Health Risk Assessments by Voice AI: HRA Completion at Scale for TPAs and Health Plans

The HRA completion problem

Every health plan and third-party administrator knows the value of a completed Health Risk Assessment. A finished HRA tells you which members need care management, which care gaps are open, and which conditions belong in the record. What every plan and TPA also knows is how hard the last step is. The instrument is fine. The list is ready. The members just will not pick up.

Live agent outreach hits a small fraction of a member list per attempt. No-answers, voicemail, disconnected numbers, and callback windows that never line up eat most of the effort. The work also surges. An enrollment period or an annual reassessment cycle can triple the volume for a few weeks, and no in-house team is staffed for the peak without being overstaffed the rest of the year. So programs settle for whatever completion rate the phone bank can reach, and the assessments that never happen quietly become missed care, missed gaps, and missed accuracy in the record.

This is exactly the shape of problem conversational voice AI is built to solve, and it is why EHVA now conducts HRAs directly, by phone, at volume.

What an HRA actually does, and why completion matters

A Health Risk Assessment is a structured conversation. It captures a member's health status, active conditions, medications, functional limitations, and social risk factors such as transportation, food, and housing. That structured picture feeds several workstreams at once:

Care management enrollment. HRA responses are how many members get identified and routed into the right care or disease management program before a small problem becomes an admission.

Care gap closure and quality. Answers surface open gaps that map to HEDIS and Stars measures, so outreach and scheduling can happen while there is still time in the measurement year.

Risk adjustment accuracy. A complete assessment supports an accurate record of a member's conditions, which is the basis of fair risk adjustment. Accuracy here is a compliance matter, not a volume game, and it depends on actually reaching the member.

Regulatory obligations. Some lines of business carry explicit HRA requirements. Medicare Advantage Special Needs Plans, for example, are required to conduct an initial assessment for new enrollees and to reassess annually. A program that cannot reach members cannot meet an obligation like that reliably.

In every one of these cases, the value is locked behind completion. An HRA that is 90 percent designed and 40 percent completed produces 40 percent of the outcome.

Why traditional HRA outreach breaks down

The failure is rarely the script. It is capacity and timing. Four things quietly cap most phone-based HRA programs:

Contact rate. A human dialer reaches a limited number of live members per hour, and most attempts end in no contact. Every miss has to be queued for another day.

Business hours. Members answer on evenings and weekends, which is precisely when an in-house team is smallest or closed.

Seasonality. Reassessment and enrollment windows create sharp peaks. Staffing to the peak is expensive, staffing to the average leaves assessments undone.

Language. A meaningful share of members are best served in a language the day's available agents may not speak, so those members get deprioritized.

You can throw people at these problems, but people are the constraint. The economics never quite work, which is why completion rates plateau where they do.

How EHVA runs Health Risk Assessments by voice

EHVA conducts the HRA as a natural spoken conversation, not a robocall reading a form. The assistant works both directions of the call. It runs persistent, scheduled outbound outreach across a member list, and it answers inbound instantly when a member calls back after a text or a voicemail. Because EHVA is built on a proprietary stack rather than stitched-together consumer tools, latency stays low enough that the conversation feels human, which is what keeps members on the line long enough to finish.

A live HRA deployment does the work of a trained assessor:

It follows your instrument. EHVA administers your exact HRA questions, in your order, with your branching logic. Conditional follow-ups fire based on answers, the same way a good human assessor would probe.

It captures structured data. Responses are recorded as discrete fields mapped to your assessment, not a freeform summary someone has to re-key. You get a completed record, a recording, and a transcript.

It writes back in real time. Through integration with your care management platform, data warehouse, or assessment vendor, the completed HRA lands in the system that needs it as soon as the call ends.

It works in multiple languages. The same assistant serves members in the language they prefer, without waiting for a bilingual agent to be free.

It never stops or tires. Every call is administered with identical consistency, at 2 p.m. on a Tuesday or 7 p.m. on a Saturday, across as many simultaneous conversations as the program needs.

This is the same pattern EHVA already runs for eligibility, benefits, and claims work at scale for TPAs. Our TPA deployments have reached 81% autonomy on in-scope structured member and provider calls, handling the routine completions autonomously and reserving people for the calls that genuinely need judgment. HRAs fit that pattern directly. You can hear how EHVA handles related member-data conversations on our real call recordings.

Escalation and the calls that need a human

Autonomy does not mean the assistant handles everything. It means the assistant handles the routine, in-scope volume so your clinicians and representatives spend their time where it counts. Escalation rules are defined before launch. If a member asks for a person, becomes distressed, or gives an answer that trips a clinical or safety flag, EHVA hands off to a live nurse or representative with the full context of the conversation already captured, or routes the case according to your protocol. A member who mentions a serious symptom is never left talking to software about it.

Compliance and member trust

HRAs involve protected health information, so the deployment has to be built for that from the start. EHVA operates under a Business Associate Agreement, applies access controls to PHI, and produces recordings and transcripts that give your compliance and quality teams a complete audit trail. For outbound member outreach, disclosures, consent handling, and calling practices are configured to align with your plan's posture and applicable rules, and EHVA is an FCC-registered carrier rather than a reseller of consumer calling platforms. Scope, disclosures, and escalation are agreed per client. For a deeper look at how PHI is handled on these calls, see our guide to HIPAA compliance for provider call AI.

What this changes for a TPA or health plan

The practical effect is a program that is no longer capacity limited. Outreach runs continuously and after hours, retries happen on schedule instead of when someone has time, inbound callbacks are answered on the first ring, and a language mix is handled by default. More conversations start, and because the conversation feels natural, more of them finish. The completed assessments then flow straight into the workflows that depend on them, care management enrollment, gap closure, and an accurate record, without a re-keying step in the middle.

And because EHVA prices by usage at $0.09 per minute with no long-term contract, the program can expand for a reassessment push and contract afterward. You are not carrying a fixed phone-bank cost through the quiet months to be ready for the busy ones.