RIDDOR: what is actually reportable, and by when

Who has to report

The duty falls on the responsible person — in practice the employer, or whoever is in control of the premises where the work is done. It is not a duty on the individual who was hurt, and it is not a duty on their supervisor personally. In a university or institute setting it almost always sits with a central safety function, which is why the practical question for a lab is usually "who do I tell, and how fast" rather than "do I file this myself".

That has a consequence worth designing for: your internal record has to reach the person who holds the duty quickly enough for them to meet the deadline. A near miss written on a whiteboard does not.

The categories, in plain English

RIDDOR groups reportable events into a small number of buckets:

• Deaths arising from a work activity.
• A defined list of specified injuries — including most fractures other than to fingers, thumbs and toes, amputations, permanent or likely-permanent sight loss, crush injuries to the head or torso, serious burns, scalpings, loss of consciousness from a head injury or asphyxia, and injuries from work in an enclosed space leading to hypothermia, heat illness, resuscitation or an overnight hospital admission.
• Injuries to a worker that leave them unable to do their normal work for more than seven consecutive days, not counting the day of the accident.
• A defined list of occupational diseases, reportable when a doctor's written diagnosis links them to the work.
• A defined list of dangerous occurrences — near-miss events serious enough to be reportable in their own right.
• Accidents to people who are not at work — visitors, students in some circumstances, members of the public — where the person is taken from the scene to hospital for treatment.

The specified-injury and dangerous-occurrence lists are exactly that: lists, published by the regulator. Read the current list rather than working from a summary — including this one.

The deadlines that catch people out

Three timescales matter, and they are different from each other:

• A death or a specified injury must be reported without delay, and a report submitted within ten days of the incident.
• A dangerous occurrence must be reported within ten days.
• An over-seven-day injury has a longer window — fifteen days from the accident — but the clock starts on the day of the accident, not on the day someone realises the absence has run past a week.
• A reportable occupational disease is reported as soon as the diagnosis is received in writing.

The shortest of these is ten days, which is why it is the sensible internal deadline to work to while the category is still being decided.

What is NOT reportable — and why you still record it

A near miss is not reportable unless it is one of the listed dangerous occurrences. A splash caught by goggles, a dropped flask, a cabinet alarm, a cut that needed a plaster — none of these normally reach RIDDOR.

They still belong in a register, for three reasons that have nothing to do with the regulator. First, accident records have to be kept: an employer keeping an accident book or equivalent is standard practice, and RIDDOR records themselves must be retained for at least three years. Second, the pattern is the point — five near misses on the same rig is information a single incident report can never carry. Third, if something serious does happen later, the register is what shows whether it was foreseeable and what was done about the warnings.

The biological and chemical cases labs ask about

Two situations come up repeatedly in research settings. An accidental release or escape of a biological agent likely to cause severe human illness is a listed dangerous occurrence. And an unintended release of a substance that could cause personal injury to any person can fall within the dangerous-occurrence list depending on the circumstances.

Neither is a judgement to make from a blog post. Both are exactly the kind of event where the incident record needs to reach your safety office the same day, with the description written while people still remember it accurately.

Why the report should be un-editable

There is a practical reason to make the initial account of an incident permanent, and it is not distrust. Memory reshapes itself, and it reshapes itself towards whatever explanation the group settled on afterwards. An account written within the hour, before anyone has decided what happened, is worth more than a tidier one written a fortnight later.

The way to get both is to separate them: fix the report — what happened, when, where, who, what was done immediately — and leave the investigation that follows fully open to being revised, extended and corrected. A correction to the report itself becomes a new record rather than an overwrite, so the change is visible.

A workable internal process

The reporting decision is not one a lab makes alone, but the process around it is entirely within a lab's control:

• Anyone can file, including about a near miss, without approval and without a form to hunt for.
• The account is captured immediately and then left alone.
• Someone is named as investigating it, with a target date, and that date is chased if it passes.
• The reportability judgement is recorded as a judgement — with who made it and when — rather than being implied by silence.
• Once reported, the regulator's reference is stored against the record, so the question "did we actually file that one?" has an answer.