Facing the possibility of losing its key weapon against drunk drivers, the Florida Department of Law Enforcement decided it would pay some of its employees to — what else? — get drunk.
So, one day in October, the state’s version of the FBI shelled out $330 on Jim Beam whiskey and other booze, along with mixers and some Doritos, and invited 15 employees to headquarters to imbibe on work time.
With a video camera recording, the FDLE crime analysts, staff assistants and Capitol Police officers drank up, and then blew into three Intoxilyzer 8000s at their Tallahassee headquarters. Their blood was then drawn and sent to a lab.
The $8,000 study, put together in three days, was part of a broader push to save the reputation of the embattled Intoxilyzer 8000, FDLE records show. And in December, FDLE’s alcohol testing guru Laura Barfield came to a Sarasota County courtroom for a hearing and presented results of the drunken employee experiment to a panel of judges, saying it proved the machines were accurate.
But the study might not even be worth the $330 bar bill.
At the hearing, judges deciding the fate of the machine in Sarasota and Manatee counties seemed skeptical of even considering the study, in part because bloodwork was still at the lab and the examination was not yet finalized.
Beyond that, statistics experts say there are concerns about how the study was conducted, whether it has any scientific validity, or whether it proves what it intended to.
“That doesn’t really address the problem,” said Dr. John Robinson, a biostatistics consultant with expertise in health care. “It’s only performed at one time, with a small group of people.”
Florida’s DWI prosecutors needed something fast to defend the Intoxilyzer 8000s.
A Herald-Tribune article in October showed how flawed machines stayed in service across the state for years, unquestioned, sometimes providing impossible results about how much breath was blown into them.
Breath-test results in about 100 Manatee and Sarasota county DWI cases were thrown out as a result. A defense expert who examined one machine in Venice found the breath volume issue was skewing breath-alcohol readings.
Although FDLE insists the machines are still accurate, a coming ruling in Sarasota and Manatee could be a bombshell that excludes the use of the breath-test results from DUI cases in those counties, and maybe across the state.
The heart of the issue is simply whether FDLE rules ensure that the Intoxilyzer 8000 is reliable.
The legal debate, however, quickly becomes mired in statistical analysis, human biology and the processes inside a machine that helps convict thousands of DWI suspects in Florida each year.
An expert hired by defense attorneys analyzed all of the tests taken on Intoxilyzer 8000s across the state since they were put online in 2005, and found serious problems with the machine at Sarasota County’s jail in Venice.
That machine reported that some drivers — whose lung capacity maxes out at about five liters — blew an impossible 10, 11 or even 12 liters of breath into the machine, according to the analysis by expert Thomas Workman.
Drivers tested on that machine were four times as likely to test a blood-alcohol level above 0.25 percent, more than three times the 0.08 percent limit at which the state considers a driver impaired. The drivers were also three times more likely to test below 0.08, which also puts the machine’s results in question.
“It was getting it wrong,” Workman said. “And the statistics show that.”
FDLE fixed the machine in August, simply saying the breath flow measurement was “calibrated.”
The agency then started testing that part of the machine for the first time in January, and so far has found around 40 percent of the machines needed similar calibration. Officials did not record how far out of calibration each machine was, or how many past tests could have been faulty.
That prompted defense attorneys from across the state to file similar motions challenging the reliability of the Intoxilyzer 8000 in their area.
Venice defense attorney Robert Harrison argues that the Venice machine proves that FDLE’s rules, which do not require testing of the breath flow, are insufficient to ensure the machine is providing reliable results.
The breath volume has always been considered important, ensuring that air comes from the deep part of the lungs, where blood and air interact inside the body. The first 1.1 liters of exhaled breath can come from the mouth, throat and esophagus and contain alcohol still in the mouth or other contaminants that can produce artificially high results.
But Barfield, at the hearing, said the breath volume is meaningless to the accuracy of the results. Each driver is tested at least two times, and if those results agree with each other then the test is accurate, she said.
Barfield, who was in charge of approving the machines for use in Florida and writing the rules for maintaining them, can point to no scientific publication that supports her view, so she commissioned the drunken office study to show breath volume is meaningless.
That stance is a complete reversal of Barfield’s position on the issue when the machines were first brought to Florida. Back in 2005, FDLE did not want to report the breath-alcohol volume for any test where the breath sample was too small.
Prosecutors complained, writing a letter to FDLE that said the information must be retained under court rules.
Barfield and FDLE caved, but with a caveat. The volume would be reported when the breath sample was too small, but after the number would be printed: “Breath Sample Not Reliable to Determine Breath Alcohol Level.”
So now, as Barfield testifies the volume is meaningless in Sarasota County, prosecutors across the state still use documents with her language that says a low volume means the sample is not reliable.
“That’s how much it meant to Laura Barfield at the time,” said Matthew Malhiot, a former Intoxilyzer 8000 inspector for FDLE who is now an expert witness on the machine. “To me it’s very contradictory.”
Every state that uses the Intoxilyzer 8000 requires a minimum breath sample, Malhiot said.
The Intoxilyzer 8000 has three safeguards to ensure reliable results, including the volume of the breath sample. “It’s like a three-legged barstool. You’ve taken one of those legs away, how stable is the stool?” Malhiot said.
At the October experiment, 15 FDLE employees were given doses of alcohol. Barfield was assigned to oversee that, even following female test subjects to the bathroom to make sure no additional alcohol was consumed.
The test set up three machines: one calibrated to accurately measure breath volume, one to measure too little breath, and one to measure too much breath.
The results, she said, showed that the machines agreed with each other and the machines were reliable.
Now, Sarasota and Manatee county judges are pondering who is correct. Rulings are expected any day.
BY TODD RUGER, SARASOTA HERALD TRIBUNE