New Hampshire's Push on Telemedicine and Insurer AI: What Insurance Teams Need to Do Now
New Hampshire lawmakers are weighing bills to expand telemedicine access and set clear guardrails for how health insurers use AI. If you operate in product, compliance, underwriting, or claims, this is your early signal to prepare. The direction is clear: parity for telehealth coverage and documented, human-overseen use of algorithms.
Telemedicine: Parity, Payment, and Practical Operations
The proposed measures focus on coverage parity, cleaner reimbursement, and fewer administrative hurdles. Expect attention on remote monitoring, continuity of care, and addressing licensing or credentialing questions-especially for providers serving rural areas.
- Benefit design and parity: Map in-person services to telehealth equivalents; update SPDs, SBCs, and internal guidelines.
- Reimbursement and coding: Align fee schedules, CPT/HCPCS codes, and modifiers; update claims edits to avoid systematic denials.
- Network and credentialing: Contract telehealth-capable providers; clarify remote monitoring policies and documentation.
- Prior authorization: Review PA criteria for virtual visits to prevent hidden barriers.
- Member communications: Clear notices on covered telehealth services, cost share, and tech requirements.
- Access considerations: Plan for rural/broadband gaps; offer alternatives and support.
AI in Insurance Operations: Guardrails You'll Likely Need
The bills target underwriting, eligibility, claims processing, and utilization review tools that use algorithms. The theme: transparency, fairness, security, documentation, and human oversight-especially for adverse determinations.
- System inventory: Catalog models, scoring tools, and decision aids used across the business and by vendors.
- Transparency triggers: Provide consumer notice when an automated system materially influences a decision.
- Human-in-the-loop: Require human review for denials and key adverse outcomes; track overrides.
- Validation and monitoring: Pre-deployment testing, periodic revalidation, drift checks, and calibration reviews.
- Fairness checks: Bias testing across protected classes; disparity metrics with thresholds and remediation plans.
- Explainability: Keep model summaries, factor importance, and case-level reason codes where feasible.
- Documentation and audit: Data lineage, training data sources, feature lists, versioning, change logs, and control owners.
- Vendor management: Contractual obligations for testing, transparency artifacts, security, and right-to-audit.
- Security and privacy: Data minimization, encryption, access controls, and retention aligned to policy.
- Reporting: Prepare to submit inventories, testing summaries, and incident reports to the state regulator.
Consumer Protections and Oversight
Expect stronger notice and appeal rights, plus regulator visibility into how decisions are made. That means better documentation, clearer member letters, and tighter turnarounds on grievances.
- Plain-language notices for decisions affected by automated tools.
- Appeal pathways with access to a human reviewer and supporting rationale.
- Centralized logs for complaints, appeals, and decision reversals.
Operational Timeline to Get Ahead
- Next 30-60 days: Build an AI/telehealth working group; complete system and policy inventories; identify quick gaps in parity and notices.
- Next 90-120 days: Draft model governance standards; pilot bias tests; update claims edits for telehealth; refresh member communications.
- Next 6-9 months: Formalize human review protocols; finalize vendor addenda; implement monitoring dashboards and audit trails.
Data and Metrics to Prepare
- Telehealth: Utilization by specialty and geography, denial reasons, PA hit rates, average time to appointment, NPS/complaints.
- Algorithms: Performance (AUC/KS/accuracy), calibration, error rates, disparity metrics, override rates, complaint patterns, drift indicators.
- Security/Privacy: Access logs, data flows, retention exceptions, third-party data sources and approvals.
Friction Points You Can Anticipate
Definitions matter. Expect debate over what counts as "AI" versus rules-based logic. Also watch costs from parity, provider capacity for virtual care, and rural broadband constraints. Clear scoping, change control, and phased rollouts reduce surprises.
Implications by Function
- Product/Actuarial: Reprice for telehealth parity; model utilization shifts; update filings if needed.
- Claims/Ops: Tune edits and workflows for virtual care; upgrade adverse determination letters; train staff for human review.
- Underwriting: Reassess any automated risk scores for bias and explainability; document overrides.
- Compliance/Legal: Own model governance policy, regulator reports, vendor clauses, and consumer notice templates.
- IT/Security: Lock down data pipelines, version control, monitoring, and incident response tied to models.
Next Steps Checklist
- Complete a telehealth coverage and reimbursement gap analysis.
- Publish an AI governance standard with roles, testing, and documentation requirements.
- Stand up human review for adverse determinations and log overrides.
- Update member notices to disclose meaningful automated influence.
- Amend vendor contracts to require testing artifacts and audit rights.
- Prepare a regulator-ready inventory with model summaries and controls.
If you need policy anchors, review the NAIC's AI principles and keep an eye on state-level updates.
Want to upskill teams on practical AI use, testing, and governance? See targeted learning paths here:
Bottom line: treat telehealth parity and AI oversight as a single, coordinated program. Build the inventory, document the controls, and put humans back in the loop where it matters. You'll reduce risk, improve decisions, and be ready when the rules go live.
Your membership also unlocks:
