John Doe, Indicted

October 2, 2026

The Jane Doe Files, Part 2 of 6

In October 1999, a Milwaukee prosecutor named Norman Gahn did something that sounded absurd. With the statute of limitations about to run out on the 1993 kidnap-rapes of three women, he filed charges against a man whose name he didn’t know. The defendant was listed as “John Doe,” identified only by a string of genetic markers (NPR).

The John Doe DNA indictment is the purest expression of what DNA was supposed to do for rape cases: the evidence knows who he is, even when nobody else does. Filing against the profile stops the clock, and the case waits until a database match supplies a name. A Wisconsin appeals court blessed the tactic in 2003 in State v. Dabney, a case where the match arrived three months after a John Doe complaint filed three days before the deadline (NIJ).

Legislatures followed. Today 31 states and DC let prosecutors bring the most serious rape charges against an adult at any time, and five more lift the deadline under conditions, most often when DNA identifies the attacker. Arkansas has the purest version of the John Doe idea: its six-year limit for rape “is eliminated if biological evidence of the alleged perpetrator is identified that is capable of producing a DNA profile.” Fourteen states still have fixed deadlines, from six years in New Hampshire to 25 in Ohio (RAINN State Law Database, 2026).

But a DNA profile only names John Doe if someone actually tests the kit. For decades, in city after city, nobody did.

The boxes on the shelf

In August 2009, 11,341 rape kits that had never been sent to a lab turned up in a Detroit police evidence warehouse, some dating to the 1980s. Prosecutor Kym Worthy made testing them a crusade. By 2022 every kit had been tested, and the county had identified 830 suspected serial rapists, with links to crimes in 39 other states (Fox 2 Detroit; NPR).

Detroit wasn’t an outlier. Memphis found 12,374 untested kits in 2013. Human Rights Watch counted at least 12,669 across Los Angeles County in 2009, and in 499 of the stranger cases, the statute of limitations had already expired (HRW). Houston tested 6,663 backlogged kits and got 850 database hits, but in 44% of those cases it was already too late to prosecute (Texas Monthly). National estimates of the backlog at its peak run from 300,000 to 400,000 kits (NO MORE / Joyful Heart).

Every one of those kits represents a woman who did what the system asked of her: went to a hospital in the worst hours of her life and submitted to a long, invasive exam. Then police left the evidence in a box.

What testing found

The most careful look at what was in those boxes comes from Cleveland. Cuyahoga County’s task force, formed in 2013, has investigated more than 8,000 old cases and indicted more than 940 defendants, with a conviction rate above 90% among those indicted (Cuyahoga prosecutor).

Researchers at Case Western Reserve University’s Begun Center studied a sample of 243 of those kits. Half (51%) were linked to serial offenders. Of those serial offenders, 60% were later arrested for another sexual assault, assaults that might not have happened if the first kit had been tested on time (Case Western).

The same team ran the numbers on cost. Testing a kit cost about $950. Testing and investigating 4,347 kits cost about $9.6 million, against roughly $48 million in harm from future rapes averted. That works out to about $8,900 in harm prevented per kit, and, on conservative assumptions, at least one rape prevented for every four convictions (Case Western; Ohio AG).

The national picture tells the same story with a sobering coda. The federal Sexual Assault Kit Initiative had, by August 2024 (its most recent published totals), tested about 97,000 kits to completion (BJA):

Roughly one tested kit in five produced a database hit. About one in sixty produced a conviction. Decades-old cases are hard to bring to trial: witnesses move, victims have rebuilt lives and don’t want to reopen them, statutes expire. The backlog’s real lesson isn’t about convicting the past. It’s about what the past tells us: the man in the kit had very often done it before, and would do it again.

How the database is supposed to work

The mechanism has two halves, and both must be filled.

1. Known offenders go in. Every state collects DNA from people convicted of felonies. Thirty-four states and the federal government also collect it at arrest (NCSL), a practice the Supreme Court upheld 5–4 in Maryland v. King (2013), comparing a cheek swab to fingerprinting. Alonzo King’s swab, taken on an assault arrest, matched an unsolved 2003 rape (Cornell LII).
2. Crime-scene evidence goes in. That’s the kit, and the evidence from underwear, clothing, and bedding. As of June 2026, the national database held more than 20.3 million offender profiles, 6.4 million arrestee profiles, and 1.49 million crime-scene profiles, and had aided more than 784,000 investigations (FBI).

The backlog broke the second half. You can have every convicted felon on file and catch no one if the evidence never reaches the lab.

The point

If most rapes were one-offs, a rape kit would be a lottery ticket: costly to her, rarely decisive. But the backlog proved that a large share of rapes are committed by men who rape repeatedly. That changes the math. Every kit tested and every offender swabbed is a tripwire for the next assault.

That leaves one gap: a DNA profile that matches no one on file. Since 2018, police have had a way around it: search his relatives instead of him.

Next in the series, Part 3: John Doe’s Family Tree. The DNA search that runs through his relatives, and why so many people won’t help it.

Research and drafting assistance for this piece, including source verification, was provided by Claude (Anthropic). All claims are cited to primary or secondary sources linked inline throughout; the analysis and opinions are my own.