By Sagar Shankaran, Founder of CallSphere
Pre-op PROMs collection is the highest-leverage AI voice workflow in surgical practices — and the one most likely to drift out of minimum necessary. Here is the 2026 design.
Key takeaways
A surgical practice with a pre-op PROMs program completed at 90%+ has structured outcome data nobody else has. An AI voice agent makes that achievable — without becoming a HIPAA liability or a clinical-judgment substitute.
flowchart TD
In[Patient interaction] --> MinNec{Minimum necessary?}
MinNec -->|yes| Process[AI process]
MinNec -->|no| Reject[Block + log]
Process --> Encrypt[(AES-256 at rest)]
Encrypt --> DB[(PostgreSQL)]
Process --> Audit[(Audit trail)]
DB --> Right[Right of access §164.524]A patient is scheduled for surgery 14, 7, 2, and 1 days out. The AI agent calls (or texts a portal link) at each touchpoint to administer the standardized PROMs (PROMIS, KOOS, HOOS, Oswestry, Eq-5D, depending on procedure), confirm pre-op instructions (NPO timing, medication holds, anticoagulation plan, transportation), capture changes in symptoms or medications, and answer common questions. Critical findings (new chest pain, signs of infection, medication misunderstandings) escalate to the surgical team in real time. Completed PROMs flow to the EHR as structured data.
Done well, the workflow lifts pre-op completion rates from 40% to 90%+ and gives the surgeon a baseline outcomes dataset. Done badly, it captures more PHI than necessary or replaces clinician judgment on a deteriorating patient.
PROMs collection is treatment under 45 CFR 164.501, fully covered by the treatment-payment-operations exception at 45 CFR 164.506(c). Minimum necessary at 45 CFR 164.502(b) governs which PROMs and which fields the agent collects — only what the surgeon and the post-op outcomes program use. Voicemail content stays generic (name, callback, "your pre-op survey is ready") under HHS guidance.
Hear it before you finish reading
Talk to a live CallSphere AI voice agent in your browser — 60 seconds, no signup.
The 2026 NPRM expectations — written technology asset inventory, MFA, encryption everywhere, annual technical safeguards verification — apply to the AI vendor as a business associate. The audit trail at 45 CFR 164.312(b) records every PROM administration and every pre-op confirmation.
When the agent escalates a critical finding (new chest pain, infection signs), the disclosure to the surgical team is treatment under 45 CFR 164.506(c). When the agent declines to give clinical advice, it is operating within the practice's protocol — the agent is not the practice of medicine.
CallSphere's Healthcare Voice Agent runs pre-op intake through the administer_proms, confirm_preop, and escalate_critical tools — 3 of 14 healthcare tools. The PROMs library carries PROMIS, KOOS, HOOS, Oswestry, Eq-5D, VR-12, and procedure-specific instruments; the surgeon picks the bundle per case type. The agent administers the PROM verbatim — wording matters for outcome scoring — captures answers as structured data, and writes them back to the EHR via FHIR R4. Pre-op instruction confirmations cover NPO timing, medication holds, anticoagulation, and transportation, with a clinician-vetted Q&A for common questions. Critical-finding keywords (chest pain, calf pain, fever, surgical-site redness, severe nausea on hold meds) escalate to the surgical team's task queue with a 30-minute SLA. Every call is captured in post-call analytics with sentiment (–1.0 to +1.0), lead score (0–100), AI summary, and audit trail in the encrypted healthcare_voice PostgreSQL database (1 of 115+ tables). HIPAA and SOC 2 aligned, 37 agents and 90+ tools across 6 verticals. Pricing on /pricing; start with 14-day trial; see /industries/healthcare for surgical specialties.
Can the AI agent give clinical advice on pre-op questions? Only what the practice protocol authorizes — typically structured Q&A on NPO, meds, and logistics. Anything beyond protocol routes to the surgical team.
Still reading? Stop comparing — try CallSphere live.
CallSphere ships complete AI voice agents per industry — 14 tools for healthcare, 10 agents for real estate, 4 specialists for salons. See how it actually handles a call before you book a demo.
What if the patient asks the agent to cancel surgery? The agent captures the request, escalates to the surgical scheduler, and follows the practice's cancellation policy. It does not act on cancellation autonomously.
Does the agent collect PROMs for outcome research? PROMs collection for treatment is TPO. Use of identifiable PROMs for research requires IRB and patient authorization unless de-identified per 45 CFR 164.514 or under a limited-data-set DUA.
Can the agent leave PROMs answers on a voicemail callback? No. PROMs answers stay in the secure record. The voicemail says "your pre-op survey is ready" with a callback or portal link.
What about the day-of-surgery confirmation? CallSphere ships a 5 a.m. NPO and arrival-time confirmation as a separate workflow — typically text only, with same minimum-necessary rules.
Written by
Sagar Shankaran· Founder, CallSphere
Sagar Shankaran is the founder of CallSphere, where he builds production AI voice and chat agents deployed across healthcare, hospitality, real estate, and home services. He writes about agentic AI, LLM engineering, and shipping voice agents that handle real calls in production.
See how AI voice agents work for your industry. Live demo available -- no signup required.
Using GPT-Realtime-2 for healthcare voice agents. BAA scope, PHI handling, retention, logging, and why a managed platform usually wins this build.
The 2024 NPRM proposes mandatory penetration tests every 12 months and vulnerability scans every 6 months. Here is how an AI voice agent should be tested in 2026.
AWS HealthScribe became the open scribe layer EHR vendors built on top of in 2026. Here's the API surface, the per-encounter pricing, the BAA terms.
Apollo, Manipal, and Narayana scaled AI agents across Bangalore in 2026. Here's the deployments across radiology, intake, and follow-up, the costs.
Notable's AI agents now handle scheduling, intake, and revenue cycle for 6,000+ clinics in 2026. Here's the multi-agent architecture, the per-clinic pricing.
Abridge raised $250M in April 2026 at a $2.7B valuation. We break down the deployment numbers, the EHR integrations across Epic and Cerner. The Q2 2026 buyer briefing.
© 2026 CallSphere LLC. All rights reserved.