I was sent this link to a NY Times article over the weekend – Dizzying Symptoms. In it we read of a patient with an unusual set of symptoms that baffles various doctors. Eventually they find out the cause and it turns out to be a vitamin deficiency that is a not uncommon side effect of a medication being taken by the patient. So, not really usual or baffling symptoms then. Let’s be clear:
- The patient was known to be taking a drug
- Which had a known side effect
- Which caused a disease known to be linked to that side effect
- Which had a classic symptom set, 2 of which (including the one for which the disease is named) the patient had
So why, exactly, is this hard? The author of the article has a book called “Every Patient Tells a Story: Medical Mysteries and the Art of Diagnosis
The real mystery is not this patient’s disease but why doctors at a major hospital like this one still don’t have any medical decision-making systems to help them. This one didn’t even need any analytics…
Comments on this entry are closed.
Right-on James. Idiom has now embarked on automating its first clinical pathway, more on the way, also NZ Blood Service. These solutions do not automate any clinical responses, just make recommendations to the medical providers. At least this one wouldn’t have gotten away – or rather, if it did it would have only done so once. After all, Decisioning in this context is about improving diagnostic memory.
James, I love the article. It gets right to the point and I agree with you. Mark Norton, long time no see, I hope you are well.
I remember speaking with someone years ago at the Harvard Medical School about a similar idea. It was more about aiding the doctor via a small wireless device to prompt the doctor to ask questions leading to an eventual diagnosis (by the doctor). I helped build an automated diagnosis application for IP networks and one for the B2 bomber’s celestial navigation unit. I know that’s not like diagnosing humans, but it was a rules-based aide to the diagnosis.
The idea that rules could be a much larger part of healthcare automation is very interesting to me. Why not leverage technology to help keep people safe or monitoring vital signs, test-results and labs to help the caregivers do even better jobs?
Thanks Mark Eastwood, hope to see you in Alexandria in October! Since my comment above, more clinical pathways have come out of the woodwork, including a mobile delivered clinical pathway in Canada, and cancer screening here in NZ. HSA Global extends this concept by promoting pathways that hand off between contributing specialists. Also came across an NZ company (SaferSleep) developing controls to help manage anesthesia in the operating theatre. This all sounds like a great story – BDMS in support of clinical professionals, with emphasis on supporting, not replacing. Your car mechanic already does it, and as Mark Eastwood observes, a wide range of other conditions that require professional diagnosis can be BDMS assisted.
I don´t have any medical background, but I read this article also and I feel that maybe doctors are sometimes looking too hard at what could be the answer, whilst also becoming distracted by another doctors inability to find out what the disease is so they look to further abstract and weird options, instead of thinking rationally. Just my opinion!
I think its interesting how much money is poured into pharmaceuticals and for R&D in terms of procedures and drugs, but very little in making hospitals and doctors more efficient . I went to the Dr the other day and when she needed to look up a code for my perscription, she turned to a thousand of thousand page book and had to flip through to find it. Arent there excel spreadsheets for that kind of thing???