Episode 35: Johannes Cordier on Machine Learning, Sepsis Detection, and Hospital Ward Allocation

This episode features Johannes Cordier, a visiting PhD student from the University of St. Gallen at Erasmus University Rotterdam.

Johannes discusses two papers using Swiss hospital administrative data. The first examines how machine learning methods can identify under-reported sepsis cases, shedding light on how coding practices and reimbursement incentives affect data quality. The second studies how patient placement across hospital wards influences mortality, combining causal machine learning with decision-theoretic approaches to account for capacity constraints.

The episode concludes with reflections on research visits, academic conferences, and practical advice for PhD students navigating research careers and industry pathways.

Guest: Johannes Cordier, Visiting PhD Student (University of St. Gallen)

Hosts: Fanny Tallgren, Prithviraj Basu Mallik

Timestamps: (01:50) Background and research focus (03:40) Sepsis under-reporting and admin data (11:35) Machine learning methods (17:40) Patient placement and mortality (31:20) Research visits and PhD career advice