A discharge summary is not a narrative — it is a structured data extraction exercise that happens to read like prose. Every time a clinician produces one from memory or from scanned ward notes, the risk of omission compounds. The new summarizer template imposes structure on that extraction without sacrificing the clinical judgment required to write it.
The Cognitive Weight of the Ward Round
By the time a clinician sits down to produce a patient summary, they have carried the full clinical picture in working memory through rounds, procedures, family consultations, and results review. The handover document is the last task, assembled from a mental state that has been degrading since morning.
The principle dx and secondary dx fields using ft_lookup types are not merely text fields — they constrain diagnosis entry to a reference set, eliminating the free-text variation that turns "community-acquired pneumonia" and "CAP" and "lobar pneumonia NOS" into three different entries that downstream analysis cannot reconcile. At scale, across a ward, across a month, that disambiguation is the difference between a searchable clinical record and a pile of unstructured text.
The past history and history separation enforces a clinical distinction that matters for handover: what the patient brought in versus what developed during the admission. Conflating them in a single field is how significant new findings get buried in the background narrative of chronic conditions.
The Architecture of Calculated Time
The template's calculated fields tell the temporal story of the admission without manual arithmetic. rbs, h1 through h6, and their corresponding c variants (likely cumulative or categorical summaries), plus days, fup, ft, ftd, fupd, and fupo — these are not decorative. They are the machine-computed spine of the admission timeline.
days calculates admission length without manual subtraction. Follow-up fields — fup, fupd, fupo — track the post-discharge obligation: when it's due, what was ordered, whether it happened. dt presumably captures discharge timestamp, anchoring the entire timeline to a fixed exit point. The calc fields eliminate the arithmetic errors that propagate when a junior doctor manually counts days-to-follow-up under time pressure.
The ip and bed calculated field is the localization anchor — knowing where the patient physically was during the admission without scanning a separate admissions ledger. At handover, the incoming team needs to know where the patient is, what they have, and what comes next. This template surfaces all three from a single record.
What a Hundred Summaries Reveal
The type of summary choice field — distinguishing between admission summaries, interim summaries, discharge summaries, and potentially referral letters — means that a single patient record can hold the complete summary chain for one admission without conflation. Filtering on type across a service gives you the admission rate, the discharge rate, and the proportion of patients with documented interim review.
At volume, the structured hospital course ft_lookup field enables retrospective audit in ways that free-text never allows. When you need to identify all patients admitted for a specific diagnosis who had a documented hospital course mentioning a particular complication, the lookup field constrains the language set enough to make that search tractable.
The ongoing tt and ds tt subheaders — ongoing treatment and discharge treatment — separate active management from the patient's exit prescription. The JavaScript fields (js type) embedded alongside these subheaders handle the conditional logic: what gets displayed depends on the summary type and admission context. A clinician filling out an interim summary doesn't see the discharge fields cluttering the interface. The form surfaces what's relevant to the current summary type and nothing else.