AGENT SKILL

openevidence-core-workflow-a

Execute OpenEvidence clinical query workflow for point-of-care decisions. Use when implementing real-time clinical decision support, building EHR-integrated evidence lookups, or point-of-care queries. Trigger with phrases like "openevidence clinical query", "point of care", "quick clinical lookup", "evidence search".

License 许可
MIT
最近更新
2026-11-15
Stars
★ 2.0k
Forks
233

01 前言 Front Matter

name
openevidence-core-workflow-a
author
Jeremy Longshore <jeremy@intentsolutions.io>
license
MIT
version
1.0.0
description
Execute OpenEvidence clinical query workflow for point-of-care decisions. Use when implementing real-time clinical decision support, building EHR-integrated evidence lookups, or point-of-care queries. Trigger with phrases like "openevidence clinical query", "point of care", "quick clinical lookup", "evidence search".
allowed-tools
Read, Write, Edit, Bash(npm:*), Grep

03 SKILL.md 正文

OpenEvidence Core Workflow A: Clinical Query

Overview

Primary workflow for real-time clinical queries at the point of care. Returns evidence-based answers in 5-10 seconds with peer-reviewed citations.

Prerequisites

  • Completed openevidence-install-auth setup
  • Understanding of clinical decision support patterns
  • Valid API credentials configured

Use Cases

  • Drug interaction checks during prescribing
  • Treatment protocol lookups
  • Differential diagnosis support
  • Dosing verification
  • Clinical guideline queries

Instructions

Step 1: Structure the Clinical Query

// src/workflows/clinical-query.ts
import { OpenEvidenceClient } from '@openevidence/sdk';

interface ClinicalQueryRequest {
  question: string;
  specialty: string;
  urgency: 'stat' | 'urgent' | 'routine';
  patientContext?: {
    age?: number;
    sex?: 'male' | 'female';
    conditions?: string[];
    medications?: string[];
  };
}

interface ClinicalQueryResponse {
  answer: string;
  citations: Citation[];
  confidence: number;
  responseTimeMs: number;
  queryId: string;
}

interface Citation {
  source: string;
  title: string;
  year: number;
  doi?: string;
  guideline?: boolean;
}

Step 2: Implement Query Service

// src/services/point-of-care-query.ts
import { OpenEvidenceClient } from '@openevidence/sdk';

const client = new OpenEvidenceClient({
  apiKey: process.env.OPENEVIDENCE_API_KEY,
  orgId: process.env.OPENEVIDENCE_ORG_ID,
  timeout: 15000, // 15 second timeout for point-of-care
});

export async function queryAtPointOfCare(
  request: ClinicalQueryRequest
): Promise<ClinicalQueryResponse> {
  const startTime = Date.now();

  const response = await client.query({
    question: request.question,
    context: {
      specialty: request.specialty,
      urgency: request.urgency,
      ...(request.patientContext && {
        patientAge: request.patientContext.age,
        patientSex: request.patientContext.sex,
        relevantConditions: request.patientContext.conditions,
        currentMedications: request.patientContext.medications,
      }),
    },
    options: {
      maxCitations: 5,
      includeGuidelines: true,
      prioritizeRecent: true, // Prefer evidence from last 3 years
    },
  });

  return {
    answer: response.answer,
    citations: response.citations.map(c => ({
      source: c.source,
      title: c.title,
      year: c.year,
      doi: c.doi,
      guideline: c.type === 'guideline',
    })),
    confidence: response.confidence,
    responseTimeMs: Date.now() - startTime,
    queryId: response.id,
  };
}

Step 3: Drug Interaction Check Example

// src/workflows/drug-interaction.ts
export async function checkDrugInteraction(
  drug1: string,
  drug2: string,
  patientContext?: { age?: number; conditions?: string[] }
): Promise<{
  hasInteraction: boolean;
  severity: 'major' | 'moderate' | 'minor' | 'none';
  details: string;
  citations: Citation[];
}> {
  const response = await queryAtPointOfCare({
    question: `What are the drug interactions between ${drug1} and ${drug2}?`,
    specialty: 'pharmacology',
    urgency: 'urgent',
    patientContext,
  });

  // Parse severity from response
  const severity = determineSeverity(response.answer);

  return {
    hasInteraction: severity !== 'none',
    severity,
    details: response.answer,
    citations: response.citations,
  };
}

function determineSeverity(answer: string): 'major' | 'moderate' | 'minor' | 'none' {
  const lower = answer.toLowerCase();
  if (lower.includes('contraindicated') || lower.includes('major interaction')) return 'major';
  if (lower.includes('moderate interaction') || lower.includes('caution')) return 'moderate';
  if (lower.includes('minor interaction')) return 'minor';
  if (lower.includes('no significant interaction') || lower.includes('no known interaction')) return 'none';
  return 'moderate'; // Default to moderate if unclear
}

