What an incident report cannot tell us on its own

(1) When a ward introduced a simpler incident-reporting form, the number of reports rose. A local newsletter described the increase as a deterioration in safety. Staff who had helped design the form were dismayed: they had wanted to make previously invisible problems easier to record. Neither interpretation could be settled by the count alone. More reports might reflect more incidents, more willingness to report, or both. Quality lead Mara Ellis argued that the first responsibility was to understand the change before turning it into either a success story or an accusation. A visible number was not automatically a self-explanatory one.

(2) The old form had asked staff to choose a single cause before describing what happened. Several interviewees said this made them reluctant to begin when responsibility was uncertain. The revised version opened with a short account of the event and allowed contributing factors to be added later. It did not suggest that causes were unimportant. It recognised that the person noticing a problem might not yet have enough information to identify them. Ellis wanted the system to preserve uncertainty long enough for investigation, rather than reward the first plausible explanation simply because a box required an answer.

(3) The team also invited reports of situations corrected before a patient was affected. A sample involved a specimen label checked and replaced before dispatch. One manager asked whether recording such events would distract from serious harm. Ellis accepted that reports needed proportionate review, but argued that a recovered error could reveal how a process failed and what prevented the failure from continuing. The successful check was part of the event, not a reason to pretend that nothing had happened. Treating every near miss as trivial would remove information about the barriers on which the service depended.

(4) As reports accumulated, a familiar phrase appeared repeatedly: staff should be more careful. It seemed reasonable and was difficult to dispute, but it gave little guidance about what to change. In one case, two containers with similar labels were routinely stored together under poor lighting. Reminding staff to read labels remained sensible; separating the containers and improving visibility addressed conditions that made confusion more likely. The team was not claiming that individual attention no longer mattered. It was asking whether the proposed action would make the next person's task easier to perform reliably, rather than simply restating the standard expected.

(5) A second difficulty was the delay between reporting and feedback. Staff sometimes heard nothing after submitting a detailed account. The investigation might have continued, but silence made it hard to distinguish careful review from neglect. The quality team began sending a brief acknowledgement explaining what would happen next, followed by an update when an action had been chosen. It avoided promising that every report would lead to a new policy. Some reports identified an existing measure that was working; others required further evidence. Feedback was intended to make the process visible, not to manufacture a dramatic response to every submission.

(6) After three months, staff surveys suggested that reporting felt easier. The total number of reports remained higher than before, while a separate review of selected records found fewer examples of one recurring labelling problem. These findings were compatible with improvement, but they answered different questions. The survey concerned experience of reporting; the record review concerned a particular failure. Neither established the ward's overall safety level. Ellis resisted combining them into a single celebratory score. She preferred to state the limited findings clearly, including the possibility that other problems had not been detected by the review.

(7) The most difficult conversations involved events in which an individual action had clearly departed from an agreed process. Some staff feared that attention to system conditions meant avoiding accountability. Ellis distinguished understanding an action from excusing it. A fair review could examine what someone did, what information they had and what conditions shaped the decision without deciding in advance that either the person or the system must carry the entire explanation. That approach required more work than assigning a category. It also offered a better chance of choosing an action relevant to the circumstances actually found.

(8) The ward continued to count reports, but stopped using the total as a direct measure of how safe it was. Counts helped identify patterns and allocate review effort; investigation and additional evidence were needed to interpret them. The reporting system was useful when it opened a question that the organisation was prepared to pursue. It was less useful when completing the form became the final action or when a rise in reports was automatically punished. A service that wanted to hear about problems had to show that hearing was followed by careful attention, and that careful attention was followed by an appropriate response.