This article — under both its original headline and its later, retitled version (“A Drug Testing Company Favored Speed Over Reliability, Some Ex-Employees Say. The Firm Says Its Work Is ‘Accurate.’”) — is built passage by passage to leave readers with one implication: that Averhealth’s laboratory produces false positives at a rate that makes its results unreliable for the custody, probation, and treatment decisions they inform.
That is simply false: no accreditation review, government investigation, or court has ever found that Averhealth reported a single materially inaccurate result. Below are some of the many ways the article cuts corners.
In 2020, a foster care supervisor in Montcalm County, Michigan, messaged her boss with concerns about drug testing. A father who was working to reunite with his children had tested positive for methamphetamine with the lab the state had a contract with, Averhealth, and the results contradicted tests ordered by other law enforcement agencies, she wrote.
• • •
Months later, another official with Michigan’s Department of Health and Human Services wrote to colleagues about similar worries. “We are struggling to do casework with Averhealth and don’t trust them,” supervisor Sara Winter wrote. “We are making BIG decisions, including having parents leave home or removal, and that’s scary to do when you don’t trust who you’re getting services from.”
NEVER PUT TO AVERHEALTH. Both opening anecdotes — the unnamed “judge” and the Sara Winter email — were in Hines’s possession before her original October 28, 2024 comment request, and neither was ever put to Averhealth in that request, or in any of the six follow-up requests over the next fifteen months. No docket, case number, or court has ever been identified for the supposed statewide judicial finding. The only version of this episode Averhealth was shown before publication was a single sentence buried among three unrelated anecdotes. (Complaint ¶¶ 73–74, 95–96)
APPLES TO ORANGES. The testing Michigan DHHS required was oral-fluid testing at very low cutoff concentrations. An oral-fluid result cannot be compared to a urine or hair result — each specimen type has its own window of detection and its own cutoffs — and one oral-fluid test cannot be compared to another for the same reason. A positive on one specimen and a negative on another is not uncommon, and it does not mean either result is wrong. They are different snapshots in time. The ‘contradiction’ the article opens with is not evidence of error; it is how specimen types work.
The cause of the discrepancies was unknown. But that year, 2021, Averhealth’s accreditor faulted its practices and placed the lab on a six-month probation, citing, among other issues, data manipulation and failed proficiency tests, which are done to ensure test accuracy.
MISLEADING BY OMISSION. Not one of CAP’s later-detailed findings concerns whether any test result was accurate. Each concerns laboratory protocol and quality-assurance processes. CAP’s own accreditation standards reserve two more severe tiers, ‘probation with jeopardy’ and ‘probation with suspension,’ for laboratories that pose a risk of harm to patients. Averhealth was never placed in either. And CAP never found that a single Averhealth result had been incorrectly reported, nor ever required any result to be re-reported. The probation was lifted within months. (Complaint ¶¶ 75, 139–152, 231–248)
In 2023, a group of parents filed a lawsuit alleging that test results were erroneous, which prevented them from seeing their children and in some cases caused them to lose custody permanently. “Averhealth’s false positive results have had a devastating impact,” according to the complaint, which cited “substantial emotional distress, including anxiety, depression, stress and sleeplessness.” Averhealth denied the allegations in the lawsuit, which it attributed to “opportunistic plaintiffs’ attorneys hoping for a windfall settlement and plaintiffs seeking to escape the consequences of their positive tests,” according to a statement. Averhealth settled the lawsuit in September.
