OpenEvidence Cost Tuning

Table of Contents

• Overview
• Prerequisites
• Instructions
• Output
• Error Handling
• Examples
• Resources

Overview

Optimize OpenEvidence API costs while maintaining clinical decision support quality. Covers usage tracking, DeepConsult optimization, user quotas, and cost reporting.

Prerequisites

• OpenEvidence billing dashboard access
• Usage metrics configured
• Understanding of pricing model
• Budget authority

Pricing Model

Instructions

Step 1: Implement Usage Tracking

Create UsageTracker with Redis-backed daily/monthly counters per user and global. Track queries and DeepConsults separately with budget alerts at configurable thresholds.

Step 2: Optimize DeepConsult Usage

Build decision logic: use regular queries for simple questions (dosage, half-life, contraindications, side effects). Reserve DeepConsult for complex research (treatment comparisons, systematic reviews, emerging therapies). Cache DeepConsult results for 1 week.

Step 3: Implement User Quotas

Define tier-based quotas (free: 10/day, professional: 100/day + 10 DeepConsults/month, enterprise: 1000/day + 100 DeepConsults/month). Enforce with middleware.

Step 4: Build Cost Reporting

Generate periodic reports with breakdowns by specialty, user, and day. Include automated recommendations for high DeepConsult ratio, low cache hits, and concentrated usage.

Output

• Usage tracking with Redis counters
• Budget alert system with configurable thresholds
• Smart DeepConsult decision logic
• Tier-based user quota enforcement
• Cost reporting dashboard with recommendations

Error Handling

Examples

DeepConsult Decision

// Simple question -> regular query (free)
"What is the dosage of metformin?" -> clinicalQuery

// Complex research -> DeepConsult (premium)
"Compare emerging therapies for treatment-resistant depression" -> deepConsult

See detailed implementation for advanced patterns.

Resources

• OpenEvidence Pricing
• OpenEvidence API Terms