PROMIS computerized adaptive tests

Fewer questions. Sharper answers.

Adaptive tests choose each question based on your previous answer, reaching research-grade precision in about 4–7 items instead of a fixed long form. They score on the PROMIS T-metric: US adult average 50, higher is better for function, lower is better for pain and distress.

Demonstration mode — adaptive engine not yet connected

How adaptive testing goes live here

The item-selection engine and IRT parameters for PROMIS CATs are licensed by HealthMeasures (Northwestern University) rather than published, so this page runs against a secure engine instead of scoring in the browser. Until one is connected, it demonstrates the flow with clearly labeled placeholder items and produces no score. Three routes to live, in order of institutional simplicity:

Route 1 · Recommended

Epic PROMIS CAT application

CATs delivered through MyChart and Hyperspace, auto-scored into the chart — the same architecture that sustained ~93% capture in the program's prior academic deployment. Licensed through Epic; nothing to host.

Route 2 · Practice-owned

Assessment Center API

Annual-fee licensed web service, self-hosted, keeps responses behind your firewall, response-pattern scoring. This page then talks to it through the included relay (cat-worker.js). Contact api@assessmentcenter.net.

Route 3 · No license

Fixed short forms

PROMIS short forms are free with HEAP permission and official scoring tables — not adaptive, same T-metric. The fallback while Routes 1–2 are procured.

Privacy stance is unchanged: in live mode, only anonymous item responses transit the relay — no name, no identifiers, nothing stored by this page. Results print exactly like the main portal summary.