name: neuro-checker description: Validates clinical decision support recommendations for neurological diagnoses. Scores content across six domains (completeness, accuracy, safety, objectivity, setting appropriateness, usability) and provides specific revision suggestions. Use after running neuro-builder to validate output quality, or when reviewing any clinical recommendation content for accuracy and safety.
Neuro Clinical Recommendation Checker¶
Validate and score clinical decision support content across six quality domains. Provides structured feedback for iterative improvement.
Expected Structure¶
Documents should have 8 sections in two groups:
Section A: Action Items (Primary) 1. Laboratory Workup (1A, 1B, 1C) 2. Diagnostic Imaging & Studies (2A, 2B, 2C) 3. Treatment (3A, 3B, 3C, 3D) 4. Other Recommendations (4A, 4B, 4C) ↠Must be diagnosis-specific
Section B: Reference (Supporting) 5. Differential Diagnosis 6. Monitoring Parameters 7. Disposition Criteria 8. Evidence & References
Expected Table Format (Option A Multi-Column)¶
Tables should use the multi-column setting-priority format:
Labs/Studies:
| Test | ED | HOSP | OPD | ICU | Rationale | Target Finding |
|------|:--:|:----:|:---:|:---:|-----------|----------------|
Treatments:
| Treatment | ED | HOSP | OPD | ICU | Dosing | Contraindications | Monitoring |
|-----------|:--:|:----:|:---:|:---:|--------|-------------------|------------|
Priority values: STAT | URGENT | ROUTINE | EXT | - (not applicable)
Validation Domains¶
Score each domain 0-10:
1. COMPLETENESS (0-10)¶
- Are all 8 sections present and populated?
- Are essential labs, imaging, and treatments included?
- Are there obvious gaps in the workup?
- Is the differential diagnosis comprehensive?
- Setting Coverage: Are all applicable settings (ED, HOSP, OPD, ICU) represented?
- Medication Coverage:
- Each medication on its own row (not grouped)?
- Complete dosing for each drug (start, titrate, max)?
- Symptomatic treatments cover relevant categories (pain, spasticity, bladder, mood, fatigue)?
- DMTs individually listed with pre-treatment requirements (if applicable)?
- Section 4 Checklist:
- 4A (Referrals & Consults) present with ≥5 relevant referrals?
- 4B (Patient Instructions) present with ≥5 actionable instructions?
- 4C (Lifestyle & Prevention) present with ≥3 recommendations?
- All Section 4 content is diagnosis-specific (not generic)?
2. ACCURACY (0-10)¶
- Are drug dosages correct?
- Are diagnostic criteria accurately stated?
- Are contraindications complete?
- Any factual errors?
- Are priority levels appropriate for each setting?
3. SAFETY (0-10)¶
- Are critical contraindications noted?
- Are drug interactions or monitoring requirements included?
- Are red flags and escalation triggers clear?
- Any potentially dangerous omissions?
- Section 4B Safety Check: Are return precautions specific and actionable?
4. OBJECTIVITY (0-10)¶
- Are claims evidence-based or clearly labeled as expert opinion?
- Is language neutral (no promotional bias)?
- Are limitations of evidence acknowledged?
5. SETTING APPROPRIATENESS (0-10)¶
- Are all four settings (ED, HOSP, OPD, ICU) appropriately covered?
- Are priorities correct for each setting (STAT in ED, ROUTINE in OPD)?
- Is "-" used correctly (truly not applicable, not just "less common")?
- Flag gaps: Missing OPD labs? Missing ED treatments?
- Would recommendations be actionable in each tagged setting?
6. USABILITY (0-10)¶
- Is format consistent and parseable?
- Is the multi-column format correctly structured?
- Could this be directly converted to JSON for clinical tools?
- Are recommendations written as directives (not suggestions)?
- Section 4 Usability: Are patient instructions written in plain language?
CRITICAL: Setting Coverage Validation¶
This is the most important new check. For each diagnosis, verify:
Setting Coverage Checklist¶
| Section | ED Coverage | HOSP Coverage | OPD Coverage | ICU Coverage |
|---|---|---|---|---|
| 1. Labs | ≥3 tests? | ≥3 tests? | ≥3 tests? | If applicable? |
| 2. Imaging | If urgent needed? | ≥1 study? | ≥1 study? | If applicable? |
| 3. Treatment | Acute therapies? | Maintenance? | Maintenance? | Critical care? |
| 4A. Referrals | If applicable? | ≥3 consults? | ≥3 consults? | If applicable? |
| 4B. Instructions | Discharge teaching? | Discharge teaching? | All applicable? | N/A |
Common Coverage Gaps to Flag¶
| Gap Type | Example | Impact |
|---|---|---|
| Missing OPD labs | Seizure workup has no OPD-tagged labs | Outpatient neurologist can't order workup |
| Missing OPD imaging | MRI only tagged for ED/HOSP | Clinic-based imaging not supported |
| Missing HOSP maintenance | Only acute treatments tagged | Admitted patients don't get home meds continued |
| Overuse of "-" | OPD marked "-" for tests that CAN be done outpatient | Limits ambulatory care |
| Missing ED consults | Referrals only tagged OPD | Delays specialty input in emergencies |
When "-" Is Appropriate¶
"-" means truly not applicable, not "less common" or "usually done elsewhere."
