Start with the question
“Please review these records” leaves the task open. A more useful opening identifies the problem and the decision: “Breathlessness has progressed over six weeks. The current question is how the pressure measurements should affect a proposed procedure.” The reader now knows why the history matters.
The summary need not contain a diagnosis. If one has already been proposed, identify it as the working diagnosis and say who proposed it. That keeps the first interpretation from being mistaken for an established fact throughout the rest of the narrative.
Build a chronology before shortening it
Make a private working timeline with a date or approximate period, a change and its supporting record. Include the onset of the main symptoms, important investigations, treatment starts or withdrawals and subsequent changes. The CARE framework for published case reports similarly emphasises patient information, a timeline, diagnostic assessment and follow-up. Here we adapt the idea for preparing a discussion, rather than claiming to create a formal case report. [1]
| When | Observation or event | Where it came from |
|---|---|---|
| Before the current illness | Previous function and relevant baseline results | Earlier records or the person's account |
| At onset | First clear change and its pace | A dated note or a clearly labelled recollection |
| At testing | Exact result, units and relevant conditions | The laboratory or imaging report |
| After treatment changed | What improved, worsened or remained unchanged | Follow-up measurements and symptoms |
| Now | Current limitations and the unresolved decision | Recent assessment |
Once the timeline is coherent, shorten repeated descriptions. Keep any detail that changes the interpretation of an important result.
Preserve the details that change meaning
An isolated number often loses its meaning when copied without units, reference interval, specimen or date. Include those when available. If a test was performed after treatment began, say so. If two results used different laboratories or methods, preserve their names instead of treating them as repeated measurements of exactly the same thing.
The same principle applies to symptoms. “Weakness” may refer to reduced measured power, fatigue, poor balance or pain-limited movement. Describe what the person could previously do and what they can do now. Keep a clinician's examination finding separate from an everyday observation.
Separate facts from conclusions
Use three clear categories in your working notes: reported symptom, measured finding and interpretation. For example, “the report called the result reassuring” is an interpretation. The original test value and method remain the underlying finding.
A summary is more useful when it preserves both. Removing all earlier opinions loses context; copying an opinion as a fact makes it harder to reassess the case. This distinction is especially important when an AI system is being asked to review an existing explanation.
Record treatment as a sequence
List the treatment, when it started, what it was intended to address and what happened afterwards. Describe symptom changes and laboratory changes separately. If several treatments started together, preserve that fact. It prevents a later reader from attributing the whole change to a single intervention without a basis for doing so.
Record a treatment response in observable terms where possible. “Could walk to the kitchen again” and “a marker returned to its reference interval” are different outcomes. Both may matter; neither should silently replace the other.
End with a short question list
Choose a small number of questions that would change the next discussion. Which finding most strongly supports the current explanation? What does the normal test leave unresolved? Which additional piece of information would distinguish the leading possibilities? What is the follow-up plan?
Healthdirect Australia's appointment resources provide another practical way to prepare questions for a clinician. They are useful companions to a case summary, particularly when the main need is deciding what to ask at the next appointment. [2]
Use the summary in MontMedical
MontMedical can use a clinical narrative as the starting point for an initial discussion and a referenced evidence review. A well-organised narrative makes it easier to inspect whether the answer has used the correct timeline and interpreted the actual tests supplied.
Remove direct identifiers that are unnecessary for the task, and check the current privacy information before submitting sensitive material to any service. Keep the source records so that you can verify the summary and bring them to the clinician. The summary helps organise the conversation; the records preserve the detail.
When the chronology, findings and question are ready, explore a clinical case in MontMedical. Then compare the answer with the original information rather than letting the answer become a replacement for it.