Risk and scoring

How the score is computed, risk bands, and clinical review

Each questionnaire can produce a score that summarizes the patient’s answers and classifies them into a risk band.

How the score is computed

The score is deterministic: it sums the weight of each selected option across multiple-choice questions. Other question types (text, numeric, boolean, date) do not contribute to the score. An option with no weight contributes 0; unanswered questions contribute 0.

Risk bands (score_class)

Bands are defined by each protocol: a name, a [min, max] range, and, optionally, a cadence adjustment (freq_override) for the next call. There is no fixed set of bands — each protocol defines its own. On publish, the bands must be contiguous, non-overlapping, start at 0, and cover the whole possible score range.

If a protocol defines no bands, the questionnaire stores the score without classifying it (score-only mode).

Risk and trend

In population views, bands are ordered from best to worst control to color risk, and you can see the score trend over time and transitions between bands. The run.previous_score variable exposes the previous call’s score for use within the flow.

Clinical review

Each questionnaire can be marked as reviewed, recording who and when, along with an optional clinical note. Lists can be filtered by reviewed / not reviewed. Clinical review of a questionnaire is distinct from acknowledging an alert.