The Convictions Outlived the Evidence

In September 2016, thousands of letters went out to people who had been convicted of drug offenses in Massachusetts. The return address belonged to RG/2 Claims Administration LLC, at a post office box in Philadelphia. Nearby were the words “IMPORTANT LEGAL NOTICE FROM THE COMMONWEALTH OF MASSACHUSETTS.”

Inside was news about the evidence against them. A state chemist named Annie Dookhan had tested drugs in their cases. She had admitted misconduct. They could challenge their convictions.

The letters also warned that a successful challenge would return a case to active status and that a district attorney might prosecute again. Recipients were directed to their original lawyers. Anyone who could no longer find that lawyer could obtain contact information from a court clerk. A website would help locate the court.

Missing was a plain statement that defendants had a right to counsel for this process, including appointed counsel if they could not afford an attorney. Missing, too, was an adequate explanation of the powerful legal presumption the state's highest court had already created in their favor. The Spanish translation was so poor that the court later concluded it might be unintelligible to people who spoke only Spanish. Bridgeman v. District Attorney for the Suffolk District, 2017, discussion of notice.

The vendor mailed 20,916 letters. It could not locate addresses for another 1,006 defendants. Of the letters sent, 5,767 came back undelivered. A further search produced 964 additional notices. By October 24, defendants had filed only twenty motions for postconviction relief in response to the mailing. Prosecutors had filed another 175.

District attorneys offered the low response as a reason to continue handling the scandal through individual applications. Perhaps, they argued, many recipients regarded their convictions as a closed chapter and felt no urgency about reopening them.

The Supreme Judicial Court rejected that interpretation. The notice was inadequate. Even its envelope created a risk that recipients would mistake it for junk mail. Four years after the scale of Dookhan's misconduct emerged, the machinery intended to correct it was still asking thousands of people to navigate a process they had not been properly told how to use. Bridgeman, 2017, background and discussion.

Meanwhile, another laboratory scandal was exposing a more direct obstruction. Evidence showing the scope of chemist Sonja Farak's misconduct had been in government hands while defendants struggled to obtain it. Court proceedings went forward on an incomplete and misleading account of how long she had been tampering with drugs.

The two scandals involved different conduct, different laboratories, and different findings about prosecutors. Together, their court records show how an institution can discover that its evidence is unreliable and still leave the people convicted on that evidence carrying much of the burden of repair.

A signature with the weight of the state

A drug certificate answers a question that can determine the course of a criminal case: what was the substance police seized, and how much of it was there?

At the William A. Hinton State Laboratory Institute in Boston, chemists examined samples submitted by law enforcement. Their work produced the scientific findings prosecutors used in drug cases. The laboratory also stored evidence and supplied expert testimony. A result could affect the offense charged, the sentence a defendant faced, and the advice a lawyer gave about pleading guilty.

The certificate carried an implied history. Someone had received the right sample, preserved its identity, performed the appropriate tests, and accurately recorded the result. Courts and lawyers could challenge scientific evidence, but they began with a document presented as the product of an actual examination.

Dookhan broke that chain. According to the findings recounted in Commonwealth v. Scott, she sometimes grouped similar-looking samples, tested only some, and reported results for the rest. This practice became known as dry labbing. In some instances, when a sample tested negative, she contaminated it with material from a sample that had tested positive and then reported a positive result. Commonwealth v. Scott, 2014, laboratory investigation.

That conduct introduced different kinds of uncertainty. A sample reported without testing might have been correctly identified by coincidence. A sample deliberately contaminated might contain a drug only because the state's chemist put it there. The surviving certificate did not reliably distinguish between those possibilities and work performed properly.

Her misconduct extended to the supporting records. Investigators found unauthorized access to samples, falsified entries, forged initials, and false documentation about equipment checks. Those were records on which another person might ordinarily rely to reconstruct what had happened. Their falsification damaged the means of checking the result as well as the result itself.

When the scandal reached the Supreme Judicial Court, that problem became central. Dookhan could not reliably identify every case in which she had cheated. Requiring each defendant to prove precisely what she had done to that defendant's sample would make relief depend on information her conduct had made unavailable.

The scientific uncertainty also affected defendants who had pleaded guilty. A plea reflects an assessment of the evidence, the available defenses, and the consequences of losing at trial. Evidence that a state chemist was fabricating results could substantially change that assessment. A person did not have to prove that every seized substance was harmless to establish that the government had corrupted an important part of the decision. Scott, 2014, discussion of the presumption and plea materiality.

