tooluniverse-clinical-risk-scoring
This skill computes and interprets eight validated bedside clinical risk scores (CHA2DS2-VASc, HAS-BLED, CURB-65, qSOFA, Child-Pugh, MELD-Na, Wells DVT/PE, ASCVD, eGFR CKD-EPI) to support point-of-care decisions on anticoagulation, admission, sepsis screening, cardiovascular prevention, and kidney function. Use it when evaluating stroke or bleeding risk in atrial fibrillation, pneumonia severity, sepsis likelihood, cirrhosis acuity, venous thromboembolism pretest probability, cardiovascular risk, or estimated glomerular filtration rate.
git clone --depth 1 https://github.com/mims-harvard/ToolUniverse /tmp/tooluniverse-clinical-risk-scoring && cp -r /tmp/tooluniverse-clinical-risk-scoring/plugin/skills/tooluniverse-clinical-risk-scoring ~/.claude/skills/tooluniverse-clinical-risk-scoringSKILL.md
# Clinical Risk Scoring
Turn a clinical scenario into the right validated risk score, compute it with a deterministic calculator tool, and interpret the number into a clinical action. All 10 backing tools are pure-compute (no network, no API key) and return `{status, data:{score, interpretation, components, ...}}`.
This skill is **decision-support only** — see LIMITATIONS. It does not replace clinical judgment.
## Step 1 — Map the scenario to the score(s)
| Clinical scenario | Score(s) | Tool(s) |
|---|---|---|
| Atrial fibrillation — stroke risk / anticoagulate? | CHA2DS2-VASc **and** HAS-BLED (pair) | `ClinicalCalc_CHA2DS2_VASc` + `ClinicalCalc_HAS_BLED` |
| Community-acquired pneumonia — severity / admit? | CURB-65 | `ClinicalCalc_CURB_65` |
| Suspected sepsis (infection + ? deterioration) | qSOFA | `ClinicalCalc_qSOFA` |
| Cirrhosis / chronic liver disease severity | Child-Pugh **and** MELD-Na (pair) | `ClinicalCalc_Child_Pugh` + `ClinicalCalc_MELD_Na` |
| Suspected DVT — pretest probability | Wells DVT | `ClinicalCalc_Wells_DVT` |
| Suspected PE — pretest probability | Wells PE | `ClinicalCalc_Wells_PE` |
| Primary CVD prevention — 10-yr risk / statin? | ASCVD | `ClinicalCalc_ASCVD_risk` |
| Kidney function / renal drug dosing / CKD stage | eGFR CKD-EPI | `ClinicalCalc_eGFR_CKD_EPI` |
When the scenario names a **pair**, always run both — one alone is misleading (e.g. stroke risk without bleeding risk, or Child-Pugh without MELD-Na).
## Step 2 — Gather the required inputs
Required vs optional inputs per tool (omitted booleans default to `false`/absent; omitted scalars are rejected when required):
| Tool | Required | Key optional booleans/values |
|---|---|---|
| `ClinicalCalc_CHA2DS2_VASc` | `age` | `chf`, `hypertension`, `diabetes`, `stroke_history`(2pt), `vascular_disease`, `female` |
| `ClinicalCalc_HAS_BLED` | `age` | `hypertension`, `renal_disease`, `liver_disease`, `stroke_history`, `bleeding_history`, `labile_inr`, `drugs`, `alcohol` |
| `ClinicalCalc_CURB_65` | `age` | `confusion`, `elevated_urea`(BUN>19), `high_resp_rate`(>=30), `low_bp` |
| `ClinicalCalc_qSOFA` | (none) | `high_resp_rate`(>=22), `altered_mentation`, `low_sbp`(<=100) |
| `ClinicalCalc_Child_Pugh` | `bilirubin`, `albumin`, `inr` | `ascites`(none/mild/moderate), `encephalopathy`(none/grade1-2/grade3-4) |
| `ClinicalCalc_MELD_Na` | `creatinine`, `bilirubin`, `inr`, `sodium` | `dialysis` (forces creatinine to 4.0) |
| `ClinicalCalc_Wells_DVT` | (none) | `active_cancer`, `immobilization`, `recent_surgery`, `localized_tenderness`, `leg_swollen`, `calf_swelling`, `pitting_edema`, `collateral_veins`, `previous_dvt`, `alternative_diagnosis`(-2) |
| `ClinicalCalc_Wells_PE` | (none) | `clinical_dvt`(3), `pe_most_likely`(3), `tachycardia`(1.5), `immobilization`(1.5), `previous_vte`(1.5), `hemoptysis`(1), `malignancy`(1) |
| `ClinicalCalc_ASCVD_risk` | `age`(40-79), `total_cholesterol`, `hdl_cholesterol`, `systolic_bp` | `bp_treated`, `smoker`, `diabetes`, `female`, `race`("white"/"black") |
| `ClinicalCalc_eGFR_CKD_EPI` | `creatinine`, `age` | `female` |
If a required value is missing, ask the user for it — do not guess. State explicitly which booleans you assumed `false`.
