Indianapolis Dock Door Repair Research
Loading Dock Accident Statistics: Fatalities, Falls, Crush Events and Forklift Incidents
· Version 0.1
Scope: Version 0.1 is a verified claim audit and methodology release. It does not yet contain the reviewed incident database or the reproduced BLS annual series. Both are planned for version 1.0.
Most loading dock accident statistics in circulation do not survive contact with their sources. We attempted to trace fourteen of the most-repeated figures to primary documents. Not one of the fourteen was both traceable to a primary source and accurate as circulated.
The best-known claim — that roughly 25% of industrial accidents happen at loading docks — we could not trace to any primary source, and OSHA’s own warehousing page makes no such claim. The second best-known — that 7% of forklift accidents involve a forklift falling off a dock — misstates the study it comes from. NIOSH recorded 39 forklift deaths from falls off a ramp or loading dock among 1,021 forklift fatalities between 1980 and 1994. That is 3.8%. Only 23 of those 39 involved a loading dock specifically, which is 2.3%. Those are fatalities, not accidents, and the underlying data is more than thirty years old.
That is the short version. The longer version is more useful, because the federal government does publish real loading dock evidence. It is just scattered across three systems that measure different things, use different classification codes, and cannot be added together. Below is what each one contains, what it can support, and where every widely circulated number came from.
Headline findings
| Finding | Figure | Source | Verified |
|---|---|---|---|
| Forklift deaths from a fall off a ramp or loading dock, 1980–1994 | 39 of 1,021 forklift fatalities (3.8%) | NIOSH NTOF; Collins et al. 1999 | |
| Of those, deaths where the forklift fell from a loading dock specifically | 23 of 1,021 (2.3%) | NIOSH NTOF; Collins et al. 1999 | |
| Deaths where a tractor-trailer rolled or pulled forward and the forklift fell between dock and trailer | 3 of 1,021 (0.3%) | NIOSH NTOF; Collins et al. 1999 | |
| OSHA accident summaries matching the exact phrase “Loading Dock” | 897 raw keyword matches | OSHA IMIS Accident Search | |
| Most recent event date in that OSHA result set | January 23, 2025 | OSHA IMIS Accident Search | |
| BLS injury source code for docks, reference years 2011–2022 | OIICS 6692, “Ramps, loading docks, dock plates” | BLS OIICS Manual 2.01, §3.3, p.407 | |
| Repeated dock claims we could trace to a primary source and confirm as stated | 0 of 14 | This audit | |
| OSHA provision requiring wheel chocks or sand shoes during dockboard use | 29 CFR 1910.26(d) | OSHA standard text |
Source: Indianapolis Dock Door Repair Research analysis of OSHA, BLS and NIOSH primary sources, version 0.1, July 18, 2026.
Before you use the 897 figure
The figure “897” is a keyword match count. It is not a count of loading dock accidents, and we can demonstrate why with the records themselves. See the OSHA methodology section for the three structural problems that separate the keyword count from a true incident total.
What do loading dock accident statistics actually measure?
There is no single federal number for how many loading dock accidents happen in the United States. This audit found no federal publication reporting one. Three separate systems capture pieces of it. They cover different populations, different periods and different definitions of an “accident,” and adding them together produces a number that means nothing.
| System | Unit measured | Fatal / nonfatal | Count or estimate | Classification | Known coverage gap | Combinable? |
|---|---|---|---|---|---|---|
| OSHA IMIS Accident Search | Investigated incidents, as narrative summaries | Both, heavily weighted to fatal and catastrophic | Neither — a selected subset of cases OSHA investigated | Free-text keywords, no injury-source code | Not a census; excludes incidents OSHA never investigated | No |
| OSHA Severe Injury Reports | Employer-reported amputations, hospitalizations, eye loss | Nonfatal, severe only | Reported case records | OIICS-coded | Federal OSHA jurisdiction only; State Plan states excluded | No |
| BLS SOII | Nonfatal injuries and illnesses with days away from work | Nonfatal | Estimate from an employer sample | OIICS source code | Sampling error; excludes some employer classes | No |
| BLS CFOI | Fatal work injuries | Fatal | Count, compiled from multiple source documents | OIICS source code | Different universe from SOII | No |
| NIOSH NTOF (Collins et al.) | Forklift-related fatalities, 1980–1994 | Fatal | Count from death certificates | Manual coding into 13 circumstance categories | Captures ∼80% of work fatalities; forklift keyword required | No |
Source: Indianapolis Dock Door Repair Research synthesis of OSHA, BLS and NIOSH program documentation, version 0.1, July 18, 2026.
A practical illustration of why they cannot be mixed. We ran the “Loading Dock” search on July 18, 2026, and the most recent event in the entire result set was dated January 23, 2025 — a gap of about eighteen months. That observed gap does not establish a fixed lag for OSHA’s whole database, but it does mean any “dock accidents this year” figure built on this result set is describing a period that is still filling in.