The public significance of the scandal therefore exceeded the question of how many individual laboratory results were chemically wrong. It concerned the state's ability to justify convictions obtained through a process whose reliability it could no longer demonstrate.

The warning inside the laboratory

Dookhan joined the Hinton laboratory in November 2003. Her unusually high output became part of the eventual inquiry into how the misconduct escaped detection. Speed was visible long before the full explanation for it was known.

In June 2011, an incident involving the unauthorized removal of approximately ninety samples prompted an internal investigation. Dookhan was removed from laboratory duties and assigned administrative work. The inquiry was initially limited. A formal Department of Public Health investigation followed in December. She resigned in March 2012.

The laboratory's oversight transferred to the State police in July 2012. A broader investigation followed. In August, Dookhan admitted conduct that went well beyond a single episode of mishandling. Investigators were confronting years of unreliable work and records connected to a vast number of criminal cases. Scott, 2014, factual background; Bridgeman, 2017, laboratory history, Bridgeman opinion.

There is a difference between finding a violation and determining its reach. The ninety-sample incident supplied a concrete reason to investigate Dookhan's handling of evidence. A narrow investigation could resolve that incident without resolving the reliability of earlier certificates bearing her name.

The subsequent court opinions describe failures in the laboratory's operating environment as well. In recounting the inspector general's review, the 2017 court identified weaknesses involving management, training, practices, security, and the handling of evidence. The laboratory's lack of accreditation formed part of that account. These findings concerned the conditions under which her work had been accepted and used. Bridgeman, 2017, discussion of the inspector general's investigation.

A high output should have generated questions about method and verification. In a laboratory supporting criminal prosecutions, the value of an additional result depends on whether the work behind it was performed. Measuring completion without adequately checking that work creates an obvious vulnerability: fabricated certainty can resemble exceptional productivity.

That is an inference from the documented failure, rather than a claim that every supervisor knew what Dookhan was doing. The distinction matters. The opinions establish misconduct and serious weaknesses in oversight. They do not support attributing her knowledge or intentions to everyone who worked with her.

Dookhan eventually pleaded guilty in November 2013 to twenty-seven charges, including evidence tampering, obstruction of justice, perjury, and falsely claiming an academic credential. She received a prison sentence of three to five years, followed by probation. Her prosecution established personal criminal responsibility. It did not, by itself, remove the convictions obtained in cases involving her work. Bridgeman, 2017, background.

The legal system now had two assignments. One was to prosecute the chemist. The other was to identify and repair the consequences for defendants. Completing the first could create the appearance that the scandal had reached its conclusion while the second remained largely unfinished.

Relief with another question attached

In 2014, the Supreme Judicial Court fashioned an important protection in Scott. A defendant whose qualifying drug certificate bore Dookhan's signature as a primary or confirmatory chemist would receive a conclusive presumption of egregious government misconduct.

The defendant would not have to reconstruct her individual fraud. The government could not defeat the presumption simply by arguing that she might have performed that particular analysis correctly.

This was a substantial departure from requiring proof of a specific act in a specific case. It recognized that the government's own agent had made the usual inquiry unworkable. The state could not demand a reliable map of misconduct from the people least able to obtain it. Scott, 2014, conclusive presumption.

For a defendant seeking to withdraw a guilty plea, however, the presumption answered only one part of the test. The defendant still had to show a reasonable probability that knowledge of the misconduct would have changed the decision to plead guilty. The court required an inquiry into the circumstances of the plea, including the available evidence and the choices the defendant faced.

That requirement reflected a judicial concern about the appropriate scope of relief. A compromised certificate could have mattered greatly in one case and less in another. Prosecutors might have other evidence. A defendant might have received a negotiated disposition that avoided other serious charges. The court sought to account for those differences.

The practical consequence was an enormous demand for individualized legal work. Someone had to locate the defendant, obtain the old file, identify the certificate, reconstruct the plea, assess other evidence, and prepare a motion. A favorable ruling could produce a new prosecution rather than an immediate and permanent dismissal.

The distinction between those outcomes was significant. Vacating a conviction removes the judgment. Dismissing the underlying charge with prejudice prevents the state from bringing that charge again. A defendant offered the first remedy could still face the uncertainty and expense of another case.

The court did not find that prosecutors had knowingly participated in Dookhan's original misconduct. Its later explanation expressly distinguished her conduct from prosecutorial participation. The duty to repair nevertheless belonged to the government. A state chemist worked within the prosecution's evidentiary apparatus even when a particular district attorney was unaware of the fraud. Bridgeman, 2017, remedial principles.

