of whom has an "office" reminiscent of a high school chemistry lab.
The office that belonged to Marcie Jardell, who started working for the lab in 2002, teems with files and books alongside the beakers, flasks and Bunsen burners.
During a lab visit, Striebel talks cheerfully about the arcane details of testing, carefully avoiding any discussion about Jardell or her status. (She's set for a meeting in mid-May with police Chief Richard Myers to appeal her March dismissal.)
From a chemist's point of view, blood alcohol tests are simple: You prepare a batch of maybe 20 samples in glass vials, each with a precise amount of blood, a couple other ingredients, and a solution containing n-propanol, a form of alcohol that is almost never ingested. Each sample gets run twice, along with more than a dozen other solutions used as calibrators and controls.
Once the vials are sealed and catalogued, a machine detects the amounts of ethanol (the stuff in booze) and n-propanol in each, which show up as peaks in a resulting graph. Since the n-propanol amount is supposed to be consistent in all vials, the blood alcohol level can be calculated from the ratio of ethanol to n-propanol.
"All of her ethanol peaks were accurate," he says, as were the n-propanol peaks for the calibrators and controls. In bad batches, however, her n-propanol peaks were consistently — and inexplicably — lower.
Striebel couldn't find a way to reproduce those results, nor any pattern that indicated why tests run on some days were good, while they were bad on others.
"She analyzed roughly half of all the blood in our lab for the last seven years," and was right most of the time, Striebel says of Jardell.
Even refrigerated, ethanol breaks down in blood samples, so Striebel was only able to retest samples from 2009. He tested two samples from each batch to determine if the results were accurate or not.
For every batch since the start of 2006, he also recorded the size of every n-propanol peak, looking for discrepancies. That's how he identified 39 errors in 2007.
Spreadsheets showing the n-propanol levels will be a standard part of documentation from each batch of blood tests performed in the lab. A new procedure also requires one sample from each batch to be re-tested with the next batch to make sure the values agree.
Such procedures may help detect mistakes in the future, but they do little to explain what happened.
Striebel says it takes a "very detail-oriented person to be consistent from beginning to end," and adds, "One of my theories is that [the] chemist became complacent."
Yet going back to 2002, supervisors consistently rated Jardell as "effective" or "excellent," with no hint of problems.
Jardell declined to speak for this story. Reached by telephone, her attorney, Jeanne Wilson, spoke only briefly, saying she doesn't yet know if there will be a lawsuit. First, Jardell will meet with the police chief. Asked if there's a different side to the story, Wilson sounded buoyant: "There definitely is."