Bad apple at DPS crime lab could spoil barrel of convictions

Potentially hundreds of people convicted of drug crimes in 36 Southeast Texas counties may have their convictions overturned after the discovery last year that a scientist who'd worked on nearly 5,000 cases had falsified results. At a meeting of the Texas Forensic Science Commission (FSC) on Friday, a number of new details emerged about the episode and its aftermath. (Since no MSM reporters were there for that portion of the meeting, this account is a Grits exclusive.)

Throughout the time Jonathon Salvador worked as a controlled substances analyst at the Texas Department of Public Safety (DPS) crime lab in Houston, he was never a particularly competent employee, said Commissioner Sarah Kerrigan who headed the FSC investigation. Salvador was terminated when it was discovered he reported the contents of a batch of pills without testing them, substituting data from another sample. (To its credit, the agency self-reported the incident to the FSC as soon as it was discovered.) Upon retesting his prior three months of casework, DPS found four other cases requiring corrective actions, including one where he'd incorrectly identified a substance as marijuana. Since no one knows how often he may have made errors or falsified evidence, all convictions based on his analyses are potentially in jeopardy.

Though a DPS official told the FSC that Salvador may have left his prior job at the Los Angeles PD crime lab "under questionable circumstances," he was considered a "valuable lab member" by his peers, Prof. Kerrigan told the commission. He was well-liked by his co-workers, was considered "responsive and compliant" by his supervisors, and was the sort of employee who regularly volunteered for unwanted tasks that helped the office function. He performed well in court and his testimony was convincing to jurors. He just wasn't a very good analyst and at some point began to take shortcuts.

Salvador was sacked last year after another employee discovered he'd issued a "fraudulent" report that "misrepresented" what tests he'd performed on a batch of Alprazolam pills. Though he'd worked for the agency since 2005, performing tests in 4,944 criminal cases, Salvador's work suffered from consistently poor documentation, technique, and decisionmaking, said Kerrigan. He struggled with his caseload and at times appeared not to fully understand the science behind the work he was assigned, though he tended to cheerfully and promptly correct mistakes whenever they were identified.

FSC staff interviewed Salvador's managers and peers, among whom there was a consensus that he did not work hard and his performance was "marginal" and of "low quality." The lab's internal oversight process  identified a high rate of errors in his work requiring corrective actions, but the lab manager was "compassionate to a fault," said Prof. Kerrigan, ignoring what in retrospect should have been obvious warning signs. In the month prior to the drylabbing incident being discovered there had been two episodes of "under-performance" cited by his supervisors: One in which he failed to find a controlled substance which was actually present, and another in which he'd allowed his instrument to become contaminated for lack of proper cleaning (which can result in false positives). The lab performs "tech reviews" of 100% of its cases, and those reviews found errors in Salvador's work more often than his peers. But tech reviews don't involve retesting and wouldn't have uncovered the sorts of falsification that resulted in his termination.

The office suffered from a "culture" that "tolerated under-performance" because of high case loads, said Dr. Nizam Peerwani. DPS managers were hesitant to record problems in an employee's file for fear such blemishes would give fodder to defense attorneys to impeach testimony in court. Salvador's annual performance reviews identified recurring problems with low output, lack of attention to detail, taking shortcuts, and too many corrections in his tech reviews. But as a general practice, employees who made it through the one-year probationary period at the lab "were dug in," said a DPS manager. Kerrigan said there was  a "perception among staff that termination was unlikely" regardless of performance.

Even so, lab workers interviewed said that Salvador felt under increased pressure to perform and was worried about management's perception of his low-quality work in the months prior to his termination. Earlier in his career, he'd more frequently asked for assistance when he struggled with scientific or technical questions beyond his understanding. But lately he'd "stopped asking for help" for fear of management's perceptions about his competence, speculated his peers.

Since Salvador's termination, DPS officials say they've implemented a new, more rigorous evaluation system that will be more likely to result in terminations for under-performing employees, and new employees will receive evaluations every two months for their first year. Changing the labs' culture is complicated, though, by the fact that DPS has too few examiners, large backlogs, and expanding caseloads. Plus, training new lab workers takes many months before they can handle cases on their own. There's no short-term solution to that dynamic and the first-cut legislative budgets for DPS did not include money for additional crime lab employees.