This distinction provides a useful measure of institutional accountability. Personal innocence of the original wrongdoing does not eliminate responsibility for correcting an institution's use of its consequences. Once prosecutors knew the certificates were compromised, their obligations included effective disclosure to the defendants whose cases depended on them.

The risk of asking for a new trial

For some defendants, reopening a case meant confronting the original charges that a plea agreement had reduced or removed. The question was whether challenging a conviction tainted by state misconduct could expose a person to a worse result.

The Supreme Judicial Court addressed that danger in 2015. Under what became known as the Bridgeman sentencing cap, a qualifying Dookhan defendant who successfully challenged a plea could not then be convicted of a more serious charge or receive a more severe sentence than under the original disposition. The protection reduced the risk of punishment for seeking relief. Bridgeman, 2017, account of the 2015 decision.

An individual case shows what was at stake before that rule existed. Angel Camacho had pleaded guilty to a reduced cocaine trafficking charge and received a sentence of six to eight years. In 2013, he successfully moved to withdraw the plea because Dookhan had been the confirmatory chemist in his case.

At the same hearing, prosecutors offered another agreement. He could plead guilty to possession with intent to distribute and receive a sentence treated as already served. According to his plea counsel, the alternative was reinstatement of the original charges and a request for the maximum penalties. The judge explained the substantial prison exposure attached to the original trafficking charge.

Camacho accepted the new plea. He also signed a waiver concerning future information about the state laboratory. Roughly seventeen months later, the Supreme Judicial Court established the Bridgeman protection. Commonwealth v. Camacho, 2019, background.

Camacho subsequently tried to withdraw his second plea. In 2019, the court held that the sentencing cap could apply retroactively, but only where a defendant had actually received a harsher sentence or been convicted of a more serious charge. Camacho had received a lesser conviction and a shorter sentence. His second plea remained in place.

That outcome should not be retold as a judicial finding that his second conviction was unlawful. The court affirmed the denial of relief. It nevertheless documented the choice he had confronted after establishing that his first plea rested on compromised government work: accept an offer resolving the case immediately, or face a renewed prosecution with serious exposure.

The later protection did not erase the earlier negotiation. Nor did it automatically reopen every second plea entered before the cap existed. The court drew a line based on the actual result, rather than the pressure of negotiating under the possibility of a worse one. Camacho, 2019, holding and analysis.

This was one reason the cleanup could not be measured simply by counting favorable appellate decisions. Each ruling operated within a changing population of cases. Some defendants were still serving sentences. Others had finished them. Some had already accepted new dispositions. A remedy arriving later could protect future applicants while leaving earlier decisions intact.

Twenty motions after twenty thousand letters

By 2016, the litigation had reached the problem of reaching people. Prosecutors and defense organizations had worked to assemble lists from records that did not readily yield a single, agreed count of affected defendants and cases.

The court acknowledged the work that went into identifying them. The parties nevertheless disputed whether the lists included people already granted relief and whether errors omitted others. A person could have more than one case. A case could involve multiple charges. Different procedural categories carried different rights.

Those distinctions make the numbers easy to misuse. The roughly 21,000 mailed notices were letters, rather than a clean total of unique defendants or convictions. The court's concern was clear even without resolving every counting dispute: close to 20,000 qualifying defendants might still be entitled to seek postconviction relief. Bridgeman, 2017, identification and notice.

The Committee for Public Counsel Services, the state's public defense agency, resisted a notice that assumed a workable supply of lawyers without addressing whether that supply existed. It argued that the volume of potential applications exceeded the resources available for individual litigation. The district attorneys favored continuing the process the court had established.

The dispute was unresolved when prosecutors proceeded with the mailing. The defense petitioners objected to its content and translation. A single justice invited delay, but mailing had already begun. An emergency request to stop further distribution failed, although the court required records of communications arising from the notice.

The result was a revealing administrative record. Thousands of notices never reached an address. Others provided incomplete explanations. Recipients were told about the risk of renewed prosecution without being adequately told about the legal findings that could help them or their entitlement to counsel.

The court's account of the Spanish translation was especially striking. The word for vacating a conviction had been rendered as physically vacating premises. Another phrase described a criminal clerk in a way that made the clerk himself a violent felon. These were consequential failures in a letter meant to explain unfamiliar legal rights. Bridgeman, 2017, notice translation and judicial assessment.

