PROMs Collection at Scale: Hitting Sustained 40 to 50 Percent Response Rates

This is a cluster article under the CMS ACCESS Model Complete Provider Readiness Guide. The pillar covers the full readiness framework. This article goes deeper on Pillar 2: patient-reported outcome measures (PROMs) collection at scale, which ACO leadership teams consistently report as the hardest organizational change in their ACCESS readiness build.

The technology piece is roughly 30 percent of this work. The other 70 percent is operating model, clinical workflow, and patient engagement design. Most teams underestimate that ratio at the start.

Why PROMs at ACCESS scale is hard

PROMs have lived in pilot programs and specialty clinics for years. They have not lived at ACO-attributed-population scale with sustained 40 to 50 percent response rates. ACCESS expects exactly that.

Three things make scaled PROMs hard.

Patient engagement at population scale. Sending a survey is easy. Getting 40 percent of an attributed Medicare population to respond every quarter, sustained over years, is a different problem. The chronic-condition cohort that ACCESS targets is older, less digitally engaged on average, and reached through a mix of channels.

Workflow integration that does not feel like a survey. PROMs that arrive as a generic "please complete this survey" message are ignored. PROMs delivered in the context of a specific clinical interaction, with clear framing of why the patient's response matters, get filled out.

Clinical loop closure. PROMs that nobody reads do not change outcomes. Surfacing PROMs results in clinician workflow with documented escalation thresholds is what turns the data collection effort into a quality and outcome score.

Instrument selection: PROMIS-29 plus condition-specific

The instrument set for ACCESS scoring blends generic and condition-specific measures.

PROMIS-29 is the standard generic instrument. Validated across most chronic conditions, manageable length (29 questions), and broadly accepted by CMS. Most ACOs deploy PROMIS-29 as the population-wide baseline, repeated quarterly for the chronic-condition cohort.

Condition-specific instruments. PHQ-9 for depression, GAD-7 for anxiety, KCCQ-12 for heart failure, CAT for COPD, AUDIT-C for alcohol risk, and similar. Layered on top of PROMIS-29 for patients with the relevant condition. These are where the clinically actionable signal lives.

The instrument selection decision is rarely close. Stick with widely-validated, CMS-accepted instruments. Custom or modified instruments create scoring and reporting headaches that are not worth the marginal value.

Delivery channels: portal, SMS, in-EHR, kiosk

Four delivery channels cover the practical population. Most ACOs deploy three of them.

Patient portal (MyChart, FollowMyHealth, athenaPatient). Highest-engagement channel for patients already enrolled in the portal. Survey completion rates above 60 percent are routine for portal-active patients. The catch is that portal enrollment among Medicare ACO populations is typically 35 to 55 percent. So portal alone caps your reachable population at roughly half.

SMS / text. Reaches the 30 to 40 percent of patients not portal-enrolled but with a working mobile number on file. Response rates are lower (15 to 30 percent) but the absolute reach is meaningful. The SMS workflow has to use a HIPAA-aligned vendor and follow content-of-message rules to stay compliant.

In-EHR survey delivery (kiosk or post-visit). Tablet handoff during the visit, post-visit summary email link, or front-desk completion. Captures patients who would not engage outside a clinical encounter. The completion rate is high but volume is gated by visit cadence.

Phone outreach. Highest cost-per-response, highest reach for the most disengaged segment. Often run by care management staff against worklists of patients who have not responded to other channels in the previous 90 days.

The portfolio that hits 40 to 50 percent sustained response rates almost always uses three of these four channels in combination, with worklist routing that escalates between them.

Achieving sustained 40 to 50 percent response rates

The mechanics that move response rates from pilot levels (often 15 to 25 percent) to sustained ACCESS-eligible levels (40 to 50 percent):

• Clinician-attributed framing. Patients respond to "Dr. Smith asked you to share how you are feeling" at 2 to 3 times the rate of "your healthcare provider would like you to complete a survey."
• Pre-visit timing for chronic-condition cohort. Survey delivered 7 to 14 days before a scheduled chronic-condition follow-up has the highest engagement. The patient sees the connection between the survey and the upcoming visit.
• Worklist-driven re-engagement. Patients who do not respond in 7 days get an SMS nudge. Those who do not respond in another 7 days get rerouted to phone outreach by care management. Response rate by attempt: ~40 percent on attempt 1, +15 percent on attempt 2, +10 percent on attempt 3.
• Care-team accountability. Panel-level response-rate dashboards reviewed weekly by clinic leadership. Named accountability for low-performing panels with documented intervention.
• Patient education on why this matters. A short video or written explainer in the patient portal that frames PROMs as "how your doctor knows what is working." Modest impact, easy to deploy.

The teams that hit 50 percent sustained response rates do all five. The teams stuck at 25 percent typically do one or two and assume the technology is the problem.

Storing PROMs as structured FHIR Observation or QuestionnaireResponse

PROMs results have to live in your data platform as structured FHIR resources, not in a separate survey-vendor silo. Two FHIR resource patterns cover most needs.

QuestionnaireResponse for the raw response data, tied to a Questionnaire resource that defines the instrument structure. This is the source of truth, preserves question-level detail, and supports re-scoring if the instrument scoring algorithm is updated.

Observation for the computed scores, with a coded reference back to the QuestionnaireResponse for traceability. Observation is the natural surface for clinical workflow consumption (worklists, dashboards, EHR widgets) because most EHR FHIR APIs surface Observations more cleanly than QuestionnaireResponses.

The pattern that wins: write both. QuestionnaireResponse as the durable record, Observation as the workflow-consumable derived view, with lineage between them.

Surfacing PROMs in clinical workflow

PROMs results stored cleanly are still useless if no clinician sees them at the right time. Three workflow integration points that turn PROMs into outcomes:

1. In-EHR pre-visit summary. PROMIS-29 trend, condition-specific scores, and any threshold-crossing alerts visible to the clinician 24 hours before the encounter. This is where the clinical conversation starts.

2. Care management worklist. Panel-level rollup of patients with concerning PROMs trends (declining functional status, rising depression score, persistent pain score above target). Care management staff drive outreach.

3. Documented escalation thresholds. PHQ-9 score >= 15 routes to behavioral health. KCCQ-12 drop > 10 points routes to heart failure care manager. Documented in clinical-content governance, not just in care management heads.

The escalation thresholds are partly a clinical-content governance product and partly a workflow design product. Both have to ship.

Build versus buy

PROMs collection is the pillar where buying a vendor platform makes the most sense. Multi-tenant SMS, portal-integration, instrument library, and worklist tooling are non-trivial to build well. Vendor platforms also typically come with response-rate optimization patterns built in.

The hybrid model: buy the PROMs collection platform (Quantros, MyOutcomes, Health Catalyst, ClearPoint, or similar). Build the FHIR integration that pulls results into your data platform as QuestionnaireResponse and Observation. Build the clinician-workflow surfacing yourself because it has to be tightly integrated with your EHR and care-team operating model.

For deeper coverage of the ACO data platform that PROMs results land in, see our ACO data platform consulting deep-dive and the Building an ACO data platform practitioner guide.

---

How DATA4AI helps: We design and build the data platform integration, FHIR storage, and clinician-workflow surfacing that turns vendor-collected PROMs into ACCESS-scoring outcomes. We also help you select between PROMs vendors based on the workflow patterns that produce sustained 40 to 50 percent response rates. Book a discovery call to talk through your PROMs build, or download the readiness checklist for the full self-assessment.