Research programmes
Each one starts from a problem somebody already had. Each page says what exists, what it does not do, and what has not been established, because a reader cannot tell those apart from the outside.
The stage label is the real one. One programme is published, one is waiting on a regional approval and is not in use in care, and one has nothing built yet.
Cancer risk after an inflammatory disease diagnosis
356,022 patients across five years of Italian hospital discharge records. Which cancers, how much more often, and when in the disease course.
Read the case study →Which subgroups carry which cancer risk
Genotype, lifestyle and socioeconomic data read together at population level. One of the algorithms delivered under CODIGE. Population associations, never an individual risk score.
Read the case study →An agent that prepares the evidence for a tumour board
A clinician enters a case, the agent searches published AIOM and ESMO guidance and returns one page the panel can read in the room. Built with an Italian local health authority. The pilot is waiting on regional approval and the agent is not in use in care.
Read the case study →An agent for the gap between rheumatology visits
Months pass between outpatient visits, and every visit begins from a reconstruction by memory. What is worth capturing in between, and does capturing it change the next one.
Read the case study →A compass for hospital medicines spending
A hospital already holds every figure it needs to see where value is leaking. It is not organised in a way that points anywhere. This reads the flows an authority already produces and points at the molecule, the channel and the ward where the money is going.
Read the case study →A pipeline for patient groups too small to analyse
When a disease affects a handful of people in a country, the statistics built for thousands stop working. We are testing what can still be learned from a cohort that small, starting with propionic acidemia in Saudi Arabia.
Read the case study →Bringing us a problem
Every programme above began with somebody who owned the problem and could not solve it. If that is you, the useful first message is the question itself rather than a proposal.