The district attorneys argued that a low response could reflect defendants' lack of interest. The court refused to use that response as evidence that the process was adequate. A recipient's silence could arise from failed delivery, misunderstanding, difficulty finding counsel, or concern about reopening a case. The record could not establish informed acceptance simply because few people filed motions.

The distinction has consequences well beyond criminal courts. Whenever an institution requires people to apply for correction, it controls part of the process that produces the apparent demand for correction. A confusing notice and an inaccessible remedy can generate a low response. Citing that response as evidence of little harm completes a self-protecting circle.

In Massachusetts, the Supreme Judicial Court broke that circle. It held the notification effort against what defendants needed to know, rather than the number of envelopes the state had mailed.

A remedy had to be possible to use

In January 2017, the court rejected the district attorneys' request to continue without substantial change. It also declined to order the automatic dismissal of every qualifying Dookhan case. Instead, it reorganized the work.

District attorneys would have to review their cases first. Within ninety days, they were to identify convictions they would move to vacate and dismiss with prejudice. That group would include cases they did not wish to pursue again and cases they could not successfully prosecute again. Cases they intended to retain would be separately identified. Bridgeman, 2017, three-phase protocol.

That changed the order of responsibility. The government would assess whether there was any reason or basis to preserve a prosecution before requiring a defendant to find a lawyer and litigate it. An individual application would no longer be the necessary first step for every conviction the state was prepared to abandon.

The remaining defendants would receive court-approved notice. The court's supervision extended to the envelope, delivery method, and content. Additional public notice could be ordered to improve the chance of reaching people who missed the mail. Indigent defendants seeking legal help would be assigned counsel through the public defense system.

The court also confronted the possibility that available attorneys would be insufficient. Low compensation, limits on billable hours, funding, and the scale of the scandal could prevent the assignment of counsel. If the system could not supply the representation to which defendants were entitled, another remedy could become necessary, including dismissal without prejudice.

The opinion did not treat a right written on paper as the end of the inquiry. A process that would take too long or overwhelm available resources could fail as a remedy even if it looked fair in an individual case. The court required practical attention to how thousands of cases would actually move. Bridgeman, 2017, counsel and workable remedies.

Most of the defendants under discussion had already served their sentences, the court observed. That fact reduced some immediate custody concerns, but it left the consequences of convictions. Criminal records can continue to affect employment, housing, immigration, and later encounters with the justice system. The court treated those continuing consequences as reasons relief still mattered.

The eventual mass dismissals were a major change in scale. More than 21,000 drug cases were dismissed in 2017 following the Dookhan litigation. That outcome came years after the misconduct emerged. Contemporaneous Axios account of the dismissals.

Those dismissals should not be described as individual scientific demonstrations that every defendant possessed no illegal substance. They reflected the state's inability or decision not to sustain prosecutions under the circumstances created by its own misconduct. The constitutional and practical failure was serious enough without converting the remedy into a claim the evidence could not establish.

While this process unfolded, the Farak cases were revealing what could happen when the government also obstructed access to the facts needed for correction.

A second laboratory, another damaged record

Sonja Farak transferred from the Hinton laboratory to the state laboratory at Amherst in 2004. She remained there until it closed in January 2013. Her misconduct differed from Dookhan's. Farak stole and consumed drugs, including laboratory standards and police-submitted evidence, and altered evidence and records to hide what she had taken.

The setting offered few effective barriers. The Supreme Judicial Court's 2018 account described broad employee access to standards, samples, and the inventory computer. Department of Public Health officials had visited only once or twice between 2006 and July 2012. Annual performance reviews stopped after a change in supervision in 2008. Coworkers generally regarded Farak as an excellent chemist. Committee for Public Counsel Services v. Attorney General, 2018, Amherst laboratory history.

Drug standards were part of the testing process. A chemist compared an unknown substance with a known one to help determine its identity. Farak began taking methamphetamine from a laboratory standard in late 2004 or early 2005. Eventually, she depleted much of the supply and added water to conceal the loss.

By 2009, the misconduct had broadened. The record included evidence of theft from police-submitted samples, along with use of other laboratory standards. In later years, Farak manipulated the computer inventory to obtain samples containing drugs she wanted. She changed recorded weights and used substitute materials to replace stolen substances.

Those acts could affect work assigned to other chemists. A certificate bearing another employee's name did not necessarily mean the sample had remained beyond her reach. Farak could alter evidence before testing, take it afterward, or replace material needed for a later examination. The court eventually made that point central to the scope of relief. CPCS v. Attorney General, 2018, evidence tampering and remedy.