| Appropriate "-" | Inappropriate "-" |
|---|---|
| Prolactin for OPD (only useful <20 min post-ictal) | Basic metabolic panel for OPD |
| ICU-only monitoring for routine outpatient | MRI brain for OPD (can be ordered from clinic) |
| Emergent intubation meds for OPD | Maintenance AEDs for ED (may need to continue) |
Output Format¶
DIAGNOSIS REVIEWED: [Name]
VERSION: [if applicable]
## SCORES
| Domain | Score | Notes |
|--------|-------|-------|
| Completeness | X/10 | [brief note] |
| Accuracy | X/10 | [brief note] |
| Safety | X/10 | [brief note] |
| Objectivity | X/10 | [brief note] |
| Setting Appropriateness | X/10 | [brief note] |
| Usability | X/10 | [brief note] |
**OVERALL SCORE: X/60 (X%)**
## SETTING COVERAGE ASSESSMENT
| Section | ED | HOSP | OPD | ICU | Gaps Identified |
|---------|:--:|:----:|:---:|:---:|-----------------|
| 1. Labs | ✓/✗ | ✓/✗ | ✓/✗ | ✓/✗ | [specific gaps] |
| 2. Imaging | ✓/✗ | ✓/✗ | ✓/✗ | ✓/✗ | [specific gaps] |
| 3. Treatment | ✓/✗ | ✓/✗ | ✓/✗ | ✓/✗ | [specific gaps] |
| 4. Other | ✓/✗ | ✓/✗ | ✓/✗ | ✓/✗ | [specific gaps] |
## SECTION 4 ASSESSMENT
| Subsection | Present | Count | Diagnosis-Specific | Notes |
|------------|---------|-------|-------------------|-------|
| 4A Referrals | Y/N | X items | Y/N | [note] |
| 4B Patient Instructions | Y/N | X items | Y/N | [note] |
| 4C Lifestyle | Y/N | X items | Y/N | [note] |
## MEDICATION ASSESSMENT
| Criterion | Met? | Notes |
|-----------|------|-------|
| Each drug on individual row (not grouped) | Y/N | [note] |
| Complete dosing (start, titrate, max) | Y/N | [list any drugs missing dosing] |
| Contraindications listed for each drug | Y/N | [note] |
| Monitoring specified for each drug | Y/N | [note] |
| Symptomatic categories covered | Y/N | [list missing: pain, spasticity, bladder, mood, fatigue, etc.] |
| Section 3B has Indication column | Y/N | [note] |
| Section 3D has Route + Pre-Treatment columns | Y/N or N/A | [note] |
## SECTION 6 MONITORING ASSESSMENT
| Criterion | Met? | Notes |
|-----------|------|-------|
| Venue columns present (ED, HOSP, OPD, ICU) | Y/N | [Section 6 must have venue columns] |
| Frequency column present | Y/N | [note] |
| Target/threshold specified for each parameter | Y/N | [note] |
| Action if abnormal specified | Y/N | [note] |
| Critical safety monitoring included (if applicable) | Y/N | [e.g., respiratory function for GBS/MG] |
| Longitudinal monitoring has OPD coverage | Y/N | [note] |
## CRITICAL ISSUES (Must Fix)
[List any dangerous or major errors - these block approval]
C1. [Section] - [Description of critical issue]
C2. [Section] - [Description of critical issue]
## SETTING COVERAGE ISSUES
[List specific setting gaps - high priority]
S1. [Section] - [Setting gap: e.g., "Labs section has no OPD coverage - outpatient workup not possible"]
S2. [Section] - [Setting gap]
## MEDICATION ISSUES
[List medication comprehensiveness gaps - high priority]
M1. [Section] - [Medication issue: e.g., "Drugs grouped as 'SSRIs' instead of individual rows"]
M2. [Section] - [Medication issue: e.g., "Gabapentin missing titration schedule"]
M3. [Section] - [Medication issue: e.g., "No bladder dysfunction treatments included"]
## RECOMMENDED REVISIONS
[Specific suggestions ranked by priority]
R1. [Section] - [Specific revision with rationale]
R2. [Section] - [Specific revision with rationale]
...
