Detroit’s rape-kit backlog was never simply a story about evidence waiting to be tested. It was about the decisions that determined which evidence was worth pursuing—and which survivors were believed. The thousands of kits left on shelves became a physical record of a systemic failure whose consequences extended far beyond the evidence room.

Read time: 15 minutes

In August 2009, representatives from the Detroit Police Department, Michigan State Police, and Wayne County Prosecutor’s Office toured a remote police property-storage facility to discuss the department’s growing volume of evidence. During the tour, an assistant prosecutor noticed rows of boxes on shelving units and asked what they contained. Police personnel identified them as sexual-assault kits—but could not say how many were there or how many had ever been tested. The boxes contained sexual-assault kits, many of which had never been submitted for forensic DNA testing. Some had been sitting in police custody since the 1980s.

A later census, conducted as part of a National Institute of Justice-sponsored research project, identified 11,219 sexual-assault kits in police custody as of November 1, 2009. Of those, 2,512 carried laboratory identification numbers, indicating that they had at least been submitted to a laboratory, though researchers could not determine how many had actually undergone DNA testing. Wayne County officials have more commonly described the backlog as 11,341 kits, the figure associated with the Detroit Rape Kit Project. The figures are not directly interchangeable: the research team applied formal census criteria to the property inventory, and 2,512 of the 11,219 kits it counted already carried laboratory identification numbers, although researchers could not determine how many had actually undergone DNA testing.

The precise count matters less than what the boxes represented. A sexual-assault kit exists because a person reported an assault and, in most cases, underwent a medical forensic examination during which evidence was collected from the body and clothing, injuries were documented, and materials were preserved—a process that can take several hours. Implicit in that process is the expectation that the evidence will be used. In Detroit, thousands of times, it was not. Thousands had been collected, labeled, and left in police custody without being submitted for forensic testing. The facility was therefore not simply a repository of forgotten boxes; it was a physical record of thousands of discrete decisions not to proceed.

What Does It Mean for a Rape Kit to Be “Backlogged”?

The vocabulary of the “backlog” can obscure what occurred. A kit may be considered backlogged under more than one definition. A crime laboratory may hold kits awaiting testing because its caseload exceeds its capacity—a laboratory backlog, which typically signals insufficient staffing, funding, equipment, or throughput. A second category, unsubmitted kits, never reached a laboratory at all, raising a different question: not why testing was delayed, but why the evidence was never sent.

Detroit’s history included both kinds of failure: thousands of kits remained in police custody without being submitted, while the institutions responsible for testing and investigating sexual-assault evidence also faced serious resource constraints. Researchers examining the backlog documented chronic understaffing and resource depletion, limited laboratory capacity, frequent turnover in police leadership, inconsistent policies, strained interagency relationships, and inadequate training. They also identified victim-blaming attitudes among police personnel as part of the institutional environment in which kits went untested. No individual failed to mail eleven thousand boxes, and no single equipment failure produced decades of accumulated evidence. The backlog was systemic—which is what makes Detroit more instructive than an account of one department’s evidence room.

A Kit Can Be Collected Without a Case Being Believed

DNA evidence is often regarded as the most objective element of a criminal investigation. A biological sample registers no assumptions about whether a complainant had been drinking, what they were wearing, whether they knew the accused, or whether they behaved as an investigator expects a victim to behave. But whether a sample is tested is not determined by the sample; it is determined by people. That was among the most consequential findings of the Detroit research.

The National Institute of Justice summarized several contributing risk factors, including victim-blaming beliefs among police personnel—assumptions that a survivor was not behaving like a “real” victim, or that an allegation amounted to a “he-said, she-said” dispute. Detroit also lacked a consistent written policy governing when sexual-assault kits should be submitted for testing. The combination is significant: where investigators exercise discretion over testing and hold preconceptions about credible sexual assault, a case that does not fit those expectations can result in a kit that stays in storage. The technology may be objective, but access to it is not. The failure was not only forensic; it was cultural.

The Myth of the “Perfect” Victim

Sexual-assault cases have long been shaped by credibility assessments that would be considered troubling in most other categories of violent crime. Investigators may encounter delayed reporting, continued contact between the complainant and the accused, inconsistencies in an account, a calm rather than visibly distressed demeanor, the involvement of alcohol or drugs, or a prior relationship between the parties—partner, acquaintance, or date rather than stranger. None of these circumstances determines whether an assault occurred, yet expectations about how a genuine survivor “should” behave can influence whether a report is taken seriously.

Detroit’s backlog offers an unusually concrete illustration of what happens when such judgments enter the evidentiary process: an examination could be completed and evidence preserved, and an investigator could nonetheless conclude that testing it was not worthwhile. That conclusion is consequential because forensic testing can reveal what an investigator could not have known when the judgment was made—identifying a previously unknown suspect, contradicting a suspect’s account, corroborating part of a complainant’s report, or linking one assault to another. The last of these became especially important once Detroit began testing its older kits.

