A carrier called me last year to get on my approved list. Seven months of active authority, three trucks. Good SAFER numbers except for the crash history — two crashes in twelve months. When I brought it up, their ops manager said the same thing I've heard a dozen times since: "Both of those are marked non-preventable."
She was right. Both crashes had the FMCSA's non-preventable determination. The SAFER snapshot showed it. So I approved them. What I didn't ask — and should have — is how they got that determination. What types of crashes were they? Did the carrier submit a Request for Data Review, or had the crashes simply aged out of the BASIC calculation on their own? Were the determinations based on reviewed evidence, or had FMCSA flagged them non-preventable because they fell into a qualifying crash category that almost always gets that result?
Those aren't gotcha questions. They're what separates someone who understands the Crash Preventability Determination Program from someone who just learned the phrase.
What the BASIC actually measures — and what it doesn't
The FMCSA's Crash Indicator BASIC measures time-weighted crash involvement in the last 24 months. It's a frequency count at the carrier level. Two fatal crashes moves the needle the same as two property-damage-only crashes — crash severity isn't weighted into the percentile calculation. That's a known limitation and one worth understanding before you lean on the number.
What the BASIC also doesn't tell you is whether the crash was the driver's fault. FMCSA is explicit: the Crash Indicator is an involvement metric, not a culpability metric. The underlying data comes from state crash reports that document a CMV was involved. They don't adjudicate fault.
That's the setup for understanding why the Crash Preventability Determination Program exists — and why it matters for your vetting.
How the CPDP actually works
FMCSA's Crash Preventability Determination Program allows carriers and drivers to request a preventability review for crashes that fit certain categories. The program started as a pilot in 2017 and expanded from there. Eligible crash types include situations where a CMV was struck in the rear, struck by a wrong-way driver, involved in a crash caused by falling cargo from another vehicle, and similar scenarios where the CMV operator arguably couldn't prevent the collision.
Here's how it works: the carrier or driver submits a Request for Data Review (RDR) to FMCSA, along with supporting documentation — police report, dashcam footage, witness statements. FMCSA reviews and returns one of three determinations: Preventable, Not Preventable, or Undecided. If the determination is Not Preventable, the crash is flagged on the SAFER snapshot and excluded from the carrier's Crash Indicator BASIC percentile calculation going forward.
The crash itself still shows up in the raw crash count. Two crashes listed is still two crashes listed. But if one has the Not Preventable flag, it doesn't count toward the SMS percentile — and the percentile is what triggers FMCSA intervention thresholds and, realistically, what most brokers look at.
The three things "non-preventable" might mean
When a carrier tells you their crashes are non-preventable, you're hearing at least one of three different things. You need to know which one.
First: the crash fits one of the RDR categories, the carrier submitted documentation, FMCSA reviewed the evidence, and the determination reflects actual review. That's a valid claim. It tells you something real.
Second: the crash fit a qualifying category and FMCSA made a determination without much friction — certain crash types carry a high likelihood of non-preventable findings based on the fact pattern alone. The carrier filed minimal paperwork and got the flag. Less meaningful than a contested review, but still a real FMCSA determination.
Third: the carrier is using "non-preventable" as shorthand for "it wasn't our fault in our opinion," with no RDR ever filed and no FMCSA determination in the system. I've had this conversation enough times to know that operations managers sometimes learn the phrase without knowing what it requires. They repeat it because it works — until someone looks at the actual SAFER snapshot.
If you can see the Not Preventable flag on SAFER, you're in scenario one or two. If the carrier is telling you something is non-preventable but the flag isn't there — that's scenario three. The practical tell: open the SAFER snapshot for the carrier while you're on the T-call. If what they're describing matches the flag, great. If the flag isn't there, follow up.
What the flag does and doesn't tell you
Even a legitimate FMCSA Not Preventable determination doesn't mean the crash was harmless. It means the program's review found the CMV operator wasn't the proximate cause. That matters for BASIC percentile purposes. It's less meaningful for your vetting decision when the crash involved serious injury or a fatality, regardless of who caused it.
Here's a concrete example. Carrier MC-1247893 / DOT-3567102 has three crashes in SAFER over 18 months. Two are flagged Not Preventable: one was a rear-end strike while stopped, one was a falling-debris incident on I-80. One crash is unflagged — a lane departure, injury to the other driver, $85,000 settlement paid. The unflagged crash doesn't have an RDR because the carrier never submitted one, and the crash type wouldn't qualify anyway.
What's the Crash Indicator BASIC percentile show? It's calculated on just the one lane departure crash, since the other two are excluded. Percentile comes back moderate. No alerts. But the actual picture includes all three crashes — and the lane departure with a paid injury settlement is the one that needs a follow-up question.
