Reviews Detail Problems at St. Paul, Minn. Crimelab

Feb. 14--Independent reviews of St. Paul police's troubled crime lab found widespread failings in staff skills, poor maintenance of testing instruments and confusing and illegible lab reports, according to documents released Thursday.

All aspects of the lab's work were reviewed last year after two public defenders challenged its scientific credibility in drug cases, prompting major changes that have slowly unfolded over the last several months. Integrated Forensic Laboratories reviewed 100 controlled substance cases tested by the lab, and found that "the analyses conducted by the St. Paul Police Department Crime Lab did not meet the minimal reporting requirements that are generally accepted by the forensic chemistry community."

"Errors were noted in the majority of case files examined, ranging from minor typographical errors to misidentification of a controlled substance," according to a report written by Integrated Forensic Laboratories (IFL). "Substances were identified using methods that were inadequate or blatantly wrong... Reports greatly overstated the results and the reporting language was not standardized."

It's unclear what effect the lab's questionable performance will have on past cases, but changes were made even before the results of the audits were made public. Police Chief Thomas Smith suspended all drug testing, removed lab director Sgt. Shay Shackle, and later suspended all of its work. Three civilian criminalists responsible mostly for narcotics testing were laid off in November, and the City Council last week approved $1 million in changes to the lab.

Chief Thomas Smith and Assistant Chief Kathy Wuorinen, who oversees the lab, said the department is committed to improving the lab and rebuilding trust with citizens and law enforcement partners.

"Yes, this has been a long, arduous and sometimes painful process," Smith said.

Smith said he'd like to being drug testing back in-house at some point, but could not say when that might happen.

Police leaders have said they'll move drug testing off-site by funding two positions at the accredited BCA lab and focus on earning accreditation for its in-house crime scene processing and response and fingerprinting work. New job classifications with more stringent qualifications were recently approved and posted for lab hires.

In addition to the one audit, a separate review by Schwarz Forensic Enterprises of the lab's latent print work found "no evidence of erroneous identifications" by latent print examiners, but did find "seriously deficient work." The harshest criticisms fell on the lab's controlled substance testing, echoing much of what was exposed in eight days of testimony last year in Dakota County district court when public defenders Lauri Traub and Christine Funk challenged the lab's results in four drug cases.

According to IFL's report: The lab's main testing instruments were "in very poor operating condition." The instruments had expired, dirty and contaminated parts. "Blanks," which are when solvents are run through the instruments to check for contamination, were not typically used for quality assurance.

"Each of these is detrimental to the analysis of a controlled substance," said IFL's report. "As a result, it is possible that the lab was unable to identify controlled substances, or even misidentify controlled substances and uncontrolled substances."

A key function of the lab's instruments is to create an "ion spectra" for suspected drugs that is then compared to library data of "ion spectra" for known drugs. Staff failed to address discrepancies between their data and library data, and "several reports were issued with unsatisfactory library matches" for drugs, including methamphetamine and morphine, IFL wrote. The lab does not possess a "sufficient" library of data to identify other drugs, such as acetyl codeine.

"This is disconcerting as there are several excellent and free libraries specifically tailored for controlled substance testing," the IFL report said.

Language use was flagged by IFL as problematic in a number of areas. In several instances, lab reports included descriptions that were subjective and inconsistent, writing was illegible and language was not scientifically sound.

"The term 'PROVES' was used repeatedly with only presumptive analysis," IFL wrote. "This is not a proper, scientific term..."

Other issues indicated that there was "little to no" review of test results before they were issued. In one case, Wikipedia was cited as a technical reference. In another, staff used a question mark to indicate a test result for suspected methamphetamine.

"There is no explanation or definition for this term," IFL's report said of the punctuation mark. "The sample should have been re-tested."

The lab also failed to use standardized language to mirror Minnesota drug statutes, which could result in the wrong charges being filed. IFL said that the lab showed "little working knowledge" of the state's controlled substance laws.

"Overall, the case files were found to be excessively complicated and difficult to review," IFL wrote.

The consultant recommended retesting of St. Paul's cases by an accredited lab. Many have been retested by the Minnesota Bureau of Criminal Apprehension. Two BCA tests have contradicted St. Paul's results, leading to the dismissal of one case in Ramsey County District Court. Public defenders are investigating whether they can seek post-conviction relief in thousands of older cases in the three counties -- Dakota, Ramsey and Washington -- that relied on St. Paul's lab.

Traub and Funk initially sought to throw out St. Paul's results in drug cases, but after the first few days of testimony last July, Dakota County prosecutors voluntarily decided to withdraw that lab's results. Traub and Funk pushed on, changing course by asking Judge Kathryn Davis Messerich to withhold any test results by the BCA, asserting that the evidence was first handled by the St. Paul lab and was possibly exposed to contamination.

Messerich has until next month to issue her decision.

Schwarz Forensic Enterprise's review of 246 randomly selected latent print cases found no evidence that lab staff made wrong identifications. But in 103 cases, about 42 percent, there was "seriously deficient" work because "latent prints of value were located where none had been reported," and because those prints were traced to the same or different subject in the case, the SFE report said.

Other issues identified by SFE include: No ongoing proficiency tests for staff, no standard operating procedures, no system to ensure that patrol and investigative officers who collected prints in the course of their normal duties were qualified to perform the work and no requirement to document where a print came from or its orientation.

Safety was also a problem. Staff were unfamiliar with safety procedures around chemicals and there was no ventilation or eye wash stations in the latent print processing area.

There were "numerous instances" where evidence processing did not follow best practices, the report said. In one case, SFE reported, staff personnel was wrongly trained to use a chemical process that took at least three hours to complete one step when a correctly trained technician could process the evidence in 10 minutes.

The latent print reports also posed problems. "The unit's reports were largely unintelligible," SFE wrote.

Reports were unclear in their results and used words that could have several definitions, such as "negative," which could mean "the subject was excluded" to "need better known prints."

"The respective meaning of these different conclusions can seriously affect the importance attached to the evidence," the SFE report said.

The consultant found problems with the lab's storage of digital images taken of evidence or at crime scenes. They were kept on one computer with no regularly scheduled backups, SFE wrote, and had no "meaningful" system for documenting or tracking the images. All the images were accessible by anyone using the terminal in the lab, which has a secure entrance with limited access. The department was moving the images to a server in the network, SFE reported.

SFE also reviewed 73 pre-chosen controlled substance cases for accuracy. The consultant found no contamination between cases, but found nine cases in which there was potential contamination from one item in one case to another item in the same case.

"The nine instances of potential contamination may or may not be actual contamination," SFE wrote. "However, it was observed that the laboratory practices by the lab to prevent and detect contamination were deficient."

Contamination was likely "not a substantial problem," but could have been a problem in specific cases, the report said.

Chao Xiong --612-270-4708 Twitter: @ChaoStrib