When a Company Deletes Its Own Safety Reports: Litigation Holds and the Evidence That Survives
Regulators require companies to keep records of what went wrong. Incident reports, near-miss logs, internal audits, maintenance histories. Those documents exist because a rule somewhere says they must. The same rules tell you how long the retention clock runs, and when it is legal to destroy them. What happens when the destruction happens early, right around the time the first lawyers’ letters start flying, is one of the most consequential questions in any injury or product case.
The legal instrument that governs this is the litigation hold. The moment a company reasonably anticipates litigation, its ordinary document destruction policies are suspended. Shredding, auto-delete, file purges, all of it must stop for anything plausibly relevant. If it does not stop, and documents vanish, courts have a specific remedy: adverse inference instructions, where the judge tells the jury it may assume the destroyed evidence would have been unfavorable to the company. In practice, that instruction can end a case.
The pattern is common enough to recognize. An incident happens. Weeks later, well before any suit is filed, a scheduled data purge deletes exactly the log covering that day. The maintenance contractor’s copy of the inspection report disappears from the shared drive. The CCTV overwrite cycle, normally 30 days, somehow drops to 7 days right after a written complaint arrives. Each of these can be innocent. Together, in sequence, they form the classic spoliation fact pattern.
What most people do not know is how much of the deleted material actually survives elsewhere, if you know to look. Safety reports are rarely single-copy. The insurance carrier that was notified keeps its own claim file, and those files are subject to insurance retention rules that do not answer to the company’s convenience. Regulators receiving mandatory incident reports, OSHA in the United States, the HSE in the United Kingdom, sector equivalents elsewhere, hold their own copies, and much of what they hold is requestable under freedom of information or disclosure rules. Contractors who performed repairs keep work orders and invoices. Equipment now carries event logs that nobody at head office remembered, Building management systems, elevator controllers, industrial PLCs, vehicle telematics boxes all timestamp events locally.
The discovery tools that reach this material are broader than most claimants expect. Interrogatories can force a company to state, under oath, what documents existed, who had access, and when each was deleted. Requests for admission can lock the company into a deletion timeline before it realizes why the timeline matters. Third-party subpoenas reach the insurer, the contractor and the regulator directly, none of whom have any incentive to protect the defendant. And forensic examination of equipment logs often produces records the company genuinely believed were gone, because the person who deleted the file deleted a shortcut, not the controller’s internal history.
Timing is everything in a spoliation fight. The duty to preserve attaches on reasonable anticipation of litigation, which courts have found in a formal demand letter, in some cases in a credible threat, and in regulator involvement. That is why sending a written notice that identifies the incident, states an intention to pursue claims, and expressly demands preservation of named document categories is worth doing early, even before you are sure you will sue. The letter does two jobs. It starts the clock, and it makes any later destruction impossible to defend as routine housekeeping.
When you suspect destruction rather than routine expiry, the first move is to freeze the timeline before anyone can revise it. Pull the company’s published retention schedules, many are public in regulatory filings or industry standards, and mark where the deleted material fell short of them. Save the auto-delete policy screenshots and IT configuration exports early, because those too can be edited after the fact. In the UK, a pre-action disclosure application under the Civil Procedure Rules can compel production of specific documents before proceedings even begin, which is often enough to surface whether the file is thin because it was thinned.
The sanctions curve runs from cost-shifting, where the destroyer pays the other side’s forensic costs, through exclusion of the company’s favorable evidence, to the adverse inference instruction and outright dismissal of the company’s own claims or entry of judgment. What moves a case up that curve is proof of intent, and intent is almost never confessed. It is inferred from convergence: a purge run out of schedule, on the specific relevant files, by a person who had just been consulted about the incident, with no explanation that survives scrutiny.
For the individual raising these questions, the practical order is simple. Document the incident yourself immediately, photos, names, dates, correspondence. Send the preservation letter naming categories, not just documents: all inspection and maintenance records for the equipment, all incident and near-miss reports, all communications internal and external concerning the event, all automated logs from involved systems. Then pursue the third-party copies in parallel, insurer, regulator, contractors, while pressing the company on its deletion timeline in writing at every step.
Deleted does not mean gone, and destroyed in breach of a duty to preserve often means worse than gone, it means the gap itself becomes evidence against whoever made it.
