feat(modules): add CKD-staging module (eGFR-driven, KDIGO G1-G5)

[aerenkaradag]

What

Adds a Synthea-style CKD-staging module (src/syntha/modules/ckd.py) that fires on the Kronik_Bobrek source flag and stages the patient G1–G5 from egfr_latest per KDIGO 2024.

For each CKD-flagged episode it emits:

Resource	When	Code
Encounter (nephrology follow-up)	always	reason = stage-specific SNOMED (or generic 709044004 if eGFR missing)
Procedure — eGFR measurement	always	SNOMED 80274001
Procedure — ACR reflex test	G3a+ (eGFR < 60)	SNOMED 271000000
Procedure — nephrology referral	G3b+ (eGFR < 45)	SNOMED 103696004
CarePlan — CKD management	always	monitoring + diet + nephrotoxic-drug avoidance (413332001)

Also adds CKD_STAGES + ckd_stage_for_egfr() to fhir/codes.py with stage-specific SNOMED CT and ICD-10 N18.1–N18.5 (replacing the generic N18.9 when eGFR is present), registers CKDModule() in the module REGISTRY, and adds 6 tests in tests/test_modules.py.

Why

Per COLLABORATE.md / ROADMAP.md v0.6, CKD is a high-prevalence condition with egfr_latest + creatinine_latest available on every source patient but no module yet. Closes #17.

Verification

• pytest -q → 108 passed (was 102; +6 new CKD tests)
• ruff check clean on all changed files
• CLI smoke test (syntha fhir on examples/sample_output/sample_episodes.csv) passes; CKD eGFR Procedure + CarePlan confirmed present in generated bundles
• New module does not fire for existing tests (default fixture has no Kronik_Bobrek), so no regression

⚠️ Clinical confirmation needed (reviewer / clinician sign-off)

I implemented the international KDIGO standard; the following are clinical-judgment calls that need a Turkish primary-care clinician's confirmation before merge:

• eGFR cut-points — G1 ≥ 90, G2 60–89, G3a 45–59, G3b 30–44, G4 15–29, G5 < 15. Confirm TR uses the same.
• Nephrology-referral stage — currently modelled at G3b and worse (eGFR < 45). Confirm TR primary-care referral threshold (some refer at G4, some at G3b with progression/albuminuria).
• ACR reflex trigger — modelled from G3a (eGFR < 60). Confirm.
• Nephrotoxic-drug avoidance is surfaced as a CarePlan note only; the module does not de-prescribe. Acceptable, or should it interact with other modules' MedicationRequests (e.g. metformin/NSAID flags at low GFR)?
• SNOMED/ICD-10 stage codes — confirm N18.3 for both G3a and G3b is acceptable (ICD-10 has no a/b split; SNOMED does, and is used).

Happy to adjust any thresholds per your guidance.

---

Submitted as an invited contributor (@aerenkaradag). The Turkish source cohort behind syntha originates from the Check-Up dataset work.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 90.19608% with 5 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/syntha/fhir/codes.py	66.6%	4 Missing ⚠️
src/syntha/modules/ckd.py	97.3%	1 Missing ⚠️
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@           Coverage Diff           @@
##            main     #46     +/-   ##
=======================================
- Coverage   91.1%   91.0%   -0.1%     
=======================================
  Files         38      39      +1     
  Lines       1551    1602     +51     
=======================================
+ Hits        1413    1459     +46     
- Misses       138     143      +5     

Flag	Coverage Δ
unittests	91.0% <90.1%> (-0.1%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/syntha/modules/__init__.py	100.0% <100.0%> (ø)
src/syntha/modules/ckd.py	97.3% <97.3%> (ø)
src/syntha/fhir/codes.py	68.1% <66.6%> (-1.9%)	⬇️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.