THE ARTICLE OMITS THE FINAL RULING. Three weeks before the settlement reported here, on August 27, 2025, a federal court:
- Entered judgment for Averhealth on all the claims of two plaintiffs
- Dismissed a third plaintiff’s claims entirely as a sanction for repeated false statements under oath
- Dismissed a fourth plaintiff’s contract claim and struck her economic damages
- Drastically constrained a fifth plaintiff’s claims after the court drew an adverse inference that she had used drugs
In fact, several plaintiffs are documented to have admitted under oath to lying about their own drug use. The words ‘summary judgment’ appear nowhere in this article. No court in that case ever found that a single Averhealth result was inaccurate. (Complaint ¶¶ 76–77, 92–93, 158–172, 249–263 — Counts Four & Nine)
ProPublica interviewed more than 50 people involved in family court or criminal proceedings who said they were tested by Averhealth and had received what they believed were inaccurate results. Some said they had tried to dispute the findings by presenting in court negative test results from other labs — one of the only ways to officially dispute a court-ordered drug test, and a strategy often recommended by attorneys. But tests from different labs are difficult to compare, and judges don’t always give credence to contradictory results.
UNVERIFIED — AND CONTRADICTED BY EVERY COMPARABLE CLAIM PROPUBLICA DID TEST. ProPublica brought only a handful of these accounts to Averhealth before publication, and every one of those was shown, on the documented record, to be false. The article says nothing about what became of them, or what that implies about the other purported sources. (Complaint ¶¶ 63, 68, 70–71, 78, 177–192, 270–286 — Counts Five & Ten)
There is next to no regulation of court-ordered drug testing by the federal government, states or the courts. At one point, McIntire said, a hospital lobbyist sought legislation in California to ban labs like McIntire’s. McIntire said he convinced lawmakers that testing doesn’t require clinical expertise. “It was determined that this is non-diagnostic in nature,” he said, meaning it isn’t used to determine medical conditions. “Therefore, it doesn’t require any special licensing — you can be a businessman.”
In 2011, McIntire sold his company to investment bankers Herzog, Jeff Herr and David Keys, who later renamed it Averhealth. The trio, who did not respond to requests for comment, had no lab experience but specialized in mergers and acquisitions. Their goal, McIntire recalls from conversations with them, was to grow and sell. In 2019, Five Arrows Capital Partners, a private equity fund and arm of Rothschild & Co., acquired a majority stake. It valued Averhealth at $150 million, according to a deposition of Herzog for the parents’ lawsuit against the company.
WHAT THE OWNERSHIP HISTORY OMITS. Averhealth’s central laboratory, established in 2014, has held CLIA certification continuously since inception and uninterrupted accreditation under CAP’s Forensic Drug Testing program since 2016 — the same accreditor whose 2021 review this article treats as its centerpiece. That accreditation was never revoked or suspended, including during the six-month probation. CAP’s program includes escalating sanctions, up to requiring a laboratory to stop testing where patient safety is at risk. None was ever imposed, which strongly suggests CAP did not regard result accuracy as the problem.
Three former lab employees — a lab tech, a chemist and a lab manager — told ProPublica that they regularly worked 12- and even 14-hour shifts and still couldn’t meet the deadlines.
Pressure to speed up delivery of results came from the top, said another lab manager, who worked at Averhealth in 2022. The manager, who asked to remain anonymous to protect their job prospects, said lab instruments were regularly in disrepair and out of service, which made meeting deadlines difficult. The manager met regularly with Delagnes, who was COO at the time, and other upper management. The manager recalls them saying, “The clients are calling and screaming, ‘What are you going to do to get these results through?’ And often my answer was, ‘The instruments are broken again. I don’t have enough people.'” The manager was unsure whether instrument maintenance affected the accuracy of test results.
ANONYMOUS, UNVERIFIED. This article ran in January 2026. The accounts in this section describe 2021 and 2022, come largely from people who asked not to be named, and are offered without a single document, test result, or client complaint to support them. The article’s own sentence gives the game away: the manager ‘was unsure whether instrument maintenance affected the accuracy of test results.’ Neither was ProPublica, but it simply printed this as fact and moved on.
Another chemist, Jennifer Picker, who worked at Averhealth from January of 2021 to July of 2022, said that when loading samples onto an instrument that confirms results, called a high performance liquid chromatograph, she was pressured to speed up to the point that she worried samples would be linked to the wrong person. “I was always extraordinarily slow and meticulous about the placement of my sample on the HPLC. And there were a lot of complaints about that I wasn’t fast enough. And I’m like, ‘No way. This is that patient’s life.'”
