For Hospital Leadership

Every note you sign is a billing document, a compliance document, and a legal document.

MedScribe AI makes sure it passes all three — and captures the APR-DRG severity your coders need, in both English and Spanish, exported MEDITECH-clean.

Designed and built by a practicing hospitalist inside Puerto Rico's hospital system — not an outside tech company retrofitting generic AI.

The Three-Test Framework

Every inpatient note must pass three independent tests. Most AI scribes only check one.

The Billing Test

Does the documentation actually support the CPT level and APR-DRG severity captured? We run every note through 46+ clinical safety rules, an evidence-anchored ICD-10 / CPT coding engine, and a deterministic MDM matrix based on AMA 2021 E/M and CMS MPFS 2023 time thresholds. No hallucinated diagnoses, no unsupported procedures, no tense errors.

The Compliance Test

Does it withstand a CMS, ASES, or internal audit? The validator blocks on missed NSAID-in-CKD cross-checks, AKI-vs-ESRD coding conflicts, unsupported sepsis claims, and 12 more hard-safety rules. Advisory-only rules surface documentation specificity gaps (CKD stage, CHF LVEF, pneumonia setting) that drive APR-DRG tier capture — the single biggest reimbursement lever no other scribe addresses.

The Legal Test

Does it hold up in a malpractice review? Every draft is explicitly a draft until the physician signs. Every AI suggestion has an accept / edit / reject audit trail in the database. PHI is redacted before any egress to OpenAI. Every PHI-touching action is logged. HIPAA-conscious architecture with staged BAA execution — we do not make compliance claims we can't back up.

The Math

Conservative ROI for a 10-hospitalist service, annualized post-pilot.

LeverPer physician / year10 physicians / year
Time savings — 40% reduction × 2 hr/day documentation × 240 shifts × $180/hr opportunity cost$34,560$345,600
Billing-score lift — MDM capture from 99232 → 99233 on ~15% of encounters × 2,000 encounters × $36 delta$10,800$108,000
APR-DRG severity tier capture — SOI tier shift on ~8% of admissions (PR-specific moat)TBD per hospitalMeasured in pilot
Reduced CDI query-back rate — ~30% fewer downstream queries, ~2 hr/wk reclaimed per coder~$20,000
Floor estimate (time + billing-score only)$453,600 / year
SHM-benchmarked median billing-score lift per hospitalist:$14,000 / year

APR-DRG lift is the largest unmeasured upside and is specific to the Puerto Rico Medicaid payment formula. We cannot quote it precisely until Phase 2 of the pilot runs on your actual case mix.

What Hospital Leadership Gets

Built for your workflow, not retrofitted from ambient AI.

APR-DRG severity awareness

Documentation prompts for severity specificity (acute vs chronic, stage, etiology) that drive SOI/ROM tier capture under Puerto Rico's payment formula. No other scribe has this.

CDI coaching built in

Evidence-anchored ICD-10 / CPT suggestions, cited from the physician's own note text — not inferred, not hallucinated. 43+ validated rules today, growing.

Bilingual clinical fluency

Dictate in Spanish, English, or mixed. Output reads as if written by a native clinician in either language. Section labels, MEDITECH exports, and CDI prompts all bilingual.

MEDITECH Expanse export

ASCII-safe, properly sectioned, copy-paste clean. No reformatting inside your EHR. Future SMART-on-FHIR integration roadmap.

46+ clinical safety rules

Every note gets permanent copilot-mode validator review before sign. NSAID in CKD, beta-blocker in decompensated CHF, AKI vs ESRD coding integrity, sepsis criteria cross-check, and growing.

Nocturnist-designed UX

Dark mode, 2-click Quick Record, auto-save, session-expiry warnings. Works on iPhone at the bedside. Built by a nocturnist for night-shift realities.

The Pilot

90 days. $0 during pilot. Staged compliance, co-authored product.

Duration
90 days (scheduled with your team)
Physicians
3-5 hospitalists on one service (nocturnists preferred for realistic stress testing)
Volume target
~500 notes across participating clinicians
Data classification
Phase 1 (days 1-30): synthetic + fully de-identified only. Phase 2 (days 31-90): real PHI, contingent on BAA execution.
Cost to the hospital
$0 during pilot. Post-pilot conversion at founder-hospital pricing to be negotiated.
Success criteria
≥40% documentation time reduction, APR-DRG tier capture improvement on target diagnoses, ≥95% clinical accuracy on adjudicated sample, NPS ≥50, zero PHI exposure incidents in Phase 1.

Compliance Posture

Staged and honest. We will not claim 'HIPAA compliant' in marketing or at the pilot table.

Encryption at rest
AES-256 (Postgres 16, Redis 7, Fernet-encrypted EHR tokens)
Encryption in transit
TLS 1.3 (Let's Encrypt, auto-renew)
Audit trail
Every PHI-touching action logged to a dedicated audit table with user ID, IP, resource, timestamp
Access control
Argon2 password hashing · JWT + Redis token blacklist · Refresh token rotation · Account lockout after 5 failed logins
De-identification
Regex PHI redaction on INPUT before any OpenAI egress (8 entity types)
Infrastructure BAA
DigitalOcean Standard Support activated (BAA document in motion)
Model provider BAA
OpenAI BAA — case under review
Pilot data policy
Synthetic or fully de-identified cases only until both BAAs are executed and filed with your compliance team

We claim 'HIPAA-conscious architecture with staged BAA execution.' That is the honest posture.

Why MedScribe — Not the U.S. Incumbents

Abridge, Nabla, Suki, DeepScribe, DAX — all outpatient-focused, English-first, and built for U.S. fee-for-service Medicare.

  • None support ASES APR-DRG severity-tier awareness.
  • None export MEDITECH Expanse clean.
  • None offer bilingual medical Spanish at clinical-fluency level.
  • None are built by a practicing Puerto Rico hospitalist using the product during real shifts.
  • None price below $100/month per physician. We will.

The Puerto Rico Medicare Advantage market is roughly $10 billion and structurally under-served by every U.S. scribe vendor. The first hospital system to deploy a bilingual, ASES-aware AI scribe gets the narrative, the case studies, and the operational head start.

Next Step

Let's talk.

Request a 30-minute call with your CMO, CFO, and CMIO. We will walk through the product, the staged-BAA pilot structure, and the specific documentation gaps in your case mix that drive APR-DRG reimbursement today.

Email Dr. Rodriguez

Dr. Hiram Rodriguez, MD

Founder & Lead Clinician · Board-certified Internal Medicine / Hospitalist / Nocturnist

drhiramrodriguez@medscribepr.com