The details were extraordinary, but their institutional meaning was straightforward. Access controls, inventory records, packaging, and retained samples are protections for everyone who later relies on a laboratory result. When one person can change several of those elements, apparently corroborating records may reflect the same concealed misconduct.

In January 2013, an evidence officer noticed missing samples assigned to Farak. A search of her work area revealed opened packaging and materials associated with substituting fake substances. The State police were notified, the laboratory closed, and a criminal investigation began. Farak was arrested on January 19.

The discovery came only months after the Dookhan scandal had become public. Massachusetts already knew that a chemist's misconduct could destabilize thousands of cases. It already had reason to separate the prosecution of an employee from the larger investigation into the reliability of convictions.

The Farak investigation would test whether the state had learned that distinction. Its initial theory sharply limited the period of misconduct. Evidence in government custody would soon undermine that theory. Getting that evidence to the people who needed it became another scandal.

What the government found in the car

Investigators searching Farak's car found drugs, evidence packaging, and envelopes containing papers. The volume of material led them initially to describe part of the collection as assorted laboratory paperwork. A closer examination revealed mental health worksheets and related material that bore directly on the timing and extent of her drug use at work.

The significance was immediate. A document incriminating Farak could help a defendant whose conviction depended on her work. Evidence that she had been using drugs and taking samples earlier than investigators first assumed could bring additional cases within the period requiring scrutiny.

On February 14, 2013, State police Sergeant Joseph Ballou emailed documents to prosecutor Anne Kaczmarek, criminal bureau chief John Verner, and another investigator. The subject was “FARAK Admissions.” Kaczmarek reviewed the material, saved a copy on her computer, and placed printed copies in her trial materials. The later disciplinary findings described the documents as plainly important to both Farak's prosecution and the defendants affected by her work. In the Matter of Foster, 2023, discovery of the worksheets.

The attorney general's office had adopted a disclosure policy under which it would send discoverable information to district attorneys, who could then provide it to affected defendants. Verner understood that the worksheets needed to be disclosed. When reviewing Kaczmarek's prosecution memorandum, he noted that they had not yet been turned over.

But the discovery packages sent to district attorneys omitted the worksheets. Farak's own defense attorney received them. Lawyers for people prosecuted on the basis of Farak's analyses did not receive them through those disclosures.

Questions about privilege formed part of the internal discussion. Mental health information ordinarily deserves careful handling, and the possibility of protected material can require a judge's review. That concern did not authorize assurances that everything relevant had been disclosed when it had not. Nor did it eliminate the significance of evidence suggesting a longer history of tampering.

The 2023 disciplinary opinion described a revealing detail in Kaczmarek's prosecution memorandum: it discussed withholding the worksheets from the grand jury out of caution about privilege while also noting that case law suggested the paperwork was not privileged. The office had enough information to recognize a disclosure question requiring resolution. Foster, 2023, prosecution memorandum and disclosures.

This was a failure of information moving through government. The material had been seized, recognized, discussed, copied, and retained. Its absence from the defendants' files was consequential because the state's preliminary account of the scandal remained much narrower than the evidence justified.

The people seeking relief could not simply supply what was missing. Their lawyers had to identify the existence and significance of documents held by the institution whose account they were challenging. The government knew where the evidence was. The defense had to litigate for the opportunity to see it.

Assurances without an examination

Lawyers for affected defendants sought subpoenas, access to evidence, and information about the scope of Farak's misconduct. Kris Foster, then relatively new to the attorney general's office, was assigned to handle responses to those requests.

The later disciplinary decision described problems in that assignment. Foster had no prior experience responding to subpoenas. Supervisors supplied examples and advice, but the work required a concrete understanding of what investigators possessed and what had already been disclosed. She did not adequately establish those facts before making representations to the court. Foster, 2023, responses to defendants' discovery requests.

In September 2013, Judge C. Jeffrey Kinder directed a review of Ballou's file and required any material claimed to be privileged to be presented for private judicial inspection. Foster subsequently sent a letter saying that, after review, every document in Ballou's possession had been disclosed and there was nothing to produce for the judge.

She had not personally examined the file. Other lawyers had supplied assurances about disclosure. Her letter gave those assurances the appearance of a completed factual review. It became part of the information on which the court acted.

The significance of her conduct requires attention to two separate proceedings. In the litigation that produced the 2018 decision, Judge Richard Carey had found that Foster and Kaczmarek's withholding and misleading conduct constituted a fraud on the court. The Supreme Judicial Court relied on the serious governmental misconduct established in that record when fashioning relief for defendants. CPCS v. Attorney General, 2018, findings following the Carey hearing.

