Crash Preventability in 2026: Rear-Ended and It Still Counts on CSA?

By the TruckerNavi compliance team, led by Dmitry Borovoy, a New Jersey-licensed insurance producer and author of Safety Manager: From Zero to Expert. Published September 25, 2026 | TruckerNavi Inc | 14 min read

The short answer: a car rear-ends your parked truck and the crash lands in your record anyway, because FMCSA counts crashes by involvement, not by fault. The Crash Preventability Determination Program (CPDP) is the one official way to fix that. FMCSA reviews 21 specific crash types, and if you prove the crash was not preventable by your driver, it is excluded from the calculation of your Crash Indicator BASIC in the Safety Measurement System.

Three things decide whether this works for you, and all three are procedural:

• You file it as a Request for Data Review (RDR) in DataQs, choosing the reason "Crash could not be prevented".
• The police accident report is mandatory. FMCSA states plainly that requests with no police accident report are closed immediately as not eligible.
• FMCSA will not review a crash older than 5 years from the crash date, because SMS and PSP only carry crash data that far back.

And the detail almost everyone gets backwards: a Not Preventable determination does not delete the crash. The crash stays published on the SMS website and stays in your record. It simply stops being counted in the Crash Indicator measure and percentile. Everything below is sourced to FMCSA's own CPDP program page, the CPDP FAQs, the December 4, 2024 Federal Register notice (89 FR 96269) and the SMS Methodology.

This is an explanation of the rules, not legal advice about your specific crash.

What does the Crash Preventability Determination Program actually do?

Start with the problem it solves. Your Crash Indicator BASIC is built from state-reported crashes that meet the reportable standard. The definition sits in 49 CFR 390.5T: an accident is an occurrence involving a commercial motor vehicle operating on a highway in interstate or intrastate commerce that results in a fatality, in bodily injury where the person immediately receives medical treatment away from the scene, or in one or more vehicles incurring disabling damage requiring a tow away. The only carve-outs are boarding or alighting from a stationary vehicle and loading or unloading cargo.

FMCSA breaks that into two tests, and one item from each must be true. Vehicle qualification: a GVWR or GCWR over 10,000 pounds used on public highways, or a vehicle designed or used to carry nine or more passengers including the driver, or any vehicle carrying hazardous materials in a placardable quantity. Severity qualification: a fatality, meaning anyone killed in or outside any vehicle involved or who dies within 30 days of an injury sustained in the crash; or an injury with immediate medical treatment away from the scene; or a tow-away of any vehicle disabled by the crash.

Read that and notice what is missing: not one word about fault. A drunk driver crossing the centerline into your trailer produces exactly the same reportable crash as your driver running a red light. At this level nobody is sorting out blame, because the data arrives automatically from fifty states in millions of rows and no human reads it.

The CPDP is the correction layer. FMCSA reviews the crash against a fixed list of 21 scenarios and issues one of three outcomes:

• Not Preventable. Excluded from the Crash Indicator BASIC measure and percentile. Listed on SMS in a separate table. Noted on the driver's PSP record.
• Preventable. Stays in the Crash Indicator, with a public note reading that FMCSA reviewed the crash and determined it was preventable.
• Undecided. The evidence did not allow a conclusive decision. Also stays in the Crash Indicator, with its own public note.

The standard FMCSA applies comes from the safety rating rules, Appendix B to 49 CFR Part 385, section II.B(e): "If a driver, who exercises normal judgment and foresight could have foreseen the possibility of the accident that in fact occurred, and avoided it by taking steps within his/her control which would not have risked causing another kind of mishap, the accident was preventable." Note what that does not say. It does not say "the other driver got the ticket". Preventability and fault are different questions, and the whole program hangs on that difference.

Which crashes are eligible, and what should you attach to each one?

FMCSA expanded the list on December 1, 2024. Crashes on or after that date are judged against the 21 types below. Crashes before that date go under the older, narrower groupings, and the 5-year cutoff has almost emptied that older list by now. The police accident report is required in every single row: the last column is what you add on top of it to move the file from "probably" to "compelling".

The two middle columns are FMCSA's own figures from the June 26, 2026 quarterly snapshot, and they show where the program is actually played: the first two crash types alone account for roughly 42 percent of every request ever filed, and no eligible category has a Not Preventable rate below 84 percent.

