One complaint, nine records: anatomy of a closed loop

The first record is a complaint. The patient is unhappy that the dressing was missed and that, when they called over the weekend, no one with the authority to act seemed to be available. The clinic acknowledges it within its policy window, investigates, and responds. The outcome is upheld in part: the dressing schedule was followed, but the weekend escalation route did not work as it should have. An apology is recorded, and the outcome is communicated back to the patient in writing.

A service that stops here has done something reasonable. It has answered the complaint. But answering the complaint and fixing what the complaint revealed are two different things, and the second one is where care actually improves.

In the course of investigating the complaint, something more serious surfaces. The wound had been deteriorating for longer than anyone realised, and the weekend call had not been escalated in time. That is not a complaint about service. That is a clinical incident with the potential for harm, and it becomes the second record.

The incident is logged and linked back to the complaint that revealed it, so anyone opening either record can see the other. It is investigated properly: who was on, what the escalation path was meant to be, where it broke. The investigation does not just record what happened, it records the thinking, because the thinking is what the next decisions will rest on. The harm is assessed as moderate, which raises the stakes for everything that follows.

Once the incident is understood, two further records open almost at once. The severity and circumstances mean the event meets the threshold for a statutory notification to CQC, which is submitted within the required window and linked to the incident, so the source of the notification is never in doubt. We cover the categories and the process in our guide to CQC notification forms.

The same facts also raise a safeguarding question, so a safeguarding concern is opened as the fourth record. The threshold is assessed with the reasoning written down, not just the conclusion, the local authority is informed, and the concern is worked through to a close. It too is linked to the incident. By this point, one weekend phone call has produced four connected records, and every one of them knows where it came from.

The investigation does not end in a finding. It ends in work. Three improvement actions are raised, each with an owner and a due date, and each linked back to the incident it came from.

The first is a refresher on the escalation protocol for the clinical team, and it links across to the training records of the two staff who needed it, so completion is evidence, not a claim. The second is a review of the post-operative observation policy, the document whose gaps the incident exposed. The third is a change to weekend senior cover, the structural problem underneath the whole event. Crucially, all three are delivered by their due dates. A finding that turns into three open actions nobody finishes is a service that noticed a problem and left it in place. A finding that turns into three delivered actions is a service that fixed it.

The policy review produces the seventh record: a new version of the post-operative observation policy. The old version is retired, a next-review date is set, and the new version is linked back to the action that prompted it. This is a small thing that matters more than it looks. A great many services decide to update a policy and never quite do, or update it without retiring the old one, so two versions circulate. Here, the change is real, it is dated, and it is traceable to the event that justified it. The reason the policy is what it is can be read straight off the record.

Step back from the single event and a pattern is visible: weekend escalation is a weak point. That is not a one-off to be closed, it is a standing risk to be managed, so it becomes the eighth record on the risk register. It is scored, and its controls point straight back at the actions already raised, the training refresher and the new weekend cover. A review is logged showing the controls are holding. The risk is linked to both the incident and the original complaint, closing the loop back to where it started.

This is the move that separates a service that reacts from one that learns. The same problem will not be rediscovered from scratch the next time it surfaces, because it is already named, owned and controlled.

The ninth record is the governance meeting where all of this is reviewed. The minutes do not say that incidents were discussed. They link to the specific complaint, the incident, the notification, the safeguarding concern, the three actions and the risk, and they record a single decision: the weekend staffing change is made permanent. That decision spawns its own small piece of follow-up, and the loop is closed, from a Saturday phone call to a structural change in how the service runs, with every step linked to the last.

It is worth naming where a chain like this most often falls apart, because none of the failures are dramatic.

The first is that the complaint is answered and closed without anyone spotting the incident inside it. The service resolves the customer-service problem and never sees the clinical one. The second is that the actions are raised but not delivered, so the finding is real and the fix never reaches anyone. The third is the quietest: every record exists, but none of them are linked, so the service is managing nine separate things instead of one connected one, and the pattern that should have become a risk never does.

Read as a chain, these nine records tell a single coherent story: a problem was raised, understood, escalated where it had to be, acted on, built into how the service runs, and reviewed at the top. Open any one of the nine and you can walk to all the rest. That is what a service understanding itself looks like, and it is the same thing a regulator reads as a learning culture, the same issue showing up in more than one place, connected by a risk and an action, proving the service saw the whole picture. But the patient on that Saturday does not benefit from the inspection. They benefit from the weekend cover that changed because of them.

This is the evidence loop in full, the idea we set out in how the evidence loop works in practice, worked through a single real-shaped case. One complaint. Nine records. A service that is measurably better at the end than it was at the start.

Verivius is built to make this chain happen as a side effect of the work, with each record linked to the next as you go. See how it works for your sector, or talk to us about a design-partner engagement.