What Happened When They Finally Tested the Evidence?

A longstanding argument against testing every kit held that testing would sometimes yield little investigative value—because the suspect was already known, because the dispute appeared to concern consent rather than identity, or because prosecution seemed unlikely. Testing Detroit’s abandoned kits allowed researchers to examine those assumptions directly. In one study, researchers tested a sample of 900 previously unsubmitted Detroit kits to determine how often they yielded DNA profiles eligible for entry into CODIS, the Combined DNA Index System, and how often those profiles matched existing records, including other sexual-assault cases. In that 900-kit sample, testing produced 259 CODIS hits; 69 involved DNA matches to another sexual-assault case.

The broader testing effort demonstrated why leaving the kits untouched carried consequences beyond any single case. When Wayne County announced in November 2022 that all 11,341 kits in the original backlog had been processed, prosecutors reported that more than 4,800 associated cases had been investigated and closed, resulting in 239 convictions and the identification of 841 suspected serial offenders across 40 states. The totals continued to grow after the testing milestone: as of August 2026, End the Backlog reports 3,169 DNA matches, 859 suspected serial offenders, and 272 convictions, with DNA from the tested kits linked to crimes in 41 states and Washington, D.C. These figures reframe the significance of the facility. Many of the boxes were not isolated evidence from isolated crimes but potential links between cases—links the justice system could not use while they remained unexamined.

The Serial-Offender Problem

Sexual assault is frequently discussed as though each report stands alone. DNA databases reveal a different picture. When the same offender’s biological evidence appears in more than one kit, testing can convert reports that seemed unrelated into evidence of repeated offending; conversely, when those kits are never submitted, the connection remains invisible, and each case continues to be treated in isolation.

The scale of that risk is not merely theoretical. Testing in Detroit repeatedly linked DNA from previously unsubmitted kits to other sexual-assault cases, demonstrating that reports treated as isolated events could involve the same suspected offender. A kit left unsubmitted therefore may bear on more than the investigation in which it was collected: a missed database match can leave repeated offending invisible to investigators. Understood this way, the backlog not only withheld investigative opportunities from the survivors already represented in the boxes but, in some cases, also removed the opportunity to identify patterns of repeated offending earlier.

The point should not be overstated. A database match does not establish guilt, and testing a kit guarantees neither an arrest nor a conviction. Some cases cannot be prosecuted for reasons unrelated to forensic evidence; some profiles do not match anyone in CODIS; some evidence cannot resolve a dispute over consent. These limitations are real, but they do not render the evidence valueless. They indicate that testing must be followed by investigation.

A Rape Kit Is Evidence, Not a Conviction

An important qualification follows. The presence of DNA in a sexual-assault kit does not, by itself, establish that an assault occurred. DNA can demonstrate biological contact or identify the likely source of biological material; whether that information is probative depends on the circumstances. Where a suspect denies ever having met the complainant, a match may be highly significant; where both parties agree that sexual contact occurred but dispute consent, the same match answers a much narrower question.

A CODIS hit, likewise, is an investigative lead rather than a verdict. The case must still be investigated, the evidence evaluated in context, and the sufficiency of charges assessed by prosecutors, with the defendant entitled to the presumption of innocence. The argument for testing does not depend on treating DNA as infallible. It depends on a simpler proposition: that evidence cannot inform an investigation that no one examines.

The Resource Problem Was Real

It would be inaccurate to attribute the backlog solely to indifference; the institutional history is more complicated. Researchers documented chronic understaffing, resource depletion, laboratory capacity issues, and extraordinary leadership instability. Detroit had nine police chiefs from roughly 1991 through 2009, with an average tenure of just over two years; three additional chiefs served during the subsequent 30-month action-research project. Researchers concluded that the turnover made sustained supervision and reform more difficult.

These conditions matter because sexual-assault investigations are labor-intensive. Testing requires laboratory personnel, equipment, reagents, facilities, quality control, database administration, and funding; and once older kits produce database hits, investigators must reopen files, locate surviving evidence, identify witnesses, find complainants who may have moved years earlier, interview suspects, and assess whether prosecution remains legally possible. Clearing a backlog is thus not a matter of sending boxes to a laboratory: testing generates investigative work, and large-scale testing can expose years of it at once.

Yet resource scarcity cannot fully account for Detroit, because the research found that institutional attitudes shaped which cases were judged worthy of the limited resources available. That is the point at which scarcity becomes a question of justice. Every system operates under constraints; the salient question is how those constraints are distributed, and whose cases are permitted to settle at the bottom of the queue.