What none of these sources measure
- The total number of loading dock incidents at U.S. facilities.
- Near misses of any kind.
- Property-damage-only events.
- Injuries never reported, never recorded, or never investigated.
- Any denominator that would let you state what share of accidents happen at loading docks.
That last one is why the most famous dock statistic falls apart.
Which loading dock statistics could not be verified?
We attempted to trace fourteen widely repeated loading dock claims to primary sources. Three led to a real primary document that the circulating version then misstates. Eight we could not trace at all. Three have a plausible route we have not yet run. None was confirmed accurate as circulated.
A note on what these verdicts mean. “Not located” does not mean “false.” It means we did not find a primary document stating the claim, and neither did the pages repeating it. If you need a defensible number, use one from a row where we reached the source. If you can point us to a primary document for anything marked not located, we will correct this table and log the change in the version history.
Verdict definitions. Traced, distorted in circulation — a real primary figure exists and the circulating version changes its numerator, denominator, date or scope. Not located — no primary document stating the claim was found in this audit. Not yet reproduced — a plausible primary route exists but we have not run it.
| Claim as circulated | Attributed to | Verdict | What we found |
|---|---|---|---|
| ∼25% of industrial accidents occur at loading docks | OSHA; “OSHA data and BLS reporting”; Industrial Safety & Hygiene News; often nothing | Not located | We found no primary OSHA or BLS document stating it. OSHA’s warehousing topic page makes no such claim — the agency instead identifies musculoskeletal disorders from lifting and lowering, and struck-by incidents involving powered industrial trucks, as the most common warehousing injuries. We also found no published national count of “industrial accidents” that would serve as the denominator. |
| ∼25% of warehouse accidents occur at loading docks | OSHA | Not located | The same figure circulates with three different denominators: “industrial accidents,” “warehouse accidents,” and “loading dock and warehouse.” Those are not interchangeable populations. A statistic whose denominator changes between sources is not a statistic. |
| ∼600 near misses for every dock accident | Usually unattributed | Not located | We found no dock-specific surveillance dataset supporting a 600:1 ratio. The structure resembles the general accident-triangle literature, which is not dock-specific and is itself contested. |
| 7% of forklift accidents occur when a forklift falls off a dock | “A study of forklift-related deaths, 1980–1994” | Traced, distorted in circulation | The study is Collins et al. (1999), NIOSH NTOF data. It found 39 deaths from a forklift falling off a ramp or loading dock among 1,021 forklift fatalities — 3.8%. Loading docks specifically account for 23, or 2.3%. The circulating claim changes the numerator, the denominator and the event type. |
| ∼100 forklift deaths and ∼34,900 serious injuries per year | “Latest estimates from OSHA” | Traced, distorted in circulation | Collins et al. attribute these figures to a 1995 U.S. Department of Labor estimate published with OSHA’s proposed rule for powered industrial truck operator training. The same paper separately cites ∼100 fatalities and ∼34,000 serious injuries to 1982 California data and a 1987 paper. We have not independently opened the 1995 notice. Either way, presenting a 1995 estimate as current is a thirty-year error. |
| 35,000–62,000 forklift injuries annually | OSHA | Not located | Repeated without a link. Appears on pages alongside 1994 and 2007 figures presented as one contemporary picture. We found no primary document publishing this as a single range. |
| 94,750 Americans injured by forklifts annually | “2007 NIOSH data” | Not located | We found no NIOSH publication reproducing 94,750. Do not describe it as NIOSH data without the actual NIOSH document. |
| Each forklift has a 1-in-10 annual chance of being in a dock accident | A dock equipment manufacturer | Not located | No fleet-exposure denominator, methodology or study is given anywhere it appears. |
| 1 in 6 U.S. workplace deaths is forklift-related | A dock equipment manufacturer | Not located | We found no primary source supporting it. Check it against CFOI totals for any recent year — forklift-related deaths are a small fraction of annual U.S. workplace fatalities, not a sixth of them. |
| ∼33,000 employees missed work from loading dock, dock plate and ramp injuries, 2015–2020; median 14 days away | BLS | Not yet reproduced | Consistent with a SOII pull on OIICS source code 6692, which is real and correctly named in the better sources. We have not run the annual tables. No total is asserted here until the ownership scope, source-versus-secondary-source selection, rounding and aggregation are published alongside it. |
| ∼6,600 employees missed work in 2018 from the same category | BLS | Not yet reproduced | Same route. The specific table to check is BLS SOII Table R57 for 2018 — nonfatal injuries and illnesses involving days away from work by source of injury, private industry — at source code 6692. We confirmed that table exists; its all-sources total for 2018 was 900,380 cases. We have not yet extracted the 6692 row. |
| 49 workers died from loading dock incidents, 2015–2020 | BLS | Not yet reproduced | Would come from CFOI. Verify whether it is source-coded to 6692 or derived another way. |
| OSHA investigated 209 dock incidents 2004–2014 / 17 in 2024 / 15 in 2021 | OSHA | Not located | These read like single-phrase IMIS keyword pulls presented as complete counts. As shown below, a single-phrase pull is neither complete nor deduplicated. None of these sources publishes its search protocol. |
| The dockboard rule is at “49 CFR 1910.26” | Multiple trade outlets | Traced, distorted in circulation | The rule exists; the citation is wrong. It is 29 CFR 1910.26. Minor, but it is an uncorrected error propagating through the same literature that carries the bigger ones. |
Source: Indianapolis Dock Door Repair Research claim audit, version 0.1, July 18, 2026. Verdicts reflect sources locatable as of that date. The row-level circulation log, with each page reviewed and each query run, publishes with the version 0.1 audit CSV.
