CQC notification mistakes that get providers downgraded

The Regulation 18 categories are precise. A service-user death from a chronic condition where nothing went wrong is a different category from a death where the cause is unclear or where an incident contributed. An allegation of abuse is a different category from a confirmed safeguarding finding. A police-involved incident is a different category from a serious injury without police involvement. Filing under the wrong category means the regulator looks at the notification with the wrong triage path.

The honest cause: most providers I saw using a single generic notification template, mapping anything notifiable to it, and selecting the category from a dropdown without checking the regulation text. Reg 18 categories are short and worth re-reading at the moment of filing rather than from memory.

The change: every notification should be filed with the specific regulation paragraph that triggers it cited in the internal record. Not the category name from a dropdown; the regulation reference. That habit forces a check at the right moment.

Most Reg 18 categories say "without delay" rather than specifying a fixed period. "Without delay" is a judgment standard, and judgment standards need to be defensible against the facts. A notification filed 11 days after an incident is not automatically late; it might be exactly right for the circumstances. But if the file contains no record of why those 11 days were reasonable, the inspector reads it as a service that does not know it has a duty.

The honest cause: services tracking the incident itself but not tracking the notification timing as a separate data point with its own reasoning record.

The change: each notification record should carry a short note about when the duty became engaged (when did the team know enough to know this was notifiable) and a short note about why the filing time was reasonable against the facts. Both fit in two sentences. Both stop the inspector from asking.

The biggest single signal an inspector reads in a notification cluster is whether the narrative submitted to CQC matches what the internal incident record says happened. Tiny discrepancies (a time off by an hour, a staff member named in one record and not the other, an outcome described slightly differently) read as either careless or evasive. The inspector cannot tell which from the records alone.

The honest cause: notifications written from memory days after the incident, rather than constructed from the structured incident record at the moment of filing.

The change: the notification narrative should be assembled from the incident record fields directly, not retyped. If the incident record says a thing, the notification should cite the incident record as its source. If the notification needs to say a thing the incident record does not say, that is the moment to fix the incident record, not to write conflicting prose.

Less common than the first three but more damaging when it happens. An incident occurs, the team handles it, an obvious notifiable trigger is present (the death of a service user, a police involvement, an allegation of abuse) and the notification is never filed. The inspector reads the incident record and asks where the matching notification is. If the answer is "we did not file one", the conversation gets difficult fast.

The honest cause: no automated check linking incident categories to notification triggers. The team relies on the registered manager remembering to file. The manager does, almost always; the missing cases are usually during high-pressure weeks when several things happened at once.

The change: each incident record should produce a mandatory notification triage step, with a notifiable / not-notifiable decision recorded against it, even when the answer is "not notifiable". The recorded non-notification decision is as important as the recorded notification, because it shows the duty was considered.

Regulation 18 (notification to CQC) and Regulation 20 (duty of candour to the patient) are different duties with different audiences. Most events that engage Reg 20 also engage Reg 18, but they require separate records and separate fulfilment. Services that treat them as one duty often fulfil Reg 18 (the CQC notification) and forget Reg 20 (the duty to the patient with the specific content elements). The inspector finds the Reg 18 record, asks for the Reg 20 record, and discovers the latter does not exist.

The honest cause: a single "duty of candour completed" checkbox in the incident template that ticks when the CQC notification is filed.

The change: explicit separation of the two duties in the recording template. The Reg 20 record should have its own field for the verbal stage (when, who, what was said) and its own field for the written stage (date, content elements delivered). The full distinction is in the Duty of Candour article.

All five mistakes have the same underlying cause: notifications are being treated as a paperwork output rather than as a structured duty with its own audit trail. The fix is the same across all five: every notification has a record that captures the triggering event, the category cited, the timing reasoning, the narrative matched to the incident record, and the relationship to Reg 20 where it applies.

None of this is hard. Most services I worked with had the substance right; they just did not have the structured recording trail around the substance. An inspector reads the trail. Without it, they read absence.

Care Quality Commission (Registration) Regulations 2009, Regulation 18, on legislation.gov.uk. It is short. Read it at the start of every quarter.