## STRENGTHS
[What was done well - reinforces good patterns]
S1. [Description]
S2. [Description]
## VERIFICATION NEEDED
[Items requiring clinical expert verification]
V1. [Item needing physician confirmation]
V2. [Item needing physician confirmation]
Checker Principles¶
- Prioritize patient safety above all - Flag any potentially dangerous omissions or errors as CRITICAL
- Validate comprehensive setting coverage - Flag gaps in ED, HOSP, OPD, or ICU coverage
- Flag unsupported claims - Note where evidence is missing or weak
- Be specific in feedback - Cite exact section/row needing revision
- Recommend verification when uncertain - Don't assume error if unclear, flag for expert review
- Check for appropriate "-" usage - Ensure "-" means truly not applicable
- Verify directive language - Recommendations should be checkbox-ready, not suggestions
- Validate Section 4 specificity - Flag generic or copy-paste content that doesn't fit the diagnosis
Common Issues to Check¶
Medication Comprehensiveness (NEW - HIGH PRIORITY): - Are all medications listed individually (not grouped as "SSRIs" or "beta-blockers")? - Does each drug have complete dosing (start dose, titration, max dose)? - Are contraindications listed for each medication? - Is monitoring specified for each medication? - For DMTs: Are pre-treatment requirements listed? - Are relevant symptomatic treatment categories covered (pain, spasticity, bladder, mood, fatigue)? - Does Section 3B include an "Indication" column? - Does Section 3D include "Route" and "Pre-Treatment Requirements" columns (if DMTs apply)?
Setting Coverage (HIGH PRIORITY): - Missing OPD coverage in Labs section - Missing OPD coverage in Imaging section - Overuse of "-" where tests/treatments ARE applicable - Inconsistent priorities across settings - Missing HOSP column for maintenance therapies
Section 6 Monitoring (HIGH PRIORITY): - Missing venue columns (ED, HOSP, OPD, ICU) - all monitoring parameters need venue columns - Missing critical safety monitoring (respiratory function for neuromuscular conditions) - Missing longitudinal monitoring (OPD coverage for ongoing parameters) - Venue interpretation: venue indicates where monitoring is ordered/initiated, not necessarily performed
Safety: - Glucose monitoring during steroids - Respiratory monitoring thresholds for myelitis - Drug-drug interactions - Contraindication completeness - Escalation criteria clarity
Accuracy: - Drug dosing (especially weight-based) - Monitoring frequencies - Evidence levels cited correctly - ICD-10 codes - Priority levels appropriate for setting
Completeness: - Missing differential diagnoses - Incomplete disposition criteria - Missing evidence references - Gaps in monitoring parameters - Missing Section 4 subsections (4A, 4B, 4C)
Objectivity: - Brand name bias - Unsupported treatment preferences - Missing alternative therapies
Usability - Directive Language: Flag any recommendations that use weak/suggestive language: - "Consider..." → Rewrite as direct order - "May benefit from..." → Rewrite as direct order - "Should be encouraged to..." → Rewrite as direct instruction
Setting Coverage Scoring Impact¶
| Issue | Scoring Impact |
|---|---|
| One setting missing entirely from a section | Setting Appropriateness: -2 points |
| Multiple settings missing coverage | Setting Appropriateness: -3 to -4 points |
| Inappropriate use of "-" (should have coverage) | Setting Appropriateness: -1 point per instance |
| OPD completely missing from Labs | Completeness: -2 points; Setting: -2 points |
| Priorities incorrect for setting (e.g., STAT in OPD) | Accuracy: -1 point per instance |
Scoring Guidelines¶
- 9-10: Excellent, minimal or no changes needed
- 7-8: Good, minor revisions only
- 5-6: Acceptable, moderate revisions needed
- 3-4: Significant issues, major revision required
- 0-2: Unsafe or fundamentally flawed, rebuild required
Overall thresholds: - ≥54/60 (90%): Ready for clinical use pending physician sign-off - 48-53/60 (80-89%): Revise and re-check - <48/60 (<80%): Significant revision needed
Change Log¶
v2.2 (January 14, 2026) - Added Section 6 Monitoring Assessment table to output format - Added venue column validation for Section 6 (ED, HOSP, OPD, ICU required) - Added Section 6 Monitoring to Common Issues to Check - Added critical safety monitoring validation (respiratory function for neuromuscular conditions) - Section 6 venue interpretation: where monitoring is ordered/initiated, not necessarily performed
v2.1 (January 13, 2026) - Added Medication Assessment table to output format - Added M-codes for medication issues (M1, M2, M3, etc.) - Added Medication Comprehensiveness checks to Common Issues section - Updated Completeness domain with medication coverage checklist - Checks for: individual drug rows, complete dosing, symptomatic categories, DMT columns
v2.0 (January 13, 2026) - Added Setting Coverage Assessment table to output format - Added Setting Coverage Issues section (S1, S2, etc.) - Added comprehensive setting coverage validation checklist - Added common coverage gaps to flag - Added guidance on appropriate vs inappropriate "-" usage - Added setting coverage scoring impact - Updated all six domain descriptions for multi-column format - Added format validation for Option A structure
v1.0 (January 13, 2026) - Initial version