Duke Institute for Health Innovation

What Duke Institute for Health Innovation does

Duke Institute for Health Innovation (DIHI) is an academic innovation institute within Duke University School of Medicine, established to catalyze transformative healthcare innovations and cultivate an innovation and entrepreneurship community across Duke University and the Duke University Health System (DUHS). DIHI operates an annual Request for Applications (RFA) process open to Duke faculty, staff, students, and trainees — with priority given to AI-enabled ideas — and runs Innovation Jam, a pitch program that has awarded $395,000 across 26 pitches since 2015. It is organizationally dotted-line separate from DUHS, reporting through the Chancellor for Health Affairs and Health System CEO, allowing different success metrics from operational units.

DIHI's product portfolio consists of independently-developed clinical and operational AI tools, each validated in Duke Health's Epic/MaestroCare EHR (via FHIR) and IRB-reviewed where applicable. Flagship capabilities include Sepsis Watch (deep learning for early sepsis detection, live since November 2018), Scout (LLM-based EHR search and synthesis that reduces chart-review tasks from hours to 1-3 minutes), the Adult Deterioration Prediction model (live across Duke intermediate/stepdown units since May 2025), Duke RadChat (LLM front-desk for radiology), and the Medical Prior Authorization Assistant. The Model Facts label framework, published in Nature Digital Medicine in March 2020, was cited multiple times in the federal HTI-1 Final Rule that took effect January 1, 2025 — establishing DIHI's influence on national AI-in-healthcare disclosure standards. DIHI also hosts the Health AI Partnership (HAIP), a multi-stakeholder collaborative launched in April 2022 with Gordon and Betty Moore Foundation funding that brings together 10 health systems and 7 ecosystem partners.

DIHI is not a commercial product company; it is funded primarily through Duke institutional support (annual innovation funds from the Chancellor for Health Affairs and Dr. Victor J. Dzau) and external foundation grants (Gordon and Betty Moore Foundation; $555,056 from Patrick J. McGovern Foundation in December 2021). It does not sell products to external customers, pricing is internal to the Duke enterprise, and successful innovations are occasionally licensed or spun out — Mente, Inc. was formed to commercialize the RFID surgical instrument tracking technology. DIHI's primary "customers" are the Duke clinician/innovator community (via RFA/Hackathon) and, through HAIP, a coalition of peer health systems seeking guidance on safe, effective AI adoption.

Name

Duke Institute for Health Innovation

Legal name

Duke Institute for Health Innovation

Website

https://dihi.org

Company type

Private

Founded year

2015

Operating status

Operating

Headcount range

11–50 employees

Short description

Duke Institute for Health Innovation (DIHI) is an academic innovation institute within Duke University School of Medicine that develops and deploys AI-enabled clinical and operational tools (e.g., Sepsis Watch, Scout LLM EHR platform, Adult Deterioration Prediction) across Duke Health, and runs the Health AI Partnership coalition of 10 health systems.

Ownership category

akta.pro rank

Product category

Healthcare AI Innovation Services

NAICS

Scientific Research and Development Services (5417), Research and Development in the Physical, Engineering, and Life Sciences (except Nanotechnology and Biotechnology) (541715)

SIC

Services-Health Services (8000), Services-Commercial Physical & Biological Research (8731), Pharmaceutical Preparations (2834)

akta.pro industry

PRIMARYEDAHAGAI

Biomedical Sciences & Medical Research Institutes (University-Based)

Education/Higher Education (Colleges & Universities)/Medical & Health Sciences Universities

SECONDARYEDAHAGAJ

Health Informatics & Digital Health Programs

Education/Higher Education (Colleges & Universities)/Medical & Health Sciences Universities

SECONDARYFSANAJAI

Healthcare Access & Life Sciences Impact Funds

Financial & Insurance/Private Equity & Venture Capital/Impact / ESG & Sustainable Investment Funds

Keywords

• Healthcare AI solutions
• Clinical decision support
• Health innovation services
• Digital health implementation
• Health AI partnerships

Headquarters

HQ city

Durham

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Services

Cost components

Personnel, Technology or R&D, Operations, Infrastructure, Others

Revenue model

1. Institutional Funding and Grants: DIHI is funded through Duke University's Chancellor for Health Affairs and an annual innovation fund from Dr. Victor J. Dzau. The institute receives support from Duke University Health System (DUHS) for operational activities.
2. Foundation Funding: The Health AI Partnership (HAIP) was launched in April 2022 with initial funding from Gordon and Betty Moore Foundation to facilitate safe, effective, and responsible diffusion of health AI software.

