Economic Evaluation of AI-enabled Early Metastasis Detection in Germany (MA)

This is a project offered in cooperation with the RWTH Aachen. For further information reach out to Lasse Falk.

Background and motivation:

Metastatic progression accounts for the majority of cancer mortality and cost. Emerging AI systems promise earlier, risk-stratified detection and monitoring across imaging, pathology, and routine clinical data. To inform adoption, decision-makers require robust, yet tractable, evidence on cost-effectiveness under German health-system conditions.

Within the BMBF-funded DECIPHER-M consortium, this thesis will focus on the development and analysis of a decision-analytic model for one clinical pathway of AI-enabled early metastasis detection.

Thesis aims:

The thesis will concentrate on a health-economic evaluation:

1. Develop a transparent state-transition (Markov) model capturing key pre-metastatic and metastatic health states, detection pathways, treatment initiation, and survival for one selected DECIPHER-M use case under German statutory health insurance (SHI) conditions.
2. Estimate clinical and economic outcomes for an AI-guided strategy versus standard care, including costs, QALYs, and an incremental cost-effectiveness ratio (ICER per QALY) from the SHI perspective (a provider perspective can be explored if time allows).
3. Conduct uncertainty analyses with:
   - deterministic sensitivity analyses as a core requirement, and
   - probabilistic sensitivity analysis if feasible within the available timeframe.

Supervision and environment:

You will work at the interface of oncology, AI, and health economics in close collaboration with clinical and data-science teams. The thesis is embedded in the DECIPHER-M consortium with specific support from Prof. Dr. med. Carolin Schneider (RWTH Aachen). Prof. Jonas Schreyögg (Universität Hamburg) will advise from the economic side. Secure compute environments and version-controlled workflows are standard, and you will have access to predefined modelling templates in R.

Candidate profile:

• Master’s student in Health Economics, Public Health, Statistics, Data Science, or a related field.
• Basic proficiency in R (e.g. data handling, simple scripts, basic statistics) and a willingness to apply R to decision modelling; prior experience with cost-effectiveness analysis is an advantage but not a prerequisite.
• Python skills are helpful but not required; the thesis can be completed fully in R.
• Interest in decision-analytic modelling and health technology assessment, and willingness to deepen methodological skills during the thesis.
• Excellent ability to communicate with multidisciplinary stakeholders.

What we offer:

• A policy-relevant topic with direct applicability to AI adoption in oncology in a leading national consortium.
• Access to real-world data pipelines, parameter sources, and mentorship in decision-analytic modelling.
• Opportunity to co-author a methods or policy-oriented output alongside the thesis, depending on progress and interest.

Deliverables:

• A documented and reproducible state-transition model in R, including model structure, assumptions, and parameter sources.
• Base-case cost-effectiveness results comparing AI-enabled early metastasis detection with standard care, with deterministic sensitivity analyses as a minimum.
• A Master thesis that clearly reports methods, results, and limitations, and a short, stakeholder-oriented summary highlighting implications for decision-makers.

Application:

Please submit:
(i) CV,
(ii) transcript of records,
(iii) a ≤300-word statement outlining your motivation, relevant prior methods training, and your preferred modelling approach (and why), and
(iv) if available, one code sample (preferably in R; Python also welcome).