Skip to main content

Understand cohort results

The Visualizer gives you several kinds of feedback as you build a cohort. This page explains each one.

The toolbar patient count​

The count beside the title is the size of your cohort:

  • All N patients before any selection — the whole dataset.
  • N of total patients selected once filters are active.

While a new result is being calculated the count dims, then brightens when it is ready.

The toolbar showing an active "N of total patients selected" count

Counts inside filter cards​

When no filters are active, a value shows its total patient count. Once other filters are active, a value can show two numbers:

included / total

For example 12/40 means:

  • 40 patients have that value in the broader data (the denominator), and
  • 12 of those also satisfy your current criteria from other cards (the numerator).

This previews how a value relates to your current cohort before you select it. A value whose numerator is 0 is disabled.

In the Selected Patients drawer​

The Selected Patients drawer adds more context:

  • The active-filter summary above the drawer lists your current filter selections, each with a per-filter patient count.
  • When the drawer is minimized, a plain-language summary describes the cohort — for example, "7 asian female patients with breast cancer."

Each filter matches patients independently; your cohort is the intersection of those matches. A value can match many patients on its own and still contribute to a small cohort once combined with other cards.

When no patients match​

If your criteria produce zero patients, the drawer explains why, in one of two ways:

  • One filter matched nobody. If a single filter matches 0 patients on its own, the Visualizer names that filter — for example, "Grade matched 0 patients before intersection. Check spelling and selected values." Review that filter's selected values.
  • The filters don't overlap. If every filter matches patients on its own but their intersection is empty, the Visualizer explains that the filters have no overlap and suggests broadening one — "Each filter matches patients independently, but their overlap is 0. Try broadening one filter."

To recover, remove or broaden your most recent criterion, or use Reset filters to start over.

Data availability​

A missing filter, value, or finding does not necessarily mean a clinical fact is absent. It may reflect the loaded data, the extraction configuration, service availability, or terminology normalization.

caution

Do not treat an absent Visualizer value as a confirmed negative clinical finding without checking the source record.