Two mechanics worth knowing. First, if you file under the wrong type but the crash fits another eligible one, FMCSA will move it for you rather than kill it, and the quarterly table is compiled after those reassignments. Second, if it fits nothing on the list, the request is closed and no review happens.

Which crashes will never be accepted?

The list of 21 is the list. Everything outside it is out, and FMCSA is explicit about why: it reviews only "less complex crash events that do not require extensive expertise to review". In practice that means the following stay in your Crash Indicator no matter how unfair they feel:

• Anything where your truck struck the other vehicle, except an animal strike or a documented rare event. The eligible list is built around the CMV being struck.
• Crashes older than 5 years. Automatically closed by the system.
• Crashes with no police accident report. Immediately closed as not eligible.
• Crashes not yet in FMCSA's MCMIS database. You must select an existing crash record. If the state has not uploaded it, save the RDR as a draft and come back.
• Anything filed with the wrong crash report number. Automatically closed.

How do you actually file the request in DataQs?

The mechanics are unglamorous and the failure points are all in the same three screens.

1. Log in. Motor carriers log into DataQs through the FMCSA Portal. Drivers and other individuals log in through Login.gov. Registration is free.
2. Start a new request and identify it as being about a crash.
3. Find the crash record. Select the state where the crash occurred, not the state where you are based, choose Crash records as the report type, and search by USDOT number or report number. FMCSA's crash report number format is the two-character state abbreviation plus ten alphanumeric characters, for example OH-xxxxxxxxx. You will find it in your own SMS results under the Crash Indicator BASIC, or in the driver's PSP report under Crash Details. The number the state printed on its own accident form may be different, and DataQs wants the FMCSA one. Note the trap: if you click "I do not see my report on the list" and continue anyway, your request will not be reviewed and will be automatically closed, because a CPDP request must be attached to a record that already exists in MCMIS. FMCSA's own submitter guide warns that it can take up to 90 days after the crash for the report to appear.
4. Choose the reason: "Crash could not be prevented."
5. Pick exactly one crash scenario from the 21. Selecting "Ineligible (None of the above)" and continuing gets the request auto-closed unreviewed.
6. Write the explanation. The free-text field is where you tell FMCSA which page and paragraph of the PAR supports which fact. Reviewers are reading a stack of files, not solving a puzzle.
7. Upload the PAR and tag it correctly. This is the single most common self-inflicted wound: you must select "Police Accident Report (PAR)" from the Document Type dropdown. Upload it as a generic supporting document and DataQs will not recognize it as a PAR, and your request stays a draft. You can fix it later by editing the document type, but only if you notice.
8. Submit. The submit option only appears once a PAR is attached. Without one you can only Save Draft, and a draft is not a submission. If you are waiting on the report, DataQs generates a bar-coded fax cover sheet that routes the PAR to your request automatically, or you can come back later, open the saved RDR from My DataQs and upload it under the Correspondence tab.

If the system itself misbehaves, the numbers are published and the calls are short: DataQs Technical Support (877) 688-2984, option 1, and FMCSA Portal registration and motor carrier PIN assistance (800) 832-5660.

Which documents are mandatory, and which just win the case?

FMCSA is blunt about where the burden sits: the burden is on the submitter to provide compelling evidence that the crash is eligible and not preventable. The Agency has no direct access to police reports. The crash fields in MCMIS are only a subset of what is on the PAR, and a preventability determination needs more than MCMIS holds. If you do not send it, it does not exist.

What are the deadlines, from the crash to the SMS update?

There is no single filing deadline, which is exactly why carriers miss the window. There are five separate clocks.

Add the worst case up: the wait for the state to upload, plus about 90 days of processing, plus up to 60 days to post the result. Half a year can pass between the crash and the notation appearing on SMS even when your side of the file is perfect. That is why the filing date matters, and it is why the next question is not academic.

Why does the state crash report take so long to reach FMCSA, and what does that mean in NY, NJ and FL?

The most common question in the whole process has nothing to do with FMCSA and nothing to do with you: the crash is not in the system yet. Crash data reaches FMCSA from the states, and FMCSA grades every state on how fast it delivers.

