Cross-National Generative Modeling of Health and Life Trajectories

Navn på bevillingshaver

Louis Boucherie

Titel

Postdoctoral Fellow

Institution

Columbia University; Princeton University

Beløb

DKK 2,705,825

År

2026

Bevillingstype

Internationalisation Fellowships

Hvad?

This project builds an AI model that learns from the complete life histories of millions of Danes - their jobs, income, diagnoses, and prescriptions - to predict and simulate how social events shape health. For instance: how does losing a job at 50 alter the risk of heart disease a decade later? The goal is to move beyond predicting outcomes toward understanding what drives them.

Hvorfor?

Doctors focus on biology, economists on social conditions, but neither examines both, even though poverty, unemployment, and education shape disease risk. Nordic countries hold uniquely complete data, linking each person's social and medical history from birth to death, making it possible to study how life shapes health at population scale and test whether patterns hold across societies.

Hvordan?

At Columbia, I will work with health AI experts to design a model that treats a person's job history and medical record as one connected story. At Princeton, I will test it on Dutch data to check whether the patterns we find are universal or specific to Denmark. In the final year, I return to Denmark to validate the model's scenarios against real-world evidence and translate findings into policy.

Tilbage til oversigtssiden