AI excellent at predicting death over last four years of life

{{ A group of Danish and American social scientists announced that with artificial intelligence and a huge data set, they were able to predict the likelihood of a person’s death within the next four years with startling accuracy. “Using Sequences of Life-Events to Predict Human Lives,” published in Nature Computational Science in December, was like a Christmas gift from Blumhouse. The internet exploded with reports of a super-creepy AIdoom calculator.” Overnight, the paper became the most famous Danish rumination on mortality since “To be or not to be.” }}

Article in Nature

https://www.nature.com/articles/s43588-023-00573-5.epdf?sharing_token=Cu7sh-JyxsO2dkd_5Sb2lNRgN0jAjWel9jnR3ZoTv0MCA8ZigxRuse4dwzhHzdeU_N0IihJNSnkNn8IuSOiDqLqG_etSzzPgqS52p7BSOQC2HAz4MvM3f8lrOkBV9oIoDdZ2sBEL7va0j0st3SeHro-oMwKpuQipR0pD6_Ntiv3tMOgHM9dj6PbsjfTtFEvEZyOU0uB7lCJ2gBGaah6ktYfM2IJvttvKsia3YZt6UeI%3D&tracking_referrer=www.washingtonpost.com

intercst

5 Likes

I skimmed the article and understand just enough to get the sense that I was reading science fiction.

1 Like

Countries that collect data from their universal healthcare systems produce research that is highly reliable because of the massive data sets that include the entire population. When n > a million the incidence of rare events happens enough to be analyzed statistically.

The power of large data sets can’t be matched in the U.S. because our fragmented healthcare system forces researchers to use relatively small numbers of people in their studies, usually in the hundreds and sometimes even less. Only the strongest signals can be teased out of noise with reasonable significance.

This is why I do medical research in the National Library of Medicine database, which includes journals from all over the world. I found data on the prognosis for bilateral breast cancer in research from Sweden.

I would like to see AI applied to the Nurses’ Health Study, which began in 1976 with 400,000 participants and is now in its third generation.

Wendy

8 Likes

What I got out of the article was that the most reliable predictor of death was cancer.

intercst

2 Likes

China is the leader in big data in the medical field. Gov’t has free access to the nation’s medical records and uses it to train and develop AI applications. Nothing like HIPPA in China.

China’s data-driven dream to overhaul health care (nature.com)

intercst

1 Like

And, back in 2000 I got my chemo doc to go with carboplatin rather than cisplatin due to a recent French study I found. Fewer side-effects, I thought. Had to go through medical library journals by hand to find it though. Different times.

Still alive so far.

-IGU-

8 Likes

That is good to hear.

We should set up an after-life board on the fool. LOL

1 Like

I did the same for my mother’s lung cancer in 2001. Had to fight the oncologist at Memorial Sloan Kettering who was using his (unsubstantiated) recipe with a different chemo agent. I will never forget looking him in the eye and saying, “How many patients have you studied? 50? 100? 500?” He backed down PDQ. It’s not easy to look me in the eye when I’m defending my mother.

Now carboplatin is the standard of care.
Wendy

5 Likes

Dad had a similar experience but he was a doctor’s doctor.

His heart rate slowed with age by 75. He blacked out while driving and went into a tree. He survived with two broken wrists and a cut on his knee down to the bone.

In the cardiac wing of the hospital the head of the department saw to him. The first day ran in and out of the room. Dad was angry.

The second day was going to be the same thing in and out of the room with nothing to say. Dad yelled get in here and sit down. You do not want to be here. The head of the department looked sheepish. Dad fired him on his second day. My BIL in Boston called a good friend a top cardiologist in Boston who recommended people he had trained in Boston who were working in the same hospital my dad was being treated in Hartford, CT.

For some reason there was a vote in cardiology for a pacemaker. The head of the department was against placing a pacemaker in my dad. He was outvoted by all the younger doctors. Dad got his pacemaker.

Dad called the head of the hospital emergency room on the third day. Dad had been in professional contact with him for close to two decades. That doctor confirmed that the head of cardiology had been a problem in the hospital for years.

Inside baseball is interesting.

If you have to run over your doctor do it. His or her feelings certainly can take it. Feel free to fire a doctor as well. But knowing where to turn next is hard for many people.

3 Likes