PatientIQ - Agentic Patient 360

Team Name: Latent Potential

App Name: PatientIQ

One Liner

PatientIQ centralizes patient data and minimizes a doctor's cognitive load by making intelligent micro decisions, freeing up more space for the life saving ones made by the experts.

Abstract

Doctors today face countless daily micro decisions, many of which are administrative and pull focus away from life saving care. PatientIQ centralizes key patient information and uses AI agents to handle routine information retrieval, giving clinicians more time to focus on analysis and delivery of patient care.

Screenshots

PatientIQ Home

PatientIQ Dashboard

Run

• make dev
• make stop

Code Formatting

• Install hooks:
  • make install-hooks

Agentc

See: /agents/AGENTS.md

API

See: /docs/api

• open docs/api/index.html

Cluster Configuration

Cluster used for testing (high level):

• AWS US East
• Couchbase Server 8.0
• 5 nodes total
• Data service group: 3 nodes
• Index/Query/Search/Eventing group: 2 nodes
• MultiAZ required for AI Functions to work
• Private Networking Enabled for Workflows, Embedding Model to Work

Database Schema

Configured via environment variables:

• COUCHBASE_BUCKET (defaults to Scripps)
• COUCHBASE_RESEARCH_BUCKET (defaults to Research)

Research (bucket)

• Pubmed (scope)
  • Pulmonary (collection)
  • questions (collection)
  • answers (collection)

Scripps (bucket)

• Notes (scope)
  • Doctor (collection)
  • Patient (collection)
  • doctors_questions (collection)
  • answers_doctors (collection)
  • patient_notes_sentiment_analysis (collection)
  • patient_masked (collection)
• People (scope)
  • Doctors (collection)
  • Patients (collection)
• Wearables (scope)
  • Patient_1 (collection)
  • Patient_2 (collection)
  • Patient_3 (collection)
  • Patient_4 (collection)
  • Patient_5 (collection)
• Questionnaires (scope)
  • patient_1 (collection)
  • patient_2 (collection)
  • patient_3 (collection)
  • patient_4 (collection)
  • patient_5 (collection)
• Messages (scope)
  • Private (collection)
  • Public (collection)
• Calendar (scope)
  • Appointments (collection)

Agent Catalog (bucket)

• agent_catalog
  • metadata
  • prompts
  • tools
• agent_activity
  • logs

Cluster Cost Estimate

Window

• Feb 8–12
• Feb 12: cluster destroyed

Assumptions used in this estimate

• Weekdays (Mon–Fri): 9am–noon (3 hours/day)
• Weekend: Sat/Sun off
• Rate (on): 2.77
• Rate (off): 0.16

Weekly estimate

Category	Calculation	Weekly hours	Cost
On time	3 hours/day × 5 days	15	$41.55
Off time	(21 hours × 5 days) + (24 hours × 2 days)	153	$24.48
Total			$66.03

Projected total (3–4 weeks): $198.09