WHAT THE ARTICLE LEAVES OUT. Before publication, Averhealth gave ProPublica three contemporaneous employment records documenting that Picker:
- Received a written warning after messaging a colleague that she should ‘get some more degrees and experience before you try this’
- Acknowledged, in a July 2022 email recapping a conversation with her own manager, an attendance problem tied to medication that made her drowsy, and a documented need to improve competence on the very lab instruments — including the HPLC discussed here — at issue in her account
- Had her ‘future with Averhealth’ expressly conditioned on improving on those instruments. None of that documentation, supplied before this article ran, appears anywhere in it. (Complaint ¶¶ 58–63, 79–81)
Picker, the chemist, shared documentation to support her allegation that standard lab practices weren’t followed at Averhealth, including a photo of a pipette maintenance log. Chemists were to complete the logs daily to verify instrument performance. But according to Picker, this didn’t happen. She noted that the photo shows each day’s required checks were signed in the same ink and handwriting. “I watched people fill this out retroactively,” she said.
THE ARTICLE ALREADY CONTAINS THE ANSWER, BURIED BELOW. The log Picker photographed is an individual chemist’s pipette log. One person keeps it, so one handwriting and one pen is exactly what a correctly kept log looks like. Averhealth explained this to ProPublica before publication, as the article concedes two paragraphs later.
Averhealth initially told ProPublica that removing samples from the lab’s biohazard area and leaving them on a desk next to personal items and snacks was not a policy violation. Picker had a “fundamental misunderstanding of the lab security and sample integrity procedure,” it said. “There was no mishandling.” When presented with photos of a sign on a lab door saying not to remove samples from the area, a company representative said: “At any company there are going to be instances where potentially policies are not adhered to from time to time. I’m not aware that this is a significant issue for Averhealth.”
SAME DISCREDITED SOURCE. The allegations above rest entirely on the same source discredited in the preceding note. Averhealth’s contemporaneous employment records — supplied to ProPublica before publication — are absent from this account, which lets the article imply, further down, that Averhealth’s denials are ‘bare say-so.’
“I had made up my own mind about the company and separating from them,” said Penn. “What I observed, that they placed speed over accuracy, didn’t have anything to do with my relationship with Jennifer.”
Penn said he didn’t bring his concerns to the company because “I didn’t think it would make a difference.”
NOT INDEPENDENT CORROBORATION. Penn worked at Averhealth less than five months (March–August 2022), raised no concerns of any kind while employed or upon resigning, telling the company only that he had found a position closer to home, and first surfaced these complaints more than three years later. He was Jennifer Picker’s roommate throughout the relevant period. This article discloses the roommate relationship but omits all evidence that undercuts the ‘second source’ framing. (Complaint ¶ 67)
The most consequential and public criticism of Averhealth has come from Sarah Riley, a professor of pathology at Saint Louis University, who was hired as the lab’s director in September 2020. After seven weeks in that role, she abruptly resigned and filed a federal whistleblower lawsuit.
WHAT PRECEDED THE LAWSUIT. Riley filed her whistleblower suit only after Averhealth declined her demand for $1.25 million for what she characterized as wrongful termination, even though she had resigned voluntarily after seven weeks. Had the payment been made, it would have resolved her claims. The article never mentions the demand, or that it came first.
The Department of Justice investigated Riley’s claims and pursued a case, but chose to intervene on only some of the allegations.