Go-to-market motion1 record

Distribution channels3 records

Marketing channels6 records

Core offering

Duke Institute for Health Innovation (DIHI) is an academic innovation institute within Duke University School of Medicine that develops and implements artificial intelligence and digital health solutions for clinical care. It partners with Duke University Health System (DUHS) to design, prototype, validate, and scale AI-enabled tools—such as early-warning systems, clinical workflow assistants, and AI governance frameworks—that are deployed inside the health system and, in some cases, shared with external healthcare partners through initiatives such as the Health AI Partnership (HAIP).

Product overview

Duke Institute for Health Innovation (DIHI) is an academic innovation institute that develops and implements healthcare AI solutions rather than a commercial product company. Its portfolio consists of distinct, independently-developed projects and tools, each addressing specific clinical or operational challenges. Key products include: Scout (LLM-based EHR search and synthesis platform); Sepsis Watch (deep learning sepsis early warning system); Adult Deterioration Prediction (ML-based deterioration alert system); Duke Chest Pain Assessment Tool/CPAT (FHIR-integrated chest pain decision support); RFID Surgical Instrument Tracking (OR efficiency optimization); Duke RadChat (LLM-based radiology triage assistant); Medical Prior Authorization Assistant (AI automation of prior authorization workflows); and Kidney Transplant Matching AI (predictive graft outcome modeling). DIHI also manages the Health AI Partnership (HAIP), a multi-stakeholder learning network for healthcare AI adoption, and publishes the Model Facts label for AI transparency in healthcare.

Differentiator

Problem solved

Functional benefit

Brands

• Health AI Partnership (HAIP): The first multi-stakeholder collaborative exclusively focused on empowering healthcare delivery systems to adopt AI safely, effectively, and equitably. Launched in April 2022 with initial funding from Gordon and Betty Moore Foundation.
• Sepsis Watch

Products and services

• Sepsis Watch An AI-enabled early-warning system developed with Duke University Health System to detect early signs of sepsis in hospitalized patients and surface alerts to clinical teams for timely intervention.
• Scout A clinical AI solution developed by DIHI to support care teams inside the Duke University Health System through data-driven insights.
• Duke Chest Pain Assessment Tool (CPAT) A digital assessment tool designed to standardize evaluation of chest pain patients and support emergency-department clinicians at Duke.
• RFID Surgical Instrument Tracking A radio-frequency identification system for tracking surgical instruments through operative workflows to improve safety, accuracy, and inventory visibility.
• Adult Deterioration Prediction Solution A predictive analytics solution designed to identify early signs of clinical deterioration in hospitalized adults and prompt escalation by care teams.
• Duke RadChat A conversational AI assistant built to support radiologists in summarizing and interacting with imaging reports and related clinical data.
• Medical Prior Authorization Assistant (MPAA) An AI tool designed to streamline and automate medical prior authorization workflows for clinicians and administrative staff.
• AI-Assisted Pediatric Psychiatry Handoff System An AI-assisted solution to support handoffs between clinicians in pediatric psychiatry settings, improving continuity and quality of care.
• Transplant Evaluation AI An AI tool that supports transplant program staff in evaluating candidate information to inform transplant decisions.
• Kidney Transplant Donor-Recipient Matching AI An AI-driven matching solution to help transplant teams identify optimal kidney donor-recipient pairings.
• Predict Behavioral Emergencies in Hospital A predictive model designed to identify inpatients at elevated risk of behavioral emergencies so clinical and security teams can intervene proactively.
• STOP-HAI A DIHI solution focused on reducing healthcare-associated infections through workflow, surveillance, or AI-enabled interventions.

Quantifiable outcome

• Sepsis Watch integrated into routine clinical care in November 2018, reshaping how local machine learning projects are executed
• +3 more outcomes

Named customers1 record

Segments2 records

Ideal customer profiles1 record

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration3 records

AI capability9 records

Feature6 records

Partnerships

Five partnerships are on record, tiered core and supporting.