One possible route to the 7% error
We cannot confirm where “7%” came from and we are not asserting an explanation. It is worth noting that the same NIOSH study contains a separate category — “forklift fell on decedent” — with 73 deaths, or 7.1% of the 1,021 total. That is a different circumstance from a forklift falling off a dock. Whether the two were confused somewhere in the citation chain, we cannot say.
How many loading dock accidents does OSHA actually record?
OSHA’s exact-phrase search for “Loading Dock” returned 897 accident summary matches when we ran it on July 18, 2026. That is a count of keyword matches, not a count of loading dock accidents. Three structural problems separate the two numbers, and all three are visible on the first page of results.
Problem one: some matches are not loading dock accidents
Summary Nr 171294.015, dated October 24, 2024, is described in OSHA’s own result list as an employee dying from cardiac arrest while docking out of a warehouse. It matches the keyword. It is a medical event, not a dock-mechanism injury. Any total including it is measuring something other than loading dock accidents.
This is not a quirk — it is what keyword search does. Records match on any appearance of the phrase anywhere in the narrative, including incidents that merely happened near a dock, and including “dock” used as a verb.
Problem two: one event can generate more than one record
The two most recent matches in the set, Summary Nrs 178473.015 and 173669.015, share the same event date of January 23, 2025 and the same Report ID, 0625400. They carry different industry codes: 484121 (general freight trucking) and 312112 (bottled water manufacturing).
That is exactly what a single fatality looks like when it happens to a visiting driver at a customer’s facility. OSHA opens inspections on more than one employer, and the same death can surface more than once in a keyword search.
The verified record behind 178473.015 reads as follows. At 6:24 p.m. on January 23, 2025, a truck driver working for a general freight service was walking along the apron near a loading dock. A coworker backing an 18-wheeler into the dock struck him and pinned him against the dock structure. He was 39 years old and died of multiple injuries. Inspection 1836830.015 was opened the following day.
A retrieval warning worth passing on
When we requested the detail page for 173669.015, OSHA’s server returned the detail page for 178473.015 instead. That behavior is consistent with the two summary numbers resolving to one underlying record — but it is also consistent with a caching or session artifact, and we could not distinguish the two. We therefore treat this pair as a confirmed flag requiring review, not as a confirmed duplicate. Anyone reconciling these records should note that the endpoint may not return the record you asked for, and should verify the summary number shown on the page they actually receive.
Problem three: Report ID is not a unique event identifier
This is the trap that will produce a wrong answer for anyone deduplicating the obvious way.
It is tempting to treat OSHA’s Report ID as an event key and merge records sharing one. It is not an event key. In the same twenty-record result set we retrieved, Report ID 0521700 appears on three different event dates — September 3, 2024, August 6, 2024 and March 26, 2024 — across three unrelated industries. Report IDs 0626600 and 0552652 each appear on two different dates.
We have not found an OSHA data dictionary defining what Report ID represents, so we make no claim about that. What the records establish is narrower and sufficient: the same Report ID appears on unrelated events, so it cannot be used alone to deduplicate. Deduplicating on it would collapse separate fatalities into single events and undercount the total.
The workable rule: matching event date plus matching Report ID is a flag for manual review, never a merge. Confirmation requires comparing the abstract, the time of day, the victim’s age and occupation, and the linked inspection numbers.
What this means for the 897
The honest statement is the one we will keep making. 897 is the number of accident summaries in OSHA’s investigation database containing the exact phrase “Loading Dock,” as of July 18, 2026.
It is neither an upper nor a lower bound on the true number of loading dock incidents. False positives and duplicate representations push the match count up. Unmatched terminology — records describing only a “dock plate,” a “leveler,” or a trailer that “pulled away” — and incidents absent from OSHA’s investigation database entirely push it down.
Version 1.0 is planned to include a deduplicated, manually reviewed corpus with an exclusion log and duplicate crosswalk. Until it publishes, no unique incident total appears on this page.
What does BLS measure, and why did the dock category change in 2023?