## Step 3 — Compute
```bash
tu run ClinicalCalc_CHA2DS2_VASc '{"age":76,"female":true,"hypertension":true,"diabetes":true}'
```
Every tool returns `data.score` plus a human-readable `data.interpretation` and a `data.components` breakdown (per-factor points). MELD/eGFR/ASCVD also return `unit`; Child-Pugh returns `child_pugh_class`; Wells PE returns `three_tier` and `two_tier`. Echo the `components` so the user can audit which factors drove the score.
## Step 4 — Interpret (per-score tables)
### CHA2DS2-VASc (stroke risk in AF, 0–9)
| Score | Stroke risk | Action |
|---|---|---|
| 0 (men) / 1 (women, sex point only) | Low | No anticoagulation |
| 1 (men) | Intermediate | Consider anticoagulation |
| **>=2 (men) / >=3 (women)** | Elevated | **Oral anticoagulation recommended** |
### HAS-BLED (major bleeding on anticoagulation, 0–9)
| Score | Bleeding risk | Action |
|---|---|---|
| 0–2 | Low–moderate | Anticoagulation reasonable |
| **>=3** | High | Caution; correct **reversible** factors (BP, labile INR, antiplatelet/NSAID, alcohol), closer follow-up — NOT an automatic contraindication |
### How to weigh CHA2DS2-VASc + HAS-BLED together
A high HAS-BLED **does not by itself withhold anticoagulation**. If CHA2DS2-VASc meets the threshold, the stroke benefit usually outweighs bleeding risk; HAS-BLED instead flags **modifiable** risk factors to fix and patients needing closer monitoring. Only a very high, non-modifiable bleeding risk shifts the decision against anticoagulation.
### CURB-65 (CAP severity, 0–5)
| Score | 30-day mortality | Disposition |
|---|---|---|
| 0–1 | Low (~1.5–3%) | Outpatient |
| 2 | Intermediate (~9%) | Short-stay / inpatient admission |
| **3–5** | High (~15–40%) | Inpatient; assess for **ICU** at 4–5 |
### qSOFA (sepsis screen, 0–3)
| Score | Meaning |
|---|---|
| 0–1 | Lower risk — does not rule out sepsis; reassess |
| **>=2** | Higher risk of poor outcome — escalate, full sepsis workup, consider full SOFA / lactate |
qSOFA is a **screen**, not a diagnosis; a low score never excludes sepsis.