Step 4: EHR Integration Pattern

// src/integrations/ehr-hook.ts
import { queryAtPointOfCare } from '../services/point-of-care-query';

// HL7 FHIR CDS Hooks integration
interface CDSRequest {
  hook: string;
  hookInstance: string;
  context: {
    patientId: string;
    encounterId?: string;
    medications?: any[];
  };
}

interface CDSResponse {
  cards: CDSCard[];
}

interface CDSCard {
  summary: string;
  detail: string;
  indicator: 'info' | 'warning' | 'critical';
  source: { label: string; url?: string };
  suggestions?: any[];
}

export async function handleCDSHook(request: CDSRequest): Promise<CDSResponse> {
  // Extract clinical context from FHIR resources
  const medications = request.context.medications?.map(m => m.medicationCodeableConcept?.text) || [];

  // Query OpenEvidence for relevant clinical information
  const evidence = await queryAtPointOfCare({
    question: buildClinicalQuestion(request.hook, medications),
    specialty: 'family-medicine',
    urgency: 'routine',
    patientContext: {
      medications,
    },
  });

  return {
    cards: [{
      summary: 'Clinical Evidence Available',
      detail: evidence.answer,
      indicator: evidence.confidence > 0.9 ? 'info' : 'warning',
      source: {
        label: 'OpenEvidence',
        url: 'https://openevidence.com',
      },
    }],
  };
}

function buildClinicalQuestion(hook: string, medications: string[]): string {
  switch (hook) {
    case 'medication-prescribe':
      return `Are there any drug interactions or contraindications for ${medications.join(', ')}?`;
    case 'order-sign':
      return `What are the clinical considerations for prescribing ${medications.join(', ')}?`;
    default:
      return `Provide clinical guidance for patient on ${medications.join(', ')}`;
  }
}

Output

  • Real-time clinical query response (5-10 seconds)
  • Evidence-based answer with peer-reviewed citations
  • Confidence score for clinical decision support
  • Query audit trail for compliance

Error Handling

Error Cause Solution
Timeout Complex query or network Increase timeout, simplify question
Low confidence Ambiguous query Rephrase with more specific context
No citations Rare condition Consider DeepConsult for deeper research
Rate limit Too many queries Implement request queuing

Performance Considerations

  • Target response time: < 10 seconds for point-of-care
  • Cache frequent queries (drug info, guidelines)
  • Pre-warm connections during low-traffic periods
  • Use streaming responses for faster perceived performance

Examples

Complete Point-of-Care Integration

// Example: Emergency department workflow
async function edClinicalSupport(chiefComplaint: string, vitals: any) {
  const queries = await Promise.all([
    queryAtPointOfCare({
      question: `What is the differential diagnosis for ${chiefComplaint}?`,
      specialty: 'emergency-medicine',
      urgency: 'stat',
    }),
    queryAtPointOfCare({
      question: `What workup is recommended for ${chiefComplaint}?`,
      specialty: 'emergency-medicine',
      urgency: 'stat',
    }),
  ]);

  return {
    differential: queries[0],
    workup: queries[1],
  };
}

Resources

Next Steps

For comprehensive research queries, see openevidence-core-workflow-b (DeepConsult).

04 同仓库更多 Skills

Dicklesworthstone
Dicklesworthstone/pi_agent_rust

theme-factory

Toolkit for styling artifacts with a theme. These artifacts can be slides, docs, reportings, HTML landing pages, etc. There are 10 pre-set themes with colors/fonts that you can apply to any artifact that has been creating, or can generate a new theme on-the-fly.

视觉设计
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

web-artifacts-builder

Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.

前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

测试与质量前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

neuropixels-analysis

Neuropixels neural recording analysis. Load SpikeGLX/OpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen/IBL curation, AI-assisted visual analysis, for Neuropixels 1.0/2.0 extracellular electrophysiology. Use when working with neural recordings, spike sorting, extracellular electrophysiology, or when the user mentions Neuropixels, SpikeGLX, Open Ephys, Kilosort, quality metrics, or unit curation.

★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

skill-creator

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.

Skill与智能体开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

notebooklm

Use this skill to query your Google NotebookLM notebooks directly from Claude Code for source-grounded, citation-backed answers from Gemini. Browser automation, library management, persistent auth. Drastically reduced hallucinations through document-only responses.

Skill与智能体开发
★ 2.0k