BLS codes the “source” of every injury and fatality using the Occupational Injury and Illness Classification System (OIICS). For reference years 2011 through 2022, loading docks sat in code 6692, “Ramps, loading docks, dock plates.” That code bundled ramps together with docks, so no BLS figure from those years isolates loading docks by themselves.
We verified the code directly in the manual. Division 6 (Structures and Surfaces) → group 66 (Floors, walkways, ground surfaces) → subgroup 669 (Other floors, walkways, ground surfaces) → 6692, on page 407 of the OIICS 2.01 manual, section 3.3.
This code is why several BLS-attributed figures in circulation are plausible. When a source says “employees who missed work because of injuries incurred on loading docks, dock plates and ramps,” it is quoting the 6692 category title almost verbatim. That does not make the number right, but it does mean whoever produced it knew where to look.
The 2023 series break
Beginning with reference year 2023, BLS implemented OIICS version 3.0, a comprehensive revision. BLS states plainly that case-characteristic data for 2023 is a break in series with prior years. Division 6 was revised, with reduced detail for building types and simplified codes for ground surfaces. The current maintenance version is OIICS 3.03, issued May 2026.
Separately — and this is often conflated — SOII case-and-demographic estimates moved to biennial reporting. That reporting change is not the same event as the OIICS 3.0 taxonomy break, and the two should not be described as one.
The practical consequence: you cannot draw one continuous trend line through BLS dock injury data across 2022 and 2023. Any chart that does is comparing two different taxonomies.
| Reference years | Classification | Dock source code | What it covers |
|---|---|---|---|
| 2011–2022 | OIICS 2.01 | 6692 | Ramps, loading docks and dock plates, combined in one category |
| 2023 onward | OIICS 3.x | Confirm against the OIICS 3.03 workbook, Source of Injury tab | Division 6 restructured and ground-surface codes simplified; ramps and docks are no longer necessarily bundled |
| Across the boundary | — | Not comparable | BLS designates 2023 a break in series for case characteristics |
Source: BLS Occupational Injury and Illness Classification Manual, versions 2.01 and 3.03. Verified .
We have not opened the OIICS 3.03 workbook ourselves, so we are not printing a current code number. On a page whose central finding is that people publish figures they have not checked, printing an unverified classification code would be self-refuting. When we read it, it goes in this table with a verification date.
Two more distinctions that get collapsed
Primary versus secondary source. OIICS records both. A dock can be the primary source of an injury — a worker falls off the edge — or a secondary source, where a forklift is primary and the dock is contributing context. Figures that do not state which they used are not comparable to figures that do.
Estimates versus counts. SOII produces estimates from a sample, subject to sampling error and rounding. CFOI produces counts. OSHA investigation records are neither. One SOII estimated case is not one OSHA record, and the three should never be summed.
What actually kills people at loading docks?
The most detailed public breakdown of dock-related forklift deaths comes from a NIOSH surveillance study of 1,021 forklift fatalities between 1980 and 1994. Falls off a ramp or loading dock accounted for 39 deaths — the seventh most common circumstance, well behind overturns (222) and pedestrians struck by forklifts (208). The study is decades old and should be cited as history, not as a current picture. It is the primary source behind the circulating 7% dock-forklift claim audited above.
Here is the full circumstance table, reproduced from the study. The article is expressly identified as a U.S. Government work in the public domain in the United States.
| Circumstance of fatal injury | Deaths | Share of 1,021 |
|---|---|---|
| Overturn | 222 | 21.7% |
| Pedestrian struck by forklift | 208 | 20.4% |
| Crushed by forklift, not elsewhere specified | 162 | 15.9% |
| Fell from forklift | 94 | 9.2% |
| Forklift fell on decedent | 73 | 7.1% |
| Forklift load fell on decedent | 62 | 6.1% |
| Forklift fell off ramp or loading dock | 39 | 3.8% |
| Forklift maintenance or repair | 16 | 1.6% |
| Forklift collision with other vehicle | 15 | 1.5% |
| Contact with a power line | 14 | 1.4% |
| Struck by a part of the forklift | 8 | 0.8% |
| Other forklift-related incidents | 42 | 4.1% |
| Insufficient detail to classify | 66 | 6.5% |
| Total | 1,021 | 100% |
Source: Collins JW, Landen DD, Kisner SM, Johnston JJ, Chin SF, Kennedy RD. “Fatal Occupational Injuries Associated With Forklifts, United States, 1980–1994.” American Journal of Industrial Medicine 36:504–512 (1999). NIOSH National Traumatic Occupational Fatality surveillance system. Counts are the study’s; percentage shares calculated by Indianapolis Dock Door Repair Research.
