When a discharge letter arrives but the message does not

(1) When a small hospital replaced its paper discharge summaries with an electronic system, the first dashboard appeared reassuring. Almost every summary was sent before the patient left. The improvement was real: staff no longer searched for a working fax machine or waited for a porter to deliver a folder. Yet community clinicians continued to telephone for information that the hospital believed it had already supplied. The project team initially treated these calls as evidence that colleagues needed help finding documents. A closer examination suggested that finding the document and finding a usable plan within it were different problems.

(2) The electronic template had encouraged completeness by giving each department its own section. That arrangement made omissions easier to detect during an internal check, but it also reproduced the hospital's organisational boundaries for the reader. A change in treatment might be described in one section, its reason in another and the required follow-up elsewhere. The community practitioner had to reconstruct the relationship between them. Consultant Leena Shah described the summary as a filing cabinet that had been posted to somebody who needed a letter. Her criticism concerned the work imposed on the recipient, not the accuracy of each individual entry.

(3) One response was to impose a word limit. The team hoped that shorter summaries would force staff to prioritise. In a small trial, however, some writers removed the reasons for changes while retaining long lists of investigations. Others shortened sentences into strings of abbreviations. The documents became smaller without necessarily becoming easier to act on. Shah did not reject concise writing. She argued that conciseness should be judged by how efficiently a reader could understand the relevant plan, rather than by treating every deleted word as an improvement. A short unexplained instruction could create more work than a slightly longer explanation.

(4) A revised template began with three prompts: what has changed, what needs to happen next and who has accepted responsibility? Hospital teams were uneasy about the last question. They could record a recommendation for community follow-up but could not assume that a named practitioner had accepted it. The template therefore distinguished an action requested from an action agreed. This distinction sometimes made the summary look less complete, since an unresolved handover remained visible. The project lead considered that discomfort useful. A blank space disguised as a completed arrangement was more dangerous to communication than an explicitly unfinished task.

(5) The hospital then invited community staff to annotate anonymised examples. Their comments were more specific than a request for less jargon. One nurse understood every technical term but could not tell whether a monitoring appointment had already been booked. Another could identify the correct medicine list but not why an earlier treatment had been stopped. These examples challenged the assumption that specialist vocabulary was the main barrier. Plain wording helped, but it could not repair a missing link between an observation and an action. The team began reviewing letters for those links as well as for factual accuracy.

(6) Early results were encouraging but difficult to interpret. Telephone queries fell after the revised template was introduced. During the same period, a liaison nurse began answering questions through a separate messaging service. Counting fewer telephone calls therefore did not establish that fewer clarifications were needed. Some hospital managers wanted to present the fall as proof of success. The evaluation team instead examined a sample of messages and asked receiving clinicians to explain the intended plan from selected summaries. This took longer than extracting a dashboard count, but it measured something closer to the problem the project was trying to solve.

(7) Recipients did not agree on a single ideal level of detail. A clinician seeing a patient for the first time needed background that a longstanding practitioner might already know. A standard core could make urgent information easier to locate, but it could not anticipate every reader's knowledge. The team retained optional background sections and asked writers to select them according to the recipient and purpose. Critics worried that this reintroduced judgement into a process the template was meant to standardise. Shah replied that removing avoidable inconsistency was worthwhile; removing the need to consider the reader was not a realistic objective.

(8) The project did not produce a universally perfect discharge letter. It produced a more demanding definition of completion. A document could be accurate, prompt and successfully transmitted while still leaving the next professional uncertain about what to do. Checking those earlier stages remained necessary, but none could stand in for checking the final one. The lesson was not that electronic systems had failed. They had made delivery easier and some omissions more visible. Their value depended on whether the organisation used that improvement to support communication, or allowed a successful transmission to become the last question it asked.