The measure is called State Safety Data Quality, Crash Timeliness, and it counts the share of a state's evaluated crash records uploaded within 45 days of the crash. The standard used to be 90 days and was cut to 45 as of the March 2024 snapshot. A state at 75 percent or better is rated Good, 50 to 74 percent is Fair, below 50 percent is Poor.

Nationally, in the July 31, 2026 snapshot (records first uploaded between May 1, 2025 and April 30, 2026), 147,320 of 171,585 evaluated crash records arrived on time, or 86 percent. By rating: 42 states Good, 7 Fair, 2 Poor. The national average tells you nothing, though, because you are waiting on one specific state. Here is the slice that matters most to carriers based in the Northeast and Florida.

One practical conclusion, and it is not a comfortable one: your waiting time is set by a state, not by a rule. The same crash on I-95 in Florida and on I-80 in Ohio reaches FMCSA's systems on very different schedules, so plan by the state where you were hit rather than by a national number.

There is a second effect that cuts both ways. SMS time-weights crashes, and the freshest ones weigh the most. A slow state means the crash lands in your record already carrying a reduced weight, but it also means you start fighting it later. The first half helps you; the second half does not.

What does a "Not Preventable" determination actually change?

This is where most articles on the subject are simply wrong, so read it slowly. FMCSA rejected requests from industry groups to pull Not Preventable crashes off SMS entirely, on the grounds that publishing all crashes gives the most complete picture of a carrier's safety performance. So:

• The crash stays in the record. It remains in MCMIS and remains published on the public SMS website, just relocated into a separate table that FMCSA's submitter guide calls "Reviewed - Not Preventable Crash Activity". The SMS Methodology puts it in one sentence: crashes found not preventable through the CPDP will be listed on the SMS website as "Reviewed - Not Preventable" but excluded from a carrier's measure and percentile in the Crash Indicator BASIC.
• The crash stops counting. It is excluded from the calculation of your Crash Indicator BASIC measure and percentile. That is the entire practical benefit, and it is a real one.
• The driver's PSP record gets the note. The Federal Register notice is explicit that only Not Preventable determinations are noted on the driver's PSP record, on the PSP Detailed Report. Preventable and Undecided are not. For the driver, that makes filing a one-way bet: either a notation in his favour or no notation at all.
• Nothing else moves. The determination does not change your safety rating or your ability to operate, produces no penalties or sanctions, imposes no obligations, and does not alter how FMCSA makes enforcement decisions.

One more thing worth naming: who actually reads it. Your SMS crash tables are public, and a broker, a shipper or an insurance broker will look at them before you get a chance to explain anything. A line reading "Reviewed - Not Preventable" is, in effect, a federal agency saying in writing that this crash was not about your driver. That changes the onboarding conversation even before the arithmetic does. The Crash Indicator percentile itself, by contrast, is not publicly displayed: to see your own you have to log into SMS under your own account.

How is the Crash Indicator calculated, and why does a small fleet lose the most?

The arithmetic is worth seeing, because it tells you which crashes are worth fighting for. The Crash Indicator BASIC measure is the sum of severity- and time-weighted crashes from the past 24 months, divided by average power units multiplied by a Utilization Factor:

Two conclusions follow. File early, because the crash you remove today is carrying a time weight of 3, and the same crash removed in 14 months is carrying 1. And small fleets have the most to gain, because the denominator is small: a single injury crash can move a five-truck carrier's percentile in a way it never could for a large one.

There is a third conclusion almost nobody writes down, and it is the most interesting one if you run two or three trucks. Under the SMS Methodology a carrier is placed in a safety event group and given a Crash Indicator percentile only if it has two or more applicable crashes in the past 24 months, and carriers with no crash recorded in the previous 12 months are removed from the ranking. Put those together: if you have exactly two crashes in two years and one of them comes back Not Preventable, you do not merely improve your percentile, you stop meeting the data sufficiency standard for having one calculated at all. For a small carrier that is a different kind of result, not a few points of improvement.

The intervention thresholds for the Crash Indicator in the SMS Methodology are the 65th percentile for general freight, the 60th for hazmat and the 50th for passenger carriers; a carrier at or above the threshold gets the alert symbol in that BASIC. The full percentile mechanics are in CSA scores explained, and the practical moves are in CSA score improvement strategies.

Why are requests refused, and why do small fleets lose more often?