The later disciplinary case examined individual professional violations on its own record. In 2023, the Supreme Judicial Court declined to find that Foster had violated the rule prohibiting dishonesty, fraud, deceit, or misrepresentation as charged. It accepted the finding that she lacked the required knowledge and was not willfully blind. It nevertheless found serious incompetence and reckless misrepresentations, and ordered a suspension of one year and one day. Foster, 2023, analysis of Foster's responsibility.

Both parts belong in the account. Quoting the earlier findings while omitting the later distinction would overstate what the disciplinary court ultimately concluded about her state of mind. Treating the later ruling as vindication would erase the violations and suspension it upheld.

The institutional damage did not depend on every lawyer sharing the same intent. An assurance to a judge can obstruct correction when the person making it has failed to verify its truth. The 2023 court connected Foster's reckless representations to Kinder's decisions denying relief because the record did not show earlier misconduct.

The state had evidence bearing on that very question. Its lawyers' account of disclosure helped prevent the evidence from doing its work.

The inspection that changed the case

Defense attorney Luke Ryan repeatedly sought access to the material seized from Farak. The record describes requests denied, efforts redirected into motions, and continued disputes over relevance and the ongoing criminal investigation.

One episode complicates any account that treats every prosecutor as a participant in concealment. Hampden County prosecutor Francis Flannery sought to arrange an inspection by defense lawyers. Kaczmarek opposed access, invoking the ongoing case against Farak and describing the proposed review as a fishing expedition. The proposed inspection did not occur at that stage. Foster, 2023, defendants' discovery requests.

Ryan continued. After further litigation, he obtained an order permitting access. On October 30, 2014, he examined the evidence and found previously undisclosed documents, including the mental health material. He then alerted the attorney general's office that the documents contradicted the limited period on which the earlier court findings had rested.

In November 2014, the office informed district attorneys that it was sending 289 pages of documentary evidence that had not previously been produced. The records had been in government hands since the investigation that began in January 2013. The defense inspection occurred more than a year after the 2013 proceedings in which access and disclosure had been contested. CPCS v. Attorney General, 2018, discovery of the records.

This chronology shows the burden in unusually concrete form. Ryan needed enough reason to keep seeking material he had not been allowed to inspect. He needed a procedural route, a favorable order, an actual inspection, and an understanding of why what he found mattered to other cases.

The task belonged to an adversarial lawyer, but the information was part of a public obligation. The Commonwealth's duty to investigate and disclose did not depend on whether a particular defendant could fund or sustain repeated attempts to obtain it.

Earlier proceedings had treated the misconduct as beginning in the summer of 2012. Defendants whose pleas predated that period lost requests for relief. Other cases were affected by those rulings. A limited factual account thus became the basis for decisions reaching beyond the first hearing.

The later Supreme Judicial Court opinion reconstructed that sequence. The consequence of incomplete disclosure was not confined to a missing document in one file. The missing information shaped the account of the scandal that courts used to decide who might deserve a remedy. CPCS v. Attorney General, 2018, Kinder findings and subsequent appeals.

That mechanism is particularly dangerous in mass misconduct cases. Once a court establishes a cutoff date, it can determine the treatment of many people who were never parties to the first factual dispute. An inaccurate boundary can be repeated through the system with the authority of a judicial finding.

Correcting it requires more than adding a document to a file. Someone must show why the earlier understanding was wrong and obtain reconsideration for the people whose cases were decided under it.

An investigation of the investigation

The Farak scandal then generated another sequence of investigations. After appellate proceedings in 2015, the attorney general's office undertook broader work on the timing and scope of the misconduct. Retired judges were appointed to investigate issues concerning the government's response. Separate grand jury proceedings examined Farak's conduct and conditions inside the laboratory.

Farak testified about extensive drug use, theft, tampering, and the manufacture of crack cocaine at her workbench. Investigators collected thousands of emails and other records. The account of a short, limited episode could no longer contain the evidence. CPCS v. Attorney General, 2018, Velis and Caldwell investigations.

The inquiries did not all reach the same conclusion about prosecutors. A March 2016 report by retired judges Peter Velis and Thomas Merrigan stated that they found no evidence of prosecutorial misconduct or obstruction by the assistant attorneys general and State police officers in the Farak matter.

