Hey,

Long-time lurker here 🙂

I’m working on a project in the UK and USA on building medical decision assistance to help doctors in hospitals with management decisions. The overarching aim is to build medical assistants that can help doctors deliver ideal medical care anywhere on the planet (or even space). These assistants will one day ensure that the clinicians taking care of us make fewer mistakes, deliver multi-specialty expert care (rather than bouncing between single-specialty doctors who can’t put everything together), understand and appraise the evidence and apply it us as individuals AND reduce the cost of medical care.

A little about me: I am a hospital physician with an engineering and technical background. I also advise the Centres for Disease Control and Prevention initiative on digitizing medicine and building clinical decision assistance.

The project is focusing initially on building assistance for stroke hyperacute thrombolysis decisions. Reasons for this choice include: it’s a binary decision (to give or not to give), has relatively few research papers to understand, access to domain knowledge, access to data and access to testing sites.

ELI5: A stroke is when an area gets damaged in the brain due to a blockage in the blood flow to that area of the brain, usually in the form of a blood clot. In some circumstances, it is possible to unblock and prevent brain damage by dissolving the clot with a strong unblocking drug (or sometimes fishing the clot out with a wire). We want to see if we can train a program that can help doctors make better-informed decisions when deciding what treatment to give the patient.

I am hoping to get together a small group to consult if anyone is interested in applying ML in the medical field. Thanks!