The question now becomes, what happens to all the defendants whose cases Salvador worked on?  In those containing errors or where the evidence has been destroyed and can't be retested, defendants could be entitled to habeas relief. Conceivably, hundreds of convicted defendants may end up having their cases overturned, either freeing them from prison or ending their probation terms. In cases that originated with the Department of Public Safety, all the evidence was kept and can be retested, Kerrigan said, but smaller agencies in those 36 counties were much more "diligent" about destroying old drug evidence and there will likely be nothing to re-test in 25-50% of cases, she estimated. What a spectacular mess!

The Texas District and County Attorneys Association helpfully advised their members last year that "For any case with a bad retest, or cases with now-destroyed evidence, [they should] request that the court appoint an attorney to take the case through a writ process if appropriate." But elected District Attorneys are under no obligation to follow that advice and, as a practical matter, various jurisdictions are approaching the situation differently.

The first habeas writs from this episode came out of Galveston, where FSC general counsel Lynn Garcia told the commission that prosecutors planned to agree to dismissal on nearly all 849 cases Salvador worked on from that county (with a handful of exceptions). But the first writs that went to the Court of Criminal Appeals were remanded for further investigation. (See one of the orders [pdf], all of which are essentially similar.) The CCA asked for "more information before addressing the claim on its merits, though they did say that "Applicant has alleged facts that, if true, might entitle him to relief." In particular, the high court ordered:

The trial court [to] supplement the writ record with the investigative report from DPS and any other documentation supporting its agreed findings of fact. The trial court shall provide support for its finding that the forensic examiner in this particular case was the examiner who was found to be unreliable by the DPS investigation. The trial court shall also determine whether there is more drug evidence that could be tested, or whether any other analyst was involved in the testing of the evidence in this case. If another analyst was involved in the testing for this case, the trial court shall make findings regarding the actions of each individual analyst. Additionally, the trial court shall determine whether any additional testing was ever done on the evidence in this case, either before or after the initial proceedings in this case. The trial court shall also make any other findings of fact and conclusions of law that it deems relevant and appropriate to the disposition of Applicant’s claim for habeas corpus relief.

I'm not a lawyer but from my reading of that passage, defendants may be entitled to relief where a) evidence was never tested by another analyst, and b) the drugs were destroyed and are not available for retesting. Since DPS typically performed only "tech reviews" of Salvador's work and did not routinely have another analyst confirm the results, the court seems to have left open the door for convictions to be overturned in cases where the evidence was destroyed. That could turn out to affect a lot of cases.

In an additional twist, the court added, "The parties are reminded that under certain circumstances, an applicant may be eligible for release on bond under Article 11.65 of the Code of Criminal Procedure." The first of those orders were issued on December 5, giving the trial judge 30 days to hold a hearing on the fact issues and 60 days to report findings back to the CCA. So we may have a better idea how these cases will shake out sooner than later.

In Harris County, the public defender office is performing "triage" and has pledged to process any viable habeas writs, though they have a relatively small number because local crime labs handle most of their drug cases. In Montgomery County, prosecutors have requested testing of all 1,281 cases Salvador worked on from that county since 2005 and will notify defendants if retesting either can't be done or uncovers more errors.

How cases from the other 33 counties will be handled varies widely. My employers at the Innocence Project of Texas (IPOT) offered to perform a "triage" role for those counties similar to that being performed by the Harris County Public Defender, the commission was told, but that process has barely moved past the discussion phase and the group only recently received a list of cases involved. (N.b., IANAL and work on policy issues for IPOT; Grits has no connection to nor inside knowledge regarding the group's legal work on these cases.)

Those intermediaries are needed because most affected defendants had court-appointed lawyers who have no obligation, incentive nor resources to pursue post-conviction claims. With so many cases involved, an ad hoc approach is untenable here but the legal mechanisms for redress simply aren't set up to handle this sort of volume.

This story has mostly flown under the MSM's radar so far, but the implications loom large and I'd expect it pick up steam in the press if and when old convictions start to be overturned and folks begin leaving prison. The Forensic Science Commission will be vetting a draft of its report on the episode at their next meeting and issue a final version in the fall.

Note: This post was corrected on January 29 to attribute a quote to Dr. Nizam Peerwai that was mistakenly ascribed to Dr. Kerrigan. Grits apologizes for the error.