THE ARTICLE’S CENTRAL FALSE IMPLICATION. Riley’s whistleblower complaint raised eighteen counts. The government investigated for years, reviewed millions of Averhealth documents, and declined to intervene on seventeen of the eighteen — including every count alleging an inaccurate test result. It intervened on exactly one: a contract-compliance question the U.S. Attorney’s own release described as ‘nonconforming services,’ not an accuracy issue. The government’s settlement announcement states outright that ‘the claims resolved by the settlement are allegations only and there has been no determination of liability.’ Presenting ‘the DOJ investigated Riley’s claims and pursued a case’ immediately after introducing Riley as Averhealth’s most consequential critic invites readers to conclude the opposite of what the government’s own filings say. (Complaint ¶¶ 82–83, 120–137, 210–228 — Counts Two & Seven)
According to the DOJ, the company had billed Michigan’s Department of Health and Human Services for tests it didn’t perform: Averhealth’s contract required it to confirm all positive drug tests using a second method; the government alleged it had tested samples only once.
Averhealth agreed to pay the DOJ $1.3 million to settle the case in June 2024. The company didn’t admit wrongdoing.
CONTEXT THE ARTICLE OMITS. The DOJ’s own settlement announcement states that the resolved claims ‘are allegations only and there has been no determination of liability.’ A partial intervention on one contract-performance count is not partial validation of the seventeen accuracy-related counts the government declined. (Complaint ¶¶ 128, 219)
“Did you bring these concerns to the attention of the management of Averhealth?” the Michigan mom’s lawyer asked.
She had, Riley said. “Their reaction was, ‘Thank you for bringing us your concerns. Please don’t change anything at this time. We have contractual time agreements to get the data out, and we just want to get the data out.'”
THE WRITTEN RECORD SAYS THE OPPOSITE. Riley raised proposed process improvements in a single email, and was told to proceed with her plan. That email was sent two business days before she resigned. The quotation above is her recollection of a conversation; the contemporaneous email is what management actually said.
The judge in the case, Lisa McCormick, declined to toss the contested Averhealth test results. “Dr. Riley’s testimony was speculative and did not provide the Court with any specific examples,” she wrote in her opinion. The judge later terminated the mom’s parental rights, according to court documents, citing missed drug tests and failure to benefit from services, among other issues. Termination of parental rights, which is typically irreversible, means children can be adopted by other families.
WORTH STATING PLAINLY. This is the presiding court’s own assessment of Riley’s testimony, in the judge’s own words, not a characterization supplied by Averhealth.
Averhealth denied Riley’s allegations about its lab practices and said that her testimony shows she misunderstood its science. Riley has estimated that Averhealth’s rate of inaccurate tests could be as high as 30%. In the 2024 deposition, she said the way the lab used calibrators was flawed and contributed to these errors. Calibrators ensure the measurements produced by mass spectrometers are correct.
THE ‘30%’ FIGURE, IN RILEY’S OWN SWORN TESTIMONY. Under oath, Riley admitted she had ‘no idea’ what the actual false-positive rate was for hair testing or urine testing; that her thirty-percent estimate rests on nothing more than the observation that a laboratory using one of three low calibrators then has ‘a one in three chance on any given day’ of a positive; that the true figure ‘could be 0’; and that she could not identify a single specific sample that had ever produced a false positive. In fact, all three calibrators at issue sit below the customer cutoff, so which one is used has no bearing on whether a sample is reported positive. (Complaint ¶ 82)
The judge in the parents’ case expressed doubts that Riley’s allegations about accuracy would hold up in court. Before the case was settled, the judge wrote in an order, “At trial the Court may end up sustaining an objection to her opinion that thirty percent of Averhealth’s tests are unreliable unless Plaintiffs lay sufficient foundation to explain why her thirty percent estimate (which presumably arises solely out of testing hair samples for THC) is reliable as to other forms of tests for other drugs.”
WORTH REPEATING. This is the presiding court’s own words questioning whether Riley’s estimate — drawn from hair testing for a single drug — could be extended to any other test at all.
