What actually breaks a PHQ-9 program at population scale? Rarely the instrument. The PHQ-9 is nine questions, a clean scoring rule, and a LOINC code most tools already know. What breaks the program is everything around the nine questions: sending the survey at the right cadence, resurveying the people who screened positive, extracting a clean Observation into whatever downstream system needs it, and doing all of that without a fresh integration project every time the ACO or the Medicaid plan changes reporting formats.

The five tools below are the ones you actually see population health teams shortlisting in 2026. Each has a different center of gravity. For more, see the conecion FHIR library, which sits next to the medical-form work.

Luma Health

Luma Health leads with engagement. The PHQ-9 lands in the same messaging channel the patient already uses for appointment reminders, which lifts completion rates against colder outreach. The scoring runs inside Luma and the total plus severity band become available to the care team through Luma's dashboards. The gap is the extraction into a FHIR store: it happens, but it is a downstream integration step rather than the primary output.

A FHIR-Native SDC Runtime

Delivery channels are usually opinionated: tools like Force Therapeutics and PatientIQ bake in ortho-specific templates, while general-purpose SDC engines let you compose PROMs from a shared Questionnaire catalog and reuse the extraction across SMS, email, and portal. For the coding side, PHQ-9 total score binds to LOINC 44249-1, and systems like Formbox rely on a terminology server such as Termbox to resolve that at populate time. For a population health team that already runs a FHIR platform, the same Questionnaire that renders on the patient's phone can drive the same Observation shape on every downstream analytics or care-gap dashboard. Teams that want to see the PHQ-9 render before any procurement work can load the standard FHIR Questionnaire JSON into the browser sandbox at form-builder.aidbox.app in an afternoon.

Phreesia

Phreesia holds ground in ambulatory intake and has extended PHQ-9 delivery into that flow. The strength is capture at the point of care: the patient completes the survey on the tablet at check-in, which lifts completion rates above any digital-only channel. The gap is the population-health cadence outside the visit; Phreesia is not the tool for the between-visit resurvey a positive PHQ-9 triggers.

Feedtrail

Feedtrail entered the PHQ-9 space through its broader patient-experience work. The delivery is competent, the reporting is aimed at the QI team rather than the care management team, and the fit is strongest for programs that already run Feedtrail for HCAHPS or clinic surveys. For a program building its PHQ-9 workflow from scratch, Feedtrail rarely wins the primary slot.

HAPI FHIR SDC

The build-your-own path. HAPI FHIR's SDC module renders the PHQ-9 Questionnaire, evaluates the calculated expression for the total score, and returns a QuestionnaireResponse the team can extract into an Observation. The licensing cost is zero. The operational cost is the team that owns the delivery layer, the resurvey scheduling, and the reporting exports. Population health programs inside academic medical centers pick this path most often, because they already have the FHIR engineering.

PHQ-9 TOOLING FOR POPULATION HEALTH PURCHASED WORKFLOW TEAM-OWNED BUILD ENGAGEMENT-FIRST EXTRACTION-NATIVE Luma Health Messaging engagement, extraction downstream FHIR-native SDC runtime Shared catalog, one extraction contract Phreesia Point-of-care kiosk, no between-visit cadence Feedtrail Experience-lineage, QI reporting bias HAPI FHIR SDC Self-hosted, zero licensing, team-owned START FROM THE REPORTING CONTRACT · NOT THE VENDOR SHEET

How to Pick for Population Health

Population health teams should start from the reporting contract rather than the vendor sheet. If the program feeds an ACO quality report, a Medicaid plan's HEDIS submission, and an internal care-gap dashboard, the PHQ-9 needs to move through a single extraction into an Observation the three systems can share. If the program only feeds a single downstream dashboard, an engagement-first tool with dashboards baked in is defensible.

For skip-logic-heavy versions of the PHQ-9 (branching to the PHQ-9M for follow-up questions on suicidality), see 7 FHIR form tools that actually handle skip logic cleanly. For the enterprise-scale evaluation across silos, best FHIR Questionnaire tools for enterprise healthcare in 2026 covers the wider tool landscape. The right pick is the one whose extraction contract survives the next change to the reporting format.

— Rowan Delgado