That conclusion did not end the judicial examination. In December 2016, Judge Carey held six days of evidentiary hearings. Witnesses included prosecutors, supervisors, investigators, former laboratory employees, and experts. His June 2017 decision reached findings sharply at odds with the earlier assessment of the government's conduct.

Carey found misconduct of systemic magnitude, serious withholding of evidence, and misleading representations to the court. He also found no adequate earlier government investigation before the broader report produced in 2016. His conclusions became central to the Supreme Judicial Court's later decision. CPCS v. Attorney General, 2018, investigative reports and Carey findings.

An investigation's existence is therefore a poor substitute for examining its methods and conclusions. Different proceedings may have different records, legal questions, and opportunities to test testimony. A report finding no misconduct can coexist with later judicial findings establishing it. The responsible account identifies the conflict and explains which decision followed.

There is also a limit to what that conflict proves. The later findings do not, by themselves, establish that the retired judges deliberately protected prosecutors. Their report's conclusion can be assessed against the subsequent record without assigning an unproven motive to its authors.

The documented sequence is damaging enough. The government possessed relevant evidence in 2013. A defense lawyer obtained crucial access in 2014. Broader investigative work followed in 2015 and 2016. The judicial findings underlying a sweeping remedy arrived in 2017 and 2018.

For defendants, those dates represented successive stages of waiting for an institution to establish the extent of its own failure. For the institution, they represented hearings, reports, appeals, and changes in legal position. An accountability story has to keep both calendars visible. A completed procedural step could be progress while a tainted conviction remained operative.

The name on the certificate was too narrow a boundary

By the time the Supreme Judicial Court considered the Farak remedy in 2018, district attorneys had agreed to vacate more than 8,000 convictions and dismiss the corresponding charges with prejudice. Remaining disputes concerned the reach of relief, including cases in which another chemist had signed the certificate.

Prosecutors argued against extending dismissal to all work performed at the Amherst laboratory during Farak's employment. They disputed whether the record justified treating other chemists' analyses as compromised. The attorney general proposed a broader period than the district attorneys, tied to the summer of 2012, when Farak acknowledged tampering with other chemists' samples. CPCS v. Attorney General, 2018, parties' positions.

The court rejected an approach confined to Farak's signature. The evidence showed that she had access to other chemists' samples, altered inventory information, and replaced substances. Even tampering after an initial analysis mattered because it could prevent a reliable retest of the original evidence.

That last point went to a defendant's ability to challenge the state's case. Retained evidence is part of the opportunity to verify a laboratory's conclusion. If the material has been stolen, diluted, or replaced, the original certificate cannot be meaningfully tested in the same way. The person whose liberty depended on that result should not have to absorb the uncertainty created by the state's loss of control.

The court also declined to dismiss every Amherst case throughout Farak's entire tenure. It found the evidence supported a more specific boundary. The expanded class included the Farak-signed cases, methamphetamine cases tied to the laboratory during her tenure, and convictions based on substances tested there from January 1, 2009, through its January 18, 2013 closure, regardless of the signing chemist. CPCS v. Attorney General, 2018, holding and definition of affected defendants.

The broader remedy rested on the combined effect of laboratory tampering and the government's response. The court found the misconduct sufficiently intentional and egregious to require dismissal with prejudice. A new round of ordinary retrials would not adequately address what had happened.

The opinion made another distinction worth preserving. It found no fault with the district attorneys' actions in transmitting the evidence they received and identifying and notifying affected defendants in the Farak matter. Its criticism of the attorney general's office should not be indiscriminately transferred to every prosecutor in Massachusetts.

The result was sweeping, but it was bounded by evidence. The court expanded relief where Farak's access and conduct undermined other employees' work, while declining the most comprehensive dismissal requested. That is a stronger account of institutional failure than a claim that every sample from every year had been proved false.

It also explained why an institution cannot define the damaged population solely by the most visible identifier. A signature was useful for finding cases. The misconduct had traveled further than the signature showed.

Responsibility, ten years later

The disciplinary decision issued on August 31, 2023, almost a decade after the contested disclosures and court representations. It assigned different consequences to three attorneys whose roles and culpability differed.

For Kaczmarek, the court adopted disbarment. It found that the lead prosecutor bore the greatest responsibility for the failure to disclose. Her conduct included knowingly obstructing access to evidence and making materially false and misleading statements. The court also considered her lack of candor and remorse during the disciplinary process. It remanded for entry of a judgment implementing the decision. Foster, 2023, Kaczmarek analysis and conclusion.