Breaking down the 39
The study goes one level deeper than almost anyone who cites it, and this is where the useful detail lives.
| Sub-circumstance within the 39 dock and ramp deaths | Deaths | Share of the 39 | Share of all 1,021 |
|---|---|---|---|
| Forklift fell from a loading dock | 23 | 59% | 2.3% |
| Forklift fell from a tractor-trailer truck | 12 | 31% | 1.2% |
| Forklift fell over the side of a flatbed trailer | 3 | 8% | 0.3% |
| Tractor-trailer rolled or pulled forward and the forklift fell between the dock and the trailer | 3 | 8% | 0.3% |
Source: Collins et al. 1999, narrative text. Counts are the study’s. The published narrative lists sub-counts of 23, 12, 3 and 3, which total 41 against a parent category of 39 fatalities; the paper does not explain the discrepancy. The study reports 59% and 31% for the first two rows; the 8% figures and all shares of 1,021 are calculated by Indianapolis Dock Door Repair Research.
That last row deserves a pause. In this 1980–1994 dataset, three of 1,021 forklift fatalities involved a tractor-trailer rolling or pulling forward while the forklift fell between the dock and the trailer. Trailer separation is written about constantly in dock safety material. This is the number the primary data actually carries for it in the one national study that counted.
Three further dock-relevant findings from the same study:
- Age. The study reports that over 40% of the 39 workers were under 24. Its two age bands — 18% aged 16 to 19 and 24% aged 20 to 24 — sum to 42%.
- Sex. All 39 were male.
- Classification uncertainty, disclosed by the authors. For the 73 deaths coded “forklift fell on decedent,” the study states the available detail was insufficient to distinguish a fall from a loading dock, a machine falling on a worker performing maintenance underneath it, or an operator thrown from an overturning forklift. Some of those may have involved loading docks. The paper does not quantify how many, if any.
One caution about a related figure. The study notes tractor-trailers were involved in half of the 15 forklift-vehicle collision deaths, “presumably in the loading dock area of the plant.” That word is the authors’ — it is an inference in the original paper, not a finding, and should not be cited as one.
What this table does and does not support
It supports a ranked picture of how forklift deaths happened across a fifteen-year national window, and a defensible correction to the circulating 7% claim.
It does not support any statement about current forklift risk, about non-fatal dock injuries, or about dock incidents that did not involve a forklift. NIOSH’s surveillance system captures roughly 80% of work-related fatalities, and this study only found cases where “forklift” or a synonym appeared on the death certificate. It is an undercount of an already narrow slice.
What does OSHA require at loading docks?
There is no single OSHA “loading dock standard.” Dock work is governed by several general industry standards, the most specific of which is 29 CFR 1910.26, “Dockboards.” A dockboard is defined in the regulation as a portable or fixed device that spans a gap or compensates for an elevation difference between a loading platform and a transport vehicle, expressly including bridge plates, dock plates and dock levelers.
| Provision | What it requires | Applies to | Effective-date boundary | Exception |
|---|---|---|---|---|
| 1910.26(a) | Dockboards capable of supporting the maximum intended load, in accordance with § 1910.22(b) | All dockboards | None | None stated |
| 1910.26(b)(1) | Designed, constructed and maintained to prevent transfer vehicles from running off the dockboard edge | Dockboards put into initial service on or after January 17, 2017 | Yes — the run-off requirement attaches to this service date | See (b)(2) |
| 1910.26(b)(2) | Run-off protection not required | All dockboards | None | Where the employer demonstrates there is no hazard of transfer vehicles running off the edge |
| 1910.26(c) | Secured by anchoring or by equipment or devices preventing movement out of a safe position; where securing is not feasible, sufficient contact between the dockboard and the surface to prevent movement | Portable dockboards | None | Infeasibility triggers the contact requirement, not an exemption |
| 1910.26(d) | Measures such as wheel chocks or sand shoes used to prevent the transport vehicle — truck, semi-trailer, trailer or rail car — from moving while employees are on the dockboard | All dockboard use | None | None stated |
| 1910.26(e) | Equipped with handholds or other means permitting safe handling | Portable dockboards | None | None stated |
Source: 29 CFR 1910.26, OSHA standard text. Verified .
Two things worth flagging for anyone writing about trailer separation. Paragraph (d) is the trailer-movement provision — it names wheel chocks and sand shoes explicitly and covers rail cars as well as road vehicles. It is regularly discussed in the dock safety literature without being cited. And paragraph (b)(1)’s run-off requirement attaches to dockboards put into initial service on or after January 17, 2017. Older dockboards remain subject to the section’s other requirements. That creates a verifiable equipment-age distinction; we have no data on what share of the installed base predates that date, and neither does anyone else we found.
Section history. § 1910.26 carries the citation history [39 FR 23502, June 27, 1974, as amended at 43 FR 49745, Oct. 24, 1978; 49 FR 5321, Feb. 10, 1984; 55 FR 32014, Aug. 6, 1990; 81 FR 82990, Nov. 18, 2016]. Dockboards have been federally regulated since 1974. The 2016 amendment is the walking-working surfaces final rule, which moved dockboards out of the former § 1910.30(a) into their own section.