When a Backlog Becomes a Measure of Whose Cases Matter

A box on a shelf appears neutral. It is not. Each kit passed through a sequence of discretionary decisions—collection, logging, the choice whether to submit it, supervisory review or its absence, the policies governing submission or their absence, staffing allocations, and budget decisions—culminating in a collective judgment about whether the growing accumulation constituted a problem at all.

That final judgment may be the most revealing. Researchers studying Detroit specifically examined why police personnel did not regard thousands of untested kits as a problem before the backlog was publicly recognized as a crisis. The system did not begin to fail when the boxes were discovered; the boxes were evidence that it had been failing for years. The discovery only made the failure visible.

The Survivor at the Other End of the Box

Aggregate figures carry an inherent difficulty: at a certain magnitude, they cease to be comprehensible. Eleven thousand kits, thousands of investigations, hundreds of suspected serial offenders, and hundreds of convictions are necessary to understand the scale of the institutional failure, but that same scale can obscure the individuals most directly affected.

Each genuine kit began with a person who went to a hospital or examination site, described what had happened, and permitted evidence to be collected—who may have been photographed, swabbed, examined, questioned, and asked to surrender clothing—and who then went home. Some waited for a call; some contacted the police themselves; some stopped expecting anything; many built lives in the intervening years. When investigators reopened Detroit’s older cases, they confronted a further difficulty: how to approach someone about an assault the justice system had neglected years earlier.

Testing Old Kits Can Reopen Old Trauma

Clearing a backlog is often treated as an unambiguous good. Testing the evidence is indeed necessary; contacting survivors is more delicate. A person who reported an assault fifteen or twenty years earlier may have long assumed the case had gone nowhere, may have deliberately set it aside, and may have moved, married, changed their name, sought counseling, or never disclosed the assault to anyone in their current life. A call reporting that the kit has been tested, that a database match exists, and asking whether they wish to participate in a prosecution does not admit of a single correct approach.

For this reason, Detroit’s action-research project did more than test evidence: researchers and practitioners developed and evaluated a victim-notification protocol to structure case processing and survivor notification. The work underscores a point easily lost in discussions of forensic reform—that correcting an institutional failure can itself cause harm if pursued without regard for the person who survived the original crime. A victim-centered response requires more than locating an individual and conveying information; it requires acknowledging that the system is arriving late, sometimes by decades.

Detroit Was Not an Isolated Problem

The warehouse rendered Detroit’s failure unusually visible, but the underlying problem was national. The federal Sexual Assault Kit Initiative, launched in 2015, has helped state, local, and tribal jurisdictions inventory, track, test, investigate, and prosecute cases involving previously unsubmitted kits. A 2022 DOJ Inspector General review reported that more than 136,000 unsubmitted kits had been identified and more than 81,500 tested at that point. The program has continued to expand: SAKI now reports more than 205,000 kits inventoried and more than 105,000 sent for testing, generating more than 20,000 CODIS hits.

The national total nevertheless remains difficult to establish because state reporting systems, definitions, inventory requirements, and testing standards are not uniform. Reform has expanded substantially—in late 2024, 33 states and Washington, D.C. had enacted laws requiring statewide rape-kit tracking systems—but gaps in counting and reporting remain, and there is still no single continuously updated federal inventory of every untested kit held by law enforcement agencies and laboratories nationwide. Because no comprehensive national inventory exists, estimates have varied widely. A Congressional Research Service review summarized earlier research that placed the backlog between roughly 90,000 and 400,000 kits—a range that illustrates the difficulty of measuring a problem defined and counted differently across jurisdictions rather than a reliable estimate of the backlog today. End the Backlog now estimates that roughly 100,000 untested kits may remain undiscovered, although no comprehensive national inventory exists to verify that figure.

Detroit’s warehouse was exceptional in scale, but not because it was the only place where sexual-assault evidence had been allowed to accumulate. It was a lens onto a considerably larger national problem.

The Backlog Was Never Just About DNA

Describing the phenomenon as a “rape-kit backlog” can make it sound primarily technical—a matter of too many samples and too few technicians. That was part of the problem, but Detroit demonstrates the inadequacy of the phrase. The backlog implicated policing, budgets, training, leadership, and forensic capacity, as well as the communication among police, prosecutors, laboratories, medical providers, and advocates—and, fundamentally, credibility: the question of whose cases institutions judged to warrant investigative effort.

A jurisdiction can acquire new equipment, hire technicians, and impose testing deadlines, and these reforms matter; but they do not by themselves dislodge the attitudes that allowed evidence to be disregarded. Where investigators continue to view certain complainants as unreliable, a testing mandate may alter what happens to the kit without altering what happens to the person. For that reason, backlog reform cannot end at the laboratory door.