That's the trap. Brokers see the percentile, see it's not alarming, and move on. They haven't looked at the raw count, identified which crashes have the flag, and thought through what the unflagged ones actually represent.
How to work through it on a T-call
Pull the SAFER snapshot and count crashes. Then note which ones carry the Not Preventable flag and which don't.
For crashes without the flag — especially any involving injury, fatality, or a lane departure/rollover — ask about them on your T-call. "I see a crash from around [time frame] in your history without a preventability determination. Can you tell me what happened?" You're not interrogating. You're building a record that you looked past the percentile and asked the right question.
For crashes with the flag, you don't need to challenge the determination. But note it specifically. "Two crashes in SAFER carry FMCSA Not Preventable determination per CPDP — reviewed snapshots" is a factual file entry. If either involves serious injury or fatality, add a line about what the crash involved even if the determination favors the carrier.
49 CFR § 385.3 defines a preventable accident as one where the driver failed to do everything reasonably possible to avoid it. The CPDP operationalizes a version of that standard. It's not a legal finding, but it is FMCSA's administrative judgment. In discovery, having that documented — rather than a vague recollection that "they said it was non-preventable" — is the difference between a credible account and a shrug.
Why this matters more after Montgomery
Before the Supreme Court's May 2026 ruling in Montgomery v. Caribe Transport II, a broker in several circuits could lean on FAAAA preemption to push back on state-court negligent selection claims. The Court ended that unanimously — the FAAAA does not preempt those claims. What a broker knew, what they documented, and whether they understood what they were looking at is now live in state court.
Plaintiff's attorneys are running through carrier files. One thing they look for is whether the broker actually understood the carrier's safety record or just clocked a percentile number and moved on. A broker who can only say "I saw the Crash Indicator at the 42nd percentile and approved them" is opening the door to: "Did you know that percentile excluded two crashes because the carrier had filed for CPDP review? And did you know there was a third crash — an injury accident — that no one ever filed on?"
That follow-up question lands differently if your file shows you identified the flagged crashes, noted the unflagged one, asked about it, and logged the answer. It lands very differently if your file is silent on all of it.
How I document this
When a carrier has crashes in SAFER, I add this to the file after pulling the snapshot:
"[Date] — SAFER shows [X] crashes in 24-month window. [Y] carry FMCSA Not Preventable determination per CPDP review — noted in snapshot. [Z] without determination: [brief description of crash type/date/circumstances]. T-call with [name/title] on [date]: [summary of explanation]. No additional safety concerns identified beyond current BASIC profile."
If a crash without a determination involves injury or a settlement, I flag it: "Unflagged crash on approximately [date] involved reported injury and [settlement/litigation, if known]. Asked carrier for explanation per T-call. [Summary]."
Two sentences to three. Sixty seconds. The file now shows you understood what the percentile includes and what it doesn't — and you asked about the gaps.
The bottom line
The Crash Indicator BASIC is one of the more useful signals in the SMS toolkit. But it's only as good as your understanding of what goes into it. The CPDP is a legitimate part of how FMCSA handles crash data — it exists because raw crash counts genuinely can overstate a safe carrier's risk profile when they keep getting struck by wrong-way drivers. The program does real work.
It also gets misused. "Non-preventable" has become a phrase that ops managers reach for when a broker raises crash history, sometimes with a real FMCSA determination behind it and sometimes without. Knowing the difference — in five minutes, by looking at the SAFER snapshot — is the whole job.
Know what the flag means. Know what it doesn't. Ask about the crashes without it. Document both.
— Mason Lavallet
Founder, DOTScreener.com
Automate your carrier vetting
DOTScreener runs every check in this article automatically — live FMCSA data, documented decisions, tamper-evident audit trail.
Related Articles
The Pre-Trip Attestation Is Evidence — Most Brokers Don't Know That
Every driver is required by federal law to inspect their truck and sign off before every trip. That signature — or its absence — shows up in discovery after a crash. Here's what it means for how you select carriers.
Broker GuidesSix Years, 18 Trucks, Never Reviewed: What the Safety Review History on SAFER Actually Tells You
Most brokers check BASIC scores and OOS rates. Almost nobody reads the Safety Review History section on the SAFER snapshot — the field that tells you whether FMCSA has ever actually looked inside this carrier's operation. The absence of a review isn't a neutral fact.
Broker GuidesThe HOS BASIC Isn't a Paperwork Score. It's a Fatigue Warning.
Brokers check the Crash Indicator BASIC because it mentions crashes. They mostly ignore the HOS Compliance BASIC because it sounds like a recordkeeping thing. That's backwards. HOS tells you where the crashes are going.