Other standards reaching dock work: 29 CFR 1910.22 (general requirements for walking-working surfaces), 1910.176 (materials handling and storage), and 1910.178 (powered industrial trucks, including operator training). ANSI MH30.2, Portable Dock Leveling Devices, is the relevant voluntary consensus standard for portable dock leveler design; it is not itself an OSHA regulation.
Why loading dock safety is under scrutiny right now
OSHA’s National Emphasis Program on Warehousing and Distribution Center Operations, directive CPL 03-00-026, began inspections on October 13, 2023. It applies agency-wide and directs enforcement resources toward warehousing, distribution centers and related operations — the settings where loading docks are.
OSHA’s own framing of warehouse hazards is worth repeating precisely, because it differs from how the topic is usually described commercially. The agency identifies the most common warehousing injuries as musculoskeletal disorders, primarily from overexertion in lifting and lowering, and being struck by powered industrial trucks and other materials handling equipment. It does not single out the loading dock as a share of all accidents.
| Date | Event | Source |
|---|---|---|
| June 27, 1974 | Dockboard requirements first published at 39 FR 23502 | 29 CFR 1910.26 citation history |
| Nov. 18, 2016 | Walking-working surfaces final rule; § 1910.26 text at 81 FR 82990 | 29 CFR 1910.26 citation history |
| Jan. 17, 2017 | Run-off protection requirement attaches to dockboards entering initial service | 29 CFR 1910.26(b)(1) |
| Oct. 13, 2023 | Warehousing NEP inspections begin under CPL 03-00-026 | OSHA Warehousing topic page |
| Reference year 2023 | OIICS 3.0 implemented; BLS case-characteristic series break | BLS OIICS manual |
| May 2026 | OIICS 3.03 issued | BLS OIICS manual |
Source: Indianapolis Dock Door Repair Research timeline assembled from OSHA and BLS primary documents, version 0.1, . The timeline reports dated regulatory and statistical events; it is not evidence of a trend in dock accidents.
How we built and verified this
Every figure presented on this page as verified was checked against a primary source on . The audit table also reproduces circulated figures that remain unverified; those are labeled as claims under audit, not as findings.
When a circulated figure could not be reproduced from a primary document, we kept it only as the exact claim being audited and marked it accordingly. We did not promote any of them to a finding.
The queries and documents
The OSHA accident search, exact phrase, keyword list enabled:
https://www.osha.gov/ords/imis/AccidentSearch.search?acc_keyword=%22Loading+Dock%22&keyword_list=onIndividual accident records, substituting the summary number:
https://www.osha.gov/ords/imis/accidentsearch.accident_detail?id=178473.015The BLS classification manual section where we verified code 6692:
https://www.bls.gov/iif/definitions/oiics-manual-2010-section-3-3.pdfThe NIOSH forklift study, hosted by CDC:
https://stacks.cdc.gov/view/cdc/196724/cdc_196724_DS1.pdfThe OSHA standard text for § 1910.26:
https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.26What we did with them
We ran the OSHA search, recorded the result count and date, reviewed the twenty summaries on the first results page, and opened the detail records cited in this analysis. We read the OIICS source code structure directly in the manual rather than relying on any description of it. We read the NIOSH study end to end, extracted its circumstance table, and calculated the percentage shares ourselves from the published counts. We read § 1910.26 in full from OSHA’s standard text.
For the claim audit, we searched for each circulating statistic, collected the pages repeating it, recorded what each attributed it to, and attempted to reach that attributed source. A claim reached a verified verdict only when we opened a primary document stating it. The row-level circulation log — each page reviewed, its publisher, and the query used — publishes with the version 0.1 audit CSV.
The version 1.0 protocol
A single phrase is not a complete search. The planned version 1.0 protocol runs each of these as a separate exact-phrase query, then merges and deduplicates at the event level:
loading dock · shipping dock · receiving dock · loading bay · dock plate · dockboard · dock board · dock leveler · bridge plate · vehicle restraint · trailer restraint · trailer separation · trailer creep · dock walk · premature departure · pinned between trailer and dock · forklift fell off dock
Inclusion:
An event qualifies when a loading, receiving or shipping dock — or its trailer interface, edge, plate, leveler, restraint or associated equipment — materially contributed to the injury mechanism.
Exclusion:
Marine and waterfront docks; “dock” used as a verb; equipment docking references; medical events that merely occurred at a dock; yard and parking incidents where the dock played no role; duplicate representations of one event.
Deduplication:
At event level, never on Report ID alone. Matching event date plus Report ID triggers manual review; confirmation requires agreement on the abstract, time, victim details and linked inspections.
Every excluded record stays in a published exclusion log. Every merged pair stays in a published duplicate crosswalk. Both ship with version 1.0 alongside the record-level CSV, a field codebook and a file checksum.
Limitations of this data
We would rather you know these than discover them after quoting us.