FMCSA publishes quarterly CPDP statistics, and they demolish the usual assumption that FMCSA is stingy with Not Preventable rulings. As of the June 26, 2026 snapshot:

Read that again: if your file reaches a reviewer, you win about 19 times out of 20. The program does not fail carriers on judgment. It fails them at the front door, on paperwork. Roughly a quarter of all requests die before anyone evaluates whether the crash was preventable, and every one of those causes is something the submitter controls.

Now the split by fleet size. FMCSA publishes it separately, and the pattern is uncomfortable and completely fixable.

A carrier with five trucks or fewer is roughly three times more likely to walk away with a Preventable or Undecided ruling than a carrier with more than 500. The crashes are not different. The files are. Large fleets run dash cameras across the fleet, have someone whose job is pulling police reports, and answer a 14-day document request on day two. A three-truck operation finds out about the deadline on day 20. The gap in that table is a documentation gap, and it is the cheapest gap in trucking to close.

One more number explains why this is worth doing at all. Of the 12,935 carriers that have ever had a crash ruled Not Preventable, 5,245 are above the alert status threshold in at least one BASIC in the current SMS snapshot and 7,690 are below it. In other words, more than forty percent of the carriers using this program are already under the system's attention, not tidying up out of boredom.

When does FMCSA rule "Preventable" on purpose?

The 1,245 Preventable rulings are not random. FMCSA states it will find a crash Preventable if there is evidence the driver or carrier could have prevented it, or was prohibited from operating the CMV at the time of the crash. That second half is the one that surprises people, because it has nothing to do with the crash dynamics:

• Out-of-service violations in effect at the time of the crash.
• License violations. FMCSA uses the MCMIS crash report to confirm the driver was properly licensed, and falls back on the CDLIS record when MCMIS is missing the data or shows the wrong license class for the vehicle. CDLIS is also used to confirm the driver was not operating under an open license withdrawal or a suspension tied to a drug or alcohol violation.
• Driver prohibitions in the Drug and Alcohol Clearinghouse.
• A positive post-accident drug or alcohol test, or a refusal to test. On a fatal crash, this makes the determination Preventable regardless of who hit whom.

In other words: a wrong-way drunk driver can hit your truck head-on, and if your driver was prohibited in the Clearinghouse that morning, the crash is Preventable. Your compliance file is part of the crash file. That is also why post-accident testing discipline matters far beyond the testing rule itself, which is covered separately in post-accident testing rules.

Which yields a rule worth saying out loud: audit your own driver before you file. Licence valid on the crash date, class matching the vehicle, medical certificate in force, Clearinghouse record clean, no out-of-service order in effect. If any of those is wrong, filing hands you a Preventable ruling with a public notation on SMS, which is worse than doing nothing. Filing a CPDP request is, among other things, a voluntary federal audit of your DQ file.

The 24-month trap: a determination that arrives too late to help

Here is a mismatch that almost nobody explains. The CPDP accepts crashes up to 5 years old. SMS only displays crashes for 24 months. FMCSA's own statistics show the consequence: of 71,446 cumulative Not Preventable determinations, only 21,197 were in SMS as of the latest snapshot. The same split shows up by company: 12,935 carriers have had a crash ruled Not Preventable over the life of the program, but only 6,203 of them can see one in the current SMS snapshot. More than two thirds of all Not Preventable rulings apply to crashes that are no longer in SMS at all.

So a determination on a four-year-old crash will do exactly nothing for your Crash Indicator, because that crash aged out of the calculation two years ago. It is not worthless: it still lands on the driver's PSP record, which carries five years of crashes, and PSP is what your next customer's insurer and your next hiring manager actually read. But if your goal is the Crash Indicator, the practical window is roughly the first 18 months after the crash: long enough for the report to exist and the review to finish, short enough that the crash is still being counted when the result posts. Do the arithmetic: up to 90 days for the crash to appear, plus about 90 days of processing, plus up to 60 days to post. A request filed at month 20 may not land before the crash falls out on its own.

Does a Not Preventable determination help with insurance or a lawsuit?

No, and this is where the marketing gets loose. FMCSA states directly that these determinations do not establish legal liability, fault or negligence by any party, that fault is generally determined in civil or criminal proceedings, and that under 49 U.S.C. 504(f) the Agency's preventability determinations may not be admitted into evidence or used in a civil action for damages and are not reliable for that purpose. A CPDP letter is not a defense exhibit.