For Foster, the sanction was a suspension of one year and one day. The court distinguished recklessness and incompetence from the particular dishonesty violation that bar counsel had sought to prove. It considered her reliance on colleagues and a supervisor, while emphasizing her own responsibility for factual representations she signed and submitted.

For Verner, the court selected a public reprimand rather than the three-month suspension recommended by the Board of Bar Overseers. His failure concerned supervision and follow-up. The court found that he had taken steps to require disclosure and had reasonably relied in good faith on Kaczmarek's assurances, although he should have checked that the work was done.

Those differences demonstrate why a collective label such as “the prosecutors” can conceal important facts. One attorney intentionally withheld and misrepresented information. Another recklessly conveyed assurances she had not adequately verified. A supervisor failed to follow through on disclosure obligations while being misled about compliance. The resulting harm involved a shared process, but the disciplinary court assessed each person's conduct separately. Foster, 2023, individual sanctions and mitigating findings.

The decision was genuine accountability. It should be reported as such. A claim that nobody faced consequences would be false. The same record shows how long the formal assignment of professional responsibility took after the evidence was withheld.

It also reveals an institutional weakness that punishment alone cannot resolve. The disclosure process depended on lawyers knowing what investigators held, accurately describing what had been sent, and checking that instructions were carried out. In the Farak matter, confident statements traveled further than verified information.

That failure calls for a practical question: can an office demonstrate disclosure through an auditable record of the materials reviewed, the decisions made, and the recipients supplied? An instruction to disclose is one event. Actual delivery is another. Supervisory confidence cannot establish that the second followed the first.

This article reports the sanctions selected in the 2023 decision. It does not assert the attorneys' present licensing status or imply that a suspension ordered then is still in effect today. Those are separate questions requiring current records. The historical finding remains: a decade after the disputed disclosures, the state's highest court ordered professional consequences for the conduct that helped prolong the scandal.

What a completed correction would have to show

The Massachusetts opinions expose several points at which a failure could continue even after the central misconduct was publicly acknowledged.

A laboratory could produce certificates faster than its oversight could verify the work. An investigation could define a period too narrowly. A disclosure system could transmit assurances while leaving crucial records behind. A notice could reach thousands of mailboxes without adequately explaining the rights it purported to announce. A remedy could require more lawyers than the state had made available.

Each problem had a different responsible actor and a different possible correction. Treating all of them as the personal misconduct of two chemists would leave much of the documented story unexplained. The government's handling of the aftermath determined how long the original damage continued to affect defendants.

The appellate courts eventually supplied important protections: presumptions that relieved defendants of impossible proof, limits on renewed punishment, court-supervised notices, and broad dismissals. Defense lawyers and public defense organizations helped force the underlying facts into view and challenged remedies that could not function at the necessary scale. Prosecutors also participated in identifying cases and dismissing convictions, and the opinions credited work that deserved credit. Bridgeman, 2017, CPCS v. Attorney General, 2018.

The history leaves a concrete standard for future failures. An institution should be able to account for the affected population, the basis for including or excluding cases, the notice actually delivered, access to representation, and the final treatment of each case. Those measures would show whether an announced remedy reached the people it was supposed to help.

They would still leave questions about the consequences that outlast a judgment. Removing a conviction cannot restore time already spent in custody. It does not establish, by itself, that every downstream record has been corrected or every financial loss repaired. Those outcomes require their own evidence. This reconstruction does not claim to have audited them.

The September 2016 mailing remains a useful place to measure the difference between acknowledging damage and repairing it. Massachusetts had identified a vast problem. Courts had recognized rights. Letters had been sent. Yet the notice failed to explain essential protections, thousands of envelopes were returned, and the small number of applications became an argument for leaving the process largely as it was.

The Supreme Judicial Court refused to accept that argument. A remedy had to give people a meaningful opportunity to use it. Its existence could not be established by the institution's activity alone.

In the Farak cases, obtaining that opportunity required exposing information the government already possessed. The people whose convictions were in question had to fight for the records that showed how extensive the state's misconduct had been.

The certificates had carried the authority of Massachusetts into their cases. Repair required Massachusetts to assume responsibility for what that authority had done.

Reporting note: This article reconstructs events from published judicial opinions, with a contemporaneous news account supporting the description of the 2017 mass dismissals. It includes analysis of the documented procedures. No new interviews or requests for comment were conducted for this article. Statements about individual conduct are attributed to the relevant proceedings. The companion Record identifies sources and counting limits.

Explore the Record: Massachusetts Drug Laboratories: The Record