- This page does not contain a national loading dock accident count, and this audit found no source that does. Version 0.1 publishes verified findings and a claim audit. The reviewed incident database publishes as version 1.0.
- The 897 figure is a raw keyword match count — neither an upper nor a lower bound, for the reasons set out above.
- OSHA investigation records are not a census. They are weighted toward fatalities and catastrophes and substantially undercount nonfatal dock injuries.
- In this result set, OSHA’s most recent event was about eighteen months old. That gap is an observation about this query on this date, not a measured lag for OSHA’s whole database.
- The NIOSH forklift study covers 1980 through 1994 and nothing since. It predates the 2017 dockboard run-off requirement, modern vehicle restraint systems and the entire e-commerce fulfillment buildout.
- The NIOSH sub-counts do not reconcile. The narrative’s four sub-categories total 41 against a parent category of 39, and the paper does not explain it. We report both numbers rather than picking one.
- BLS case-characteristic data breaks in series at reference year 2023. Figures on either side of that boundary are not directly comparable.
- BLS SOII figures are sample-based estimates, subject to sampling error and rounding, and cannot be summed with OSHA record counts.
- We have not reproduced the BLS 6692 annual series. Three circulated BLS-attributed figures remain unreproduced for that reason.
- We have not opened the OIICS 3.03 workbook, which is why the current dock code is absent from the crosswalk rather than guessed.
- We could not independently retrieve OSHA record 173669.015, for the reason described above.
- Percentage shares calculated from the NIOSH tables are ours, not the study’s. The counts are the study’s; the arithmetic is reproducible from the tables as printed.
Frequently asked questions
- How many loading dock accidents happen each year in the United States?
- No federal agency publishes that number, and this audit found no federal publication reporting a national total. OSHA records investigations of selected incidents, BLS estimates nonfatal injuries by source category and counts fatalities separately, and none of the three produces a national loading dock total. A credible national figure would need to identify its dataset, its denominator and its extrapolation method.
- What percentage of industrial accidents happen at loading docks?
- The widely circulated answer is 25%. We could not trace it to any primary source, and OSHA’s own warehousing page does not make the claim. We also found no published national count of “industrial accidents” to serve as the denominator, so the percentage cannot be calculated from public data.
- Is it true that 7% of forklift accidents happen when a forklift falls off a loading dock?
- No. The study behind that claim, Collins et al. (1999) using NIOSH surveillance data, found 39 deaths from forklifts falling off a ramp or loading dock among 1,021 forklift fatalities from 1980 to 1994 — 3.8%. Loading docks specifically account for 23, or 2.3%. The claim also swaps “fatalities” for “accidents,” which are very different populations.
- What is the most common cause of forklift deaths?
- In the NIOSH study period, overturns, with 222 of 1,021 deaths (21.7%), followed by pedestrians struck by forklifts at 208 (20.4%). Falls off a ramp or loading dock ranked seventh at 39 (3.8%). Those figures describe 1980 through 1994 and should not be presented as current.
- How many loading dock incident records does OSHA have?
- OSHA’s accident investigation search returned 897 summaries containing the exact phrase “Loading Dock” on . That is a keyword match count, not an incident count — it includes at least one non-traumatic medical event and can represent a single fatality more than once.
- What is OIICS code 6692?
- It is the BLS injury source code for “Ramps, loading docks, dock plates,” used for reference years 2011 through 2022. Because it bundles ramps with docks, no BLS figure from those years isolates loading docks alone.
- Why did the BLS loading dock category change after 2022?
- BLS implemented OIICS version 3.0 with reference year 2023, restructuring the classification system. Division 6, which contains dock codes, was revised and ground-surface codes were simplified. BLS designates 2023 a break in series, so figures on either side of that boundary are not comparable.
- Does OSHA have a loading dock standard?
- Not a single one. The most dock-specific provision is 29 CFR 1910.26, “Dockboards,” covering dock plates, bridge plates and dock levelers. It requires load capacity, run-off protection for dockboards entering service on or after January 17, 2017, securing of portable dockboards, measures such as wheel chocks or sand shoes to prevent the transport vehicle from moving while employees are on the dockboard, and handholds on portable dockboards. Dock work is also reached by 1910.22, 1910.176 and 1910.178.
- Does OSHA require wheel chocks at loading docks?
- 29 CFR 1910.26(d) requires measures such as wheel chocks or sand shoes to prevent the transport vehicle — truck, semi-trailer, trailer or rail car — from moving while employees are on a dockboard. Requirements for powered industrial truck operations appear separately at 1910.178. For obligations at a specific facility, consult a qualified safety professional or your OSHA area office.
- What is trailer creep?
- Trailer creep is unintended gradual movement of a trailer away from the dock during loading or unloading, which can open a gap between trailer and dock. In the NIOSH study, three of 1,021 forklift fatalities involved a tractor-trailer rolling or pulling forward while the forklift fell between the dock and the trailer; the paper does not classify all three as trailer creep specifically.