### Child-Pugh (cirrhosis severity, class A/B/C)
| Class | Score | 1-yr survival (approx) | Meaning |
|---|---|---|---|
| A | 5–6 | ~100% | Well-compensated |
| B | 7–9 | ~80% | Significant functional compromise |
| C | 10–15 | ~45% | Decompensated; high surgical/anesthetic risk |
### MELD-Na (90-day mortality / transplant priority, 6–40)
| MELD-Na | 90-day mortality (approx) | Transplant relevance |
|---|---|---|
| <=9 | ~2% | Low priority |
| 10–19 | ~6% | |
| 20–29 | ~20% | Rising allocation priority |
| 30–39 | ~50% | High priority |
| **>=40** | **>50%** | Highest priority |
Pair with Child-Pugh: Child-Pugh class anchors chronic severity / surgical risk; MELD-Na drives short-term mortality and transplant listing.Install and configure ToolUniverse for any use case — MCP server (chat-based), CLI (command line with 9 subcommands), or Python SDK (Coding API with 3 calling patterns). Covers uv/uvx setup, MCP configuration for 12+ AI clients (Cursor, Claude Desktop, Windsurf, VS Code, Codex, Gemini CLI, Trae, Cline, etc.), full CLI reference (tu list/grep/find/info/run/test/status/build/serve), Coding API quickstart, agentic tools, code executor, API key walkthrough, skill installation, and upgrading. Use when user asks how to set up ToolUniverse, which access mode to use (MCP vs CLI vs SDK), configuring MCP servers, using the CLI, troubleshooting installation, upgrading, or mentions installing ToolUniverse or setting up scientific tools. Also triggers for "how do I use ToolUniverse", "what's the best way to access tools", "command line", "tu command", "coding API", "tu build".
Systematic ACMG/AMP germline variant classification with all 28 criteria (PVS1, PS1-4, PM1-6, PP1-5, BA1, BS1-4, BP1-7) for clinical significance. Produces 5-tier verdict (Pathogenic / Likely Pathogenic / VUS / Likely Benign / Benign) with cited evidence per criterion. Use for variant interpretation, VUS resolution, and pathogenicity assessment. Combines ClinVar, gnomAD, computational predictors, and gene-mechanism context.
Comprehensive ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) profiling for drug candidates. Integrates ADMET-AI predictions, SwissADME drug-likeness, PubChemTox experimental toxicity, ChEMBL clinical data, Lipinski rule-of-five, and CYP interaction data. Use for drug-likeness assessment, BBB penetration, bioavailability, hepatotoxicity prediction, ADME/PK profiling, or screening compound libraries before lab testing.
Detect and analyze adverse drug event signals using FDA FAERS reports, drug labels, and disproportionality statistics (PRR, ROR, IC). Generates quantitative safety signal scores (0-100) with evidence grading. Use for post-market surveillance, pharmacovigilance, drug safety assessment, regulatory submissions, and detecting rare AE signals not visible in clinical trials.
Map environmental and industrial chemicals to adverse outcome pathways (AOPs) — molecular initiating event to organ-level toxicity. Uses AOPWiki, GHS classification, IARC carcinogen status, and LD50 data. Use for environmental/industrial chemical risk assessment, regulatory-grade hazard characterization, and AOP stressor mapping. Distinct from drug-safety analysis (use tooluniverse-pharmacovigilance for drugs).
Aging biology, cellular senescence, and longevity research. Covers senescence markers (p16/CDKN2A, SASP, SA-beta-gal), aging hallmarks, senolytic drug discovery (dasatinib+quercetin, fisetin, navitoclax), epigenetic clocks, telomere biology, and longevity GWAS. Use for senescence-pathway analysis, age-related disease genetics, senolytic-target discovery, and centenarian-genetics queries. Distinguishes correlative vs causal evidence (knockout, intervention).
Therapeutic antibody engineering and optimization, lead-to-clinical-candidate. Covers sequence humanization (germline alignment, framework retention), affinity maturation, developability (aggregation, stability, PTMs), structure modeling (AlphaFold/PDB CDR analysis), immunogenicity prediction, and manufacturing feasibility. Use for biologic-drug optimization, mAb design review, biosimilar engineering, and clinical-precedent comparison.
Discover novel small-molecule binders for protein targets using structure-based and ligand-based screening. Covers druggability assessment, known-ligand mining (ChEMBL, BindingDB), similarity expansion, ADMET filtering, and synthesis feasibility. Use for hit identification, virtual screening, target-to-compounds workflows, and lead-finding before commit-to-medchem.