Nor does it touch your safety fitness. The CPDP does not affect preventability determinations made during FMCSA safety investigations under 49 CFR Part 385, does not change the safety rating standard, and produces no penalties or sanctions. Anyone quoting you a specific premium reduction for filing RDRs is quoting themselves: FMCSA declares no link between SMS data and what an underwriter charges you, and neither FMCSA nor state insurance regulators publish any rule on how a carrier's public SMS page must be read.

What it does deliver is narrow and genuinely worth having: a Crash Indicator built from crashes your driver could actually have prevented, fewer reasons for FMCSA to prioritize you for an intervention, and a driver whose PSP record shows the record instead of just the collision. For a small fleet where one crash swings the whole measure, that is enough.

What do you do if the answer comes back Preventable or Undecided, and what does the filing checklist look like?

You are not finished, but you get a limited number of moves.

1. Reopen the RDR, once. The Federal Register notice states that an RDR that came back Preventable or Undecided, or that was closed for failure to submit requested documents, may be re-opened once. Reopening without adding anything new gets it re-closed unchanged, so bring evidence. The statistics back that up: of 6,005 reopened requests, 3,947 still closed as not eligible.
2. Or file a new RDR with the additional documents or evidence. FMCSA states it will reconsider a request when the submitter provides additional documentation to support it.
3. Fix the crash type if that was the problem. If your request died under the video catch-all because the crash actually fits one of the other 20 types, resubmit under the correct type and include the video as support. FMCSA spells this out in its FAQ: that category is reserved for crashes that meet none of the other 20.
4. Chase the missing document specifically. An Undecided ruling usually means one identifiable gap: the medical certificate as of the crash date, the drug and alcohol result on a fatal crash, a legible PAR page. Find that one item.
5. Remember it works in both directions. Anyone with documentation or data to refute a determination may send it to crash.preventability@dot.gov, and determinations may be changed as a result. A Not Preventable is not permanently bulletproof either.

And keep the two review types separate. If the underlying record is wrong rather than the preventability judgment, that is an ordinary DataQs data-correction request against the state agency, and the mechanics of those are a different exercise entirely.

And now the checklist that keeps a request alive. It is short, because all the real losses happen in the same ten places.

1. Treat every reportable crash as a filing candidate on day one, not when a broker mentions it. Order the police report immediately, while the records office is quick and the crash is fresh.
2. Pull and preserve the camera footage the same day. Most systems overwrite. Save the relevant minutes, not the shift.
3. Photograph the scene and the damage geometry. Impact points tell the reviewer who hit whom when the narrative is ambiguous.
4. On a fatal crash, run post-accident testing by the book, and if a test could not be done, write and keep the record explaining why. FMCSA will ask.
5. Check MCMIS monthly until the crash appears. Do not click "I do not see my report on the list".
6. Pick the crash type deliberately. Use the video catch-all only when nothing else fits.
7. Tag the PAR as a PAR in the dropdown. Then confirm it shows under Police Accident Report Documents, not under Supporting Documents.
8. Put the 14-day clock in a calendar with an owner's name on it. That single habit closes most of the gap between three-truck and 500-truck outcomes.
9. Keep the vehicle-condition paperwork clean anyway. A crash file sits next to your inspection history, and the defect trail runs through your daily reports, covered in DVIR rules.
10. Do not stop at the crash. The same investigator who sees a preventable crash sees your inspection record, which is why the CVSA inspection blitzes and winter state chain laws belong in the same plan as your crash paperwork.

FAQ

Related articles

• CVSA Roadcheck and Brake Safety Week: What Actually Happens During the Blitzes
• DVIR Rules: Daily Vehicle Inspection Reports Done Correctly
• Chain Laws by State: Who Requires What and When
• The Bridge Formula and Axle Weight Limits
• Post-Accident Drug and Alcohol Testing: When It Is Required
• CSA Scores Explained: BASICs, Thresholds, Methodology
• CSA Score Improvement Strategies That Actually Move the Number
• DataQs Challenge: How to Dispute an Inspection Violation
• Crash Preventability Determination Program: русская версия этой статьи