- Can I download this data?
- The verified findings and the claim audit are published on this page as tables, and the claim audit ships as a CSV with the version 0.1 release. The record-level incident database, exclusion log and duplicate crosswalk publish with version 1.0.
How to cite this page
Default:
Indianapolis Dock Door Repair Research. “Loading Dock Accident Statistics: Fatalities, Falls, Crush Events and Forklift Incidents.” Version 0.1, last updated . https://indianapolisdockdoorrepair.com/research/loading-dock-accident-statistics/
APA:
Indianapolis Dock Door Repair Research. (2026). Loading dock accident statistics: Fatalities, falls, crush events and forklift incidents (Version 0.1). https://indianapolisdockdoorrepair.com/research/loading-dock-accident-statistics/
Chicago:
Indianapolis Dock Door Repair Research. “Loading Dock Accident Statistics: Fatalities, Falls, Crush Events and Forklift Incidents.” Version 0.1. Last modified . https://indianapolisdockdoorrepair.com/research/loading-dock-accident-statistics/
Each table has a plain-text source line beneath it. The claim-audit table additionally links each verified claim to the primary document we opened. Section anchors are stable, so individual evidence blocks can be referenced directly.
Collins et al. (1999) is expressly identified as a U.S. Government work in the public domain in the United States. The compilation, claim audit and calculated shares on this page are licensed CC BY 4.0. Third-party material, if any, retains its own rights.
Primary sources used in version 0.1
OSHA
- Accident Investigation Search (IMIS) — https://www.osha.gov/ords/imis/AccidentSearch.search
- 29 CFR 1910.26, Dockboards — https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.26
- 29 CFR Part 1910 Subpart D, Walking-Working Surfaces — https://www.ecfr.gov/current/title-29/subtitle-B/chapter-XVII/part-1910/subpart-D
- Warehousing topic page — https://www.osha.gov/warehousing
- National Emphasis Program on Warehousing, CPL 03-00-026 — https://www.osha.gov/enforcement/directives/cpl-03-00-026
Bureau of Labor Statistics
- Occupational Injury and Illness Classification Manual — https://www.bls.gov/iif/definitions/occupational-injuries-and-illnesses-classification-manual.htm
- OIICS 2.01, Section 3.3, Source of Injury code titles — https://www.bls.gov/iif/definitions/oiics-manual-2010-section-3-3.pdf
- OIICS 3.0 summary of major changes — https://www.bls.gov/iif/definitions/oiics-version-3-major-changes.htm
- SOII Table R57, 2018 — https://www.bls.gov/iif/nonfatal-injuries-and-illnesses-tables/soii-case-and-demographic-characteristics-historical-data/2018/case-and-demographic-characteristics-table-r57-2018.htm
NIOSH
- Collins JW, Landen DD, Kisner SM, Johnston JJ, Chin SF, Kennedy RD. “Fatal Occupational Injuries Associated With Forklifts, United States, 1980–1994.” American Journal of Industrial Medicine 36:504–512 (1999) — https://stacks.cdc.gov/view/cdc/196724/cdc_196724_DS1.pdf
Official data systems not yet incorporated
These are relevant and planned for version 1.0. No finding on this page draws from them.
- OSHA Severe Injury Reports — https://www.osha.gov/severeinjury
- OSHA Fatality Inspection Data — https://www.osha.gov/fatalities
- BLS Census of Fatal Occupational Injuries — https://www.bls.gov/iif/fatal-injuries-tables.htm
Version history
| Version | Date | Changes |
|---|---|---|
| 0.1 | Initial publication. Verified OSHA query count, OIICS 6692 classification, NIOSH forklift fatality breakdown, complete 29 CFR 1910.26 provisions and section history, and Warehousing NEP inspection date. Published fourteen-claim audit with verdict definitions. Reviewed incident database not yet included. |
Planned for version 1.0: the deduplicated OSHA incident corpus with record-level CSV, exclusion log, duplicate crosswalk and codebook; the reproduced BLS 6692 annual series; the confirmed OIICS 3.x dock source code.
Review schedule. OSHA query counts and the claim audit are re-run quarterly. BLS series are re-checked on each annual release. Classification codes are re-checked on each OIICS revision. The last-verified date at the top of this page changes only when we have actually re-verified.
About this research
This page is published by Indianapolis Dock Door Repair, a commercial loading dock equipment service company. We have an obvious business interest in the subject of loading dock safety, and you should weigh that when reading anything we publish about it.
That is precisely why the page is built this way. Every query is printed. Every verified source links to the issuing agency. Every calculation we performed is reproducible from figures shown on the page. Where we could not verify something, we left it unverified rather than filling the gap with a number that sounded right. You do not have to trust us — you can check all of it, and if you find an error we will correct it and log the correction.
This is educational research, not safety, legal or compliance advice. For questions about obligations at a specific facility, consult a qualified safety professional or your OSHA area office.