In January of 2021, Michigan’s Department of Health and Human Services hired an independent firm, Wagner Toxicology Associates, to audit Averhealth’s lab and address judges’ concerns about Riley’s allegations. Two toxicologists, who were unaware of Riley or her complaints, spent two days in the lab “to confirm that the laboratory personnel were performing their laboratory work in accordance with their laboratory manual,” they wrote. “A relatively small number of reports was audited during the visit, and the reporting process was observed.” Wagner Toxicology issued a nine-page report stating that Averhealth’s scientists were competent and followed procedure, and that the lab’s results could be “scientifically supported and forensically defended in court.”
AN INDEPENDENT AUDIT THIS ARTICLE MENTIONS ONLY IN PASSING. Michigan’s own outside evaluator, unaffiliated with Averhealth and unaware of Riley’s complaint when it inspected the lab, reached this conclusion after direct, on-site review of the laboratory’s work.
CAP substantiated Riley’s allegations. In a letter, it stated that those allegations were: “concern regarding unacceptable quality assurance of mass spectrometry confirmatory testing”; “failure to follow procedures as written”; “concern regarding the manipulation of instrument calibrations”; and “concern regarding the review of quality control results.” CAP informed Averhealth it was on probation “based on the laboratory’s lack of continuous compliance” with “quality management” standards.
NONE OF THESE FOUR FINDINGS CONCERNS ACCURACY. Each quoted finding goes to laboratory protocol and quality-assurance process — not to whether any reported result was right or wrong. This article itself concedes as much a few paragraphs later: ‘The purpose of the investigation was to check whether the lab followed proper protocols, not to verify the accuracy of all the tests Averhealth had reported.’ CAP never found a single incorrect result, never required a result to be re-reported, and never moved Averhealth into its more serious ‘probation with jeopardy’ or ‘probation with suspension’ categories, reserved for laboratories presenting a risk of harm to patients. (Complaint ¶¶ 140, 147–149, 232, 239–241)
The purpose of the investigation was to check whether the lab followed proper protocols, not to verify the accuracy of all the tests Averhealth had reported. But the question of false positives and false negatives did come up.
Internal emails show CAP scientists discussed how problems in the lab might impact test results but didn’t reach a definite conclusion about whether the issues they found would influence accuracy. “The laboratory does not seem to have a problem with false positive results,” one toxicologist wrote, adding that its process for determining the quantity or concentration of drugs “is at times unacceptable and they do not seem to take the timely resolution of such problems seriously enough.” Another toxicologist involved in the audit seemed to disagree. “Qualitative reporting errors could exist,” that toxicologist wrote, citing “imprecision and lack of quantitative accuracy.” A qualitative reporting error is a negative test that is reported as a positive, or vice versa.
READ WHAT WAS ACTUALLY FOUND. The one toxicologist who addressed false positives directly found the laboratory ‘does not seem to have a problem’ with them. The criticism in this passage concerns quantitation — the numerical concentration reported alongside a result. Quantitative values in urine and oral fluid have no clinical significance; many accredited laboratories do not report them at all, only positive or negative. A speculative ‘could exist,’ reasoning from quantitative imprecision to a qualitative error, is not a finding that any result was wrong.
Companion piece, same series
“A Hospital Said Kaitlin Tested Positive for Codeine, But the Military Would Have Said the Test Was Negative…” — ProPublica, by Alice Hines
In a statement, Averhealth, the lab that processed Herrera’s tests, said the mismatch in results that concerned Michigan administrators “in no way calls into question the accuracy or reliability” of its testing. “Inconsistencies occurred when different types of tests were conducted (saliva or hair) or when the individual was tested days later,” the company said, noting that “different types of testing have different limitations.” The company said its test results “simply attest to whether a drug is present in a specimen and, if so, in what quantity. It is left to the courts to decide what, if any consequences, follow.”
CONSISTENT WITH AVERHEALTH’S GENERAL POSITION. Averhealth’s quoted explanation here — that cross-lab and cross-specimen-type comparisons are inherently limited, and that a result attests only to presence and quantity, not to real-world consequence — tracks the company’s consistent public position that no accreditation review, government investigation, or court has ever found it reported a materially inaccurate result.