• Health AI Partnership (HAIP) - Health System PartnerscoreStrategic or Co-development PartnerMulti-stakeholder collaborative including Duke Health, Hackensack Meridian Health, Jefferson Health, Kaiser Permanente, Mayo Clinic, NewYork-Presbyterian, OCHIN, PCCI, UCSF, and University of Michigan focused on safe and equitable AI adoption in healthcare delivery.
• Health AI Partnership (HAIP) - Ecosystem PartnerssupportingStrategic or Co-development PartnerEcosystem partners include American Medical Association, DLA Piper, Gordon and Betty Moore Foundation, IDEO.org, Patrick J McGovern Foundation, PLOS Digital Health, and UC Berkeley providing expertise across healthcare, policy, ethics, and technology domains.
• Mente, IncsupportingOEM/ Whitelabel/ Licensing PartnerCompany formed to translate RFID surgical instrument tracking technology developed by DIHI. License for the technology is currently in negotiation.
• Duke UniversitycoreStrategic or Co-development PartnerDIHI is organizationally connected to Duke University, gaining from the freedom to research, teach, and learn, and aligns with the three-part academic mission to advance healthcare, research, and education.
• Duke University Health System (DUHS)coreStrategic or Co-development PartnerDIHI is dotted-line separate from DUHS allowing for different success metrics. Benefits from IT and healthcare knowledge of DUHS. Originally funded directly by Duke's Chancellor for Health Affairs with direct line from Health System CEO.

Scale indicators5 records

Recent moves7 records

Expansion highlights6 records

Direct peers

• Stanford Center for Clinical Informatics: Stanford School of Medicine's clinical informatics group similarly translates AI/ML research into Epic-integrated clinical tools inside an academic medical center. Closest direct comparison in mission, governance, and operating model to DIHI.
• UCSF Bakar Computational Health Sciences Institute: UCSF's institute for computational health sciences develops EHR-integrated AI and predictive models within an academic medical center. HAIP partner with near-identical model to DIHI of bridging research and clinical implementation.
• Geisinger Steele Institute for Health Innovation: Geisinger's innovation institute translates research into deployed clinical AI and operations tools across an integrated delivery network. Closely matches DIHI's mission of converting academic research into live clinical workflows.
• Johns Hopkins inHealth: Johns Hopkins Medicine's precision medicine and AI initiative integrates data science into clinical care across the health system. Analogous academic-medical-center vehicle for AI deployment at scale.
• Vanderbilt Department of Biomedical Informatics: Vanderbilt's BMI group pioneered EHR-based AI (e.g., My Health Team at Vanderbilt, prior authorization tools) and is widely cited as the model for academic clinical AI. Strong functional peer on NLP, predictive models, and clinical decision support.
• Northwestern Center for AI in Healthcare (AIM-HI): Northwestern's AI in Healthcare center partners with Northwestern Medicine to translate ML research into clinical practice. Direct structural and operational peer to DIHI's model.

Broad incumbents

• Mayo Clinic Platform: Mayo Clinic's enterprise platform for AI-enabled care and biopharma partnerships. HAIP partner to DIHI; comparable in deploying clinical AI inside a large health system but operates at much larger scale and scope.
• Mass General Brigham Data Science Office: MGB's enterprise data science and clinical AI organization operates at the scale of a multi-hospital AMC. Comparable in deploying validated clinical AI into Epic workflows; broader footprint and budget than DIHI.

Emerging players

• Cohere Health: Commercial AI platform focused on prior authorization and utilization management, directly overlapping with DIHI's Medical Prior Authorization Assistant (Ocrevus, Botox, Pluvicto). Comparable use case but as a venture-backed vendor rather than an academic institute.
• LeanTaas: Applied AI platform for hospital operations, capacity, and surgical optimization (e.g., iQueue). Comparable to DIHI's RFID OR optimization and operational AI portfolio, but as a commercial vendor competing for the same health system budgets.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xx.com/dukeinnovate
• YouTubeyoutube.com/channel/UCIT40hRUoV03dC8eigVC4fg
• LinkedInlinkedin.com/company/dukeinnovate

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Duke Institute for Health Innovation