Abundance Decision Systems — back to topAbundance DS — back to topContact

AI Systems for Health Economics & Outcomes Research

We advance AI capabilities in HEOR by developing new tools, workflows and evaluation methods.

Contact us
Explore our approach

Approach

Diagnosis

AI’s potential in HEOR remains largely unrealised.

AI is often applied within workflows designed for a different era, with little adaptation to its capabilities and limitations and without rigorous evaluation.

Mission

We work to raise the quality and speed of HEOR delivery as AI capabilities advance.

We build AI workflows, redesign existing systems around AI capabilities, and develop evaluation frameworks. Our solutions are built and operated with rigorous testing and expert human oversight.

Services

  • Strategy and training: advice, workshops and practical support for AI adoption
  • Evaluation: assessment of AI solutions in realistic HEOR environments
  • Custom development: AI tools and workflows, built from scratch or based on completed work samples
  • HEOR project delivery: end-to-end delivery using AI workflows and human expert review

Call for collaboration

We are seeking research partners to collaborate on the development of AI workflows through replay pilots.

A replay uses the inputs and outputs of completed work as the basis for a new AI workflow. We then test the system on a separate (holdout) case and evaluate it against agreed quality criteria.

Discuss a replay pilot

Initiatives

Organising knowledge for AI

Today’s workspaces are designed for people. Knowledge is scattered across documents, spreadsheets and applications, making it difficult for AI agents to retrieve relevant context.

Autonomous work needs knowledge to be reorganised for machines. Graphs give agents connected information they can query, trace and act on.

More knowledge graphsNICE-Graph ↗AMNOG-Graph ↗

EQ-Graph

A graph of EuroQol-funded research

HEOR Bench

Currently, AI can support HEOR work only under close human guidance and review. Greater autonomy requires evidence of which tasks AI systems can perform reliably and where human intervention remains necessary.

Rigorous evaluations are the foundation for comparing existing AI solutions and building new performant, continuously improving AI workflows.

Read the briefing
Task: Extracting outcome data — HEOR BenchThree source publications feed a structured record with endpoint, value, provenance and uncertainty. The record is graded one criterion at a time. Checkmarks indicate that each illustrated criterion passes; the list continues beyond the frame.Task: Extracting outcome datasource documentssource coordinatesstructured outputEndpointValueProvenanceUncertaintyGrading specificationResultTreatment arm and estimandValue and uncertaintyPopulation and time pointProvenanceDerived calculation

mim devtools series

A set of open source tools for AI-native knowledge work we developed and use internally.

mimir

AI workspace for teams

mim dictate

Local voice-to-text AI

mim awake

Keeps your computer awake while agents work

mim review

MCP for robust AI peer review

mim clip

Encrypted clipboard history for macOS

mim ultra

Experimental AI-native Linux environment

Team

Scientific advisors

Heike Adel-Vu

Professor of AI, Stuttgart Media University

Profile

Sören Auer

Director of TIB, Professor at Leibniz University Hannover

Profile

Contact

We welcome enquiries about AI in HEOR, project opportunities and research collaboration.

contact@abundanceds.com

Street address

Lohmühlenstraße 65
12435 Berlin, Germany
Berlin — Lohmühlenstraße 65. Open Google Maps in a new tab.
© OpenStreetMap

Supported by

Federal Ministry
for Economic Affairs
and Energy

Co-funded by the
European Union

EXIST, from science to business

Abundance Decision Systems receives the EXIST Startup Grant from the Federal Ministry for Economic Affairs and Energy, co-funded by the European Union through the ESF Plus.