Try free →
Nephrology specialty

Nephrology AI scribe 2026: CKD staging, dialysis documentation, transplant coordination, and the lab trend reality

May 8, 2026 · 6 min read

Nephrology workflow has three modes that don't share a template: outpatient CKD progression management, dialysis monthly capitation visits with strict CMS documentation, and transplant coordination across pre/post phases. All three depend heavily on lab trends. General AI scribes built for primary care don't capture the structured lab integration or score the KDIGO staging that drives nephrology decisions.

The 2026 nephrology-aware AI scribe stack handles four things: CKD staging with KDIGO category assignment, dialysis adequacy documentation (Kt/V, URR) for ESRD billing, transplant evaluation and post-transplant immunosuppression management, and the lab trend integration that makes nephrology decisions defensible.

CKD staging and progression

KDIGO categoryeGFR rangeDocumentation needs
G1 (normal/high)≥ 90With evidence of kidney damage (albuminuria, etc.)
G2 (mild)60-89With damage; not CKD without damage
G3a (mild-moderate)45-59HCC v28 capture relevant
G3b (moderate-severe)30-44HCC; nephrology referral threshold
G4 (severe)15-29RRT planning
G5 (failure)< 15 (or dialysis)ESRD if dialysis

Albuminuria categories (A1/A2/A3) modify the prognosis grid. The KDIGO staging guides drug dose adjustment, nephrology referral timing, and HCC capture. AI scribes should compute and document this automatically from current labs.

Dialysis monthly capitation visits

CMS bundles ESRD payment under monthly capitation that requires per-visit documentation:

The AI scribe with structured monthly visit format helps document everything CMS requires for the bundled payment.

The nephrology-aware system prompt

You are documenting a nephrology encounter.

INPUT:
- Encounter audio transcript
- Patient: age, sex, primary kidney dx, dialysis status (HD/PD/none)
- Last 12 months: BMP/CMP, eGFR, urine ACR, hemoglobin, calcium, phosphorus, iPTH
- For dialysis: Kt/V, URR, vascular access type, dry weight series, ESA dose
- For transplant: type (living/deceased), date, current immunosuppression, recent rejections
- Current medications

OUTPUT structured note based on visit type:

For CKD outpatient:
1. KDIGO category (G stage + A category)
2. eGFR trend (4-6 values), albuminuria trend
3. Comorbidity management (DM, HTN, lipid)
4. Drug dose adjustments per renal function
5. Nephrology referral / RRT planning if G4-5
6. Vaccination status

For dialysis MCP visit:
1. Adequacy (Kt/V or URR)
2. Vascular access status
3. Anemia management with ESA dose rationale
4. CKD-MBD parameters with management
5. Volume status + dry weight
6. Comorbidity management
7. Transplant candidacy review

For transplant:
1. Time post-transplant
2. Current immunosuppression doses (tacrolimus level, MMF dose, prednisone)
3. Recent rejection episodes if any
4. Infection surveillance (CMV, BK, EBV)
5. Cancer surveillance
6. Comorbidity management

For each clinical decision, cite lab trend or transcript. For HCC capture, ensure CKD stage is at maximum specificity (N18.31 not N18.3).

Vendor and DIY paths

For solo or small nephrology practices, the DIY stack with structured KDIGO staging and dialysis adequacy documentation outperforms general scribes meaningfully. The high-volume dialysis monthly capitation visit format is the highest-ROI use case — a structured note schema saves 5-10 minutes per visit.

BAA chain

Practice + EHR + dialysis facility EHR (Davita Falcon, FMC eCube, etc.) + lab vendor + transcription vendor + LLM vendor. 5-6 BAA documents.

When to start

If your nephrology practice has dialysis monthly capitation patients (almost all do), the structured monthly visit documentation pays back the AI scribe stack in the first month. The HCC v28 capture for CKD stages G3a/G3b on the outpatient panel is additional revenue lift on Medicare Advantage panels.

Nephrology DIY scribe stack on LessRec

$0.05/min Whisper. Build KDIGO staging + dialysis adequacy + transplant immunosuppression schema. First 10 minutes free.

Try LessRec free →