From 11,000 to Zero

Detroit ultimately achieved something notable. By April 2022, laboratory analysis had been completed on all 11,341 kits in Wayne County’s original tracked backlog, and the project had produced investigations, prosecutions, convictions, and the identification of suspected serial offenders across multiple states.

Even the milestone carried a qualification. The original discovery also did not prevent new failures. Hundreds of additional Detroit kits collected in 2010 and 2011 were later found to have gone untested and were not sent for testing until 2015. In at least one case, DNA from a 2010 kit was associated with a suspected serial offender who remained uncharged for years and was later accused of another assault. The term “cleared” generally refers to the older, decades-deep accumulation rather than to recently collected kits still moving through overburdened laboratories, which is why the backlog is better understood as an ongoing condition than a finite quantity to be exhausted.

“Zero untested kits” answers one question without erasing the years in which the number was in the thousands. Testing evidence two decades late cannot recover those two decades. A conviction obtained long after an assault may still matter a great deal—a survivor may value the knowledge that a case was finally deemed worth investigating, a match may provide answers, a prosecution may prevent further offending—but delayed justice carries costs that conviction totals do not capture: witnesses become unavailable, memories change, records are lost, statutes of limitation may intervene, complainants and suspects die, and evidence deteriorates, while individuals who might have been identified earlier may remain at liberty in the interim. Clearing a backlog addresses the evidence that survives; it cannot reconstruct the opportunities lost while that evidence went untouched.

What Detroit Actually Revealed

The central lesson of Detroit is not simply that kits should be tested, though they should. The deeper lesson concerns how institutions fail. Systemic failure is often imagined as dramatic—corruption, conspiracy, the deliberate destruction of evidence, a single official making an indefensible choice. More often, it is mundane. A detective judges that a case will not proceed; a supervisor does not question the judgment; no written submission policy exists; the laboratory is overwhelmed; the unit loses staff; leadership changes; and the next case arrives, and the next.

Each decision may appear individually defensible, and yet the boxes accumulate into the thousands. This is what makes systemic failure difficult to perceive as it unfolds: no single decision needs to produce the outcome, which emerges instead from the accumulation of ordinary decisions within a system that has ceased to examine its own assumptions. Detroit eventually counted the boxes, but the figure of eleven thousand was never the essential point. The essential point was that a justice system could collect evidence from thousands of people, retain it for decades, and fail to recognize the shelves themselves as evidence that something had gone seriously wrong.

Why It Still Matters

Sexual-assault kits occupy an unusual position within the justice system: they are physical evidence, but they are also evidence of institutional choices. Whether a kit is collected reflects access to medical care and to reporting; whether it is submitted reflects investigative practice; whether it is tested reflects laboratory resources and policy; whether the results are pursued reflects staffing and prosecutorial priorities; and whether a survivor is contacted reflects victim-service practices. At each stage, a decision is made. The backlog is therefore not reducible to a problem of DNA; it is a problem of systems.

Detroit’s warehouse made that system legible because its failures accumulated in a single location, close enough together to be counted. Most institutional failures are harder to document, residing in cases never opened, calls never returned, evidence never submitted, reports closed prematurely, policies never written, budgets never funded, and assumptions never examined. The boxes gave those failures a form, and once the facility was opened, the question was no longer whether the evidence existed but why it had gone unexamined for so long.

Federal support remains central to rape-kit reform. The Bureau of Justice Assistance continues to administer SAKI, while the Debbie Smith Act—another major source of federal DNA-backlog funding—has been reauthorized through fiscal year 2029. Continued progress nevertheless depends on appropriations, state implementation, laboratory capacity, investigative staffing, and sustained victim-service resources rather than on testing grants alone.

The conditions that filled the Detroit facility were never confined to Detroit. Although state reforms have expanded considerably, uneven inventory, reporting, testing, and tracking requirements continue to make the full scale of the problem difficult to measure.

Final Thoughts

The ethical weight of the backlog rests on a straightforward observation. The survivors represented in Detroit’s kits had done what the system asked of them: they reported, they submitted to examination, and they provided evidence. The failure was not theirs, and it cannot be explained by technical limitations alone. Detroit’s own research documented staffing shortages, resource depletion, leadership instability, and laboratory constraints—but also a deeper failure of belief: institutional practices in which assumptions about credibility helped determine whose evidence was considered worth pursuing.

Testing a kit cannot undo an assault or restore the years during which the evidence sat unexamined, but it can identify an offender, connect the dispersed cases of a serial predator, and, at times, prevent a subsequent assault. That is what the unopened kits contained all along: not only answers for the people already represented in them, but a measure of protection for those not yet harmed. The lesson Detroit paid for, and that much of the country is still learning unevenly, is that evidence is only as useful as the willingness to examine it.