The Patient Story Replaces the Chart Page

The chart page is organized around where data came from. A patient view should be organized around the question being asked about this person right now.

A conventional patient page is a set of tabs named after source systems: vitals, notes, messages, claims, time. The reader assembles the story. Every reader assembles a slightly different one, and none of them is recorded.

The alternative generates the narrative against the current question. Escalating for widening variability produces a different opening paragraph from reviewing month-end evidence, over the same patient and the same records. Underneath, timeline lanes separate what the device reported, what the patient reported, what staff did and what the system did, so the provenance of each event is visible without leaving the story.

Two details carry most of the weight. Missing days are drawn as gaps rather than as flat lines, because a missing reading is not a normal reading. And the page reasons explicitly about care-program fit: what the program requires, its typical operating burden, the observed effort for this patient, and whether the two are consistent.

That clinicians are trained on the chart and that a generated narrative is a documentation risk; that the legal record must be reproducible exactly; and that a story built for a question may omit the thing that mattered for a different question.

The reproducibility objection stands unless the narrative is versioned, grounded and stored alongside the question it answered. A story you cannot regenerate identically is not a record.

Comprehension testing under time pressure: give matched clinicians the same underlying record through both interfaces and measure what each group correctly reports about the patient. Alongside it, a grounding audit, since a narrative with any ungrounded sentence fails regardless of how well it reads.

Whether the question a narrative was generated for should be visible on its face. How to show what the narrative deliberately left out. Whether program-fit reasoning that names revenue alongside clinical need can be presented to clinical users without distorting clinical judgment.