We had something similar with my father in law who just passed from pancreatic cancer.

The doctors each saw one small part of the case, he was being passed between multiple doctors and NPs. There was no coherence. At one point he ended up on so many medications that made no sense he passed out. Only at the hospital when a doctor looked at everything together did they realize what a mess had been created and got him back on his feet.

I ended up giving all of his records to an agent and making it the coordinator. It found a bunch of problems. He was being given the wrong pain medication on the wrong schedule, the moment that was fixed he felt a lot better (unfortunately he lost so much weight on the wrong schedule it's what got him in the end). They also use voice transcription software, which led to a key error about what genetic markers he had. Fixing this changed his treatment plan!

In general, I wish we had been much more aggressive with this. Last time we had a major medical issue an NP almost killed my wife. At the time there was no AI, so we just had to recognize things were in decline and start being aggressive about getting in with the right doctors (the moment we found the right one she was diagnosed, had surgery and was cured in weeks).

He started doing better when the AI became the central coordinator and doctors/tests/etc. became their inputs. It was an excellent advocate. And while we have science backgrounds, we aren't doctors. But it could present evidence for actions in ways that the doctors appreciated (his oncologist was impressed and called the AI a member of the team).

Unfortunately his physical reserves were very drained at that point. He probably would have had another 6-12 months with his family if it wasn't for the disorganized mess of a care system that we have.