I'm glad you brought up the A1c. I have many older adult patients with A1cs that = "pre-dabetes". Yet, I have not seen a single 70- or 80-something with one of these A1cs go on to frank diabetes, needing medication for high blood sugars, etc. These folks are worrying for nothing IMHO!
We probably ought not call it “care”, more like mechanical diagnosis and treatment. Hard to see the future income stream that is sustainable for human doctors. A partnership, AI and human, may work. Some people will always want the human interface. There’s one thing AI will never be able to do, that is be human. They can talk the talk, but never walk the walk.
Satisfaction scores depend on getting all the patients to fill out the satisfaction score questionaire. When the care is bad, giving a bad number does not provide information on why the care was bad. If you give a bad rating, no one calls to find out why you gave a bad score.
A few years ago, I underwent an endoscopy of my left ureter to work up gross bleeding.
The urologist tore my ureter with the scope and had to place a stent. In the recovery room I
had incredible pain. It felt like someone was hitting my testicle with a hammer very hard./ The pain was so bad I was biting the bed rail . screaming and vomiting. The recovery room nurse said "it is time to go home now". No pain meds were given and the surgeon was not called. Several days later. I received a satisfaction questionaire. I thought what possible good could come of this.? This is a joke.
Does Homeland Security collect satisfaction scores on people they abuse ( for protesting !)?
Absolutely love and appreciate this post. When I was the Chief Medical Officer at UHC, we eliminated prior auth and replaced it with coordinating care--REALLY coordinating care at a personal level. It worked. Readmissions went down 42% However, employer customers wanted data on our value proposition and the financial impact of what we prevented (since we couldn't measure denials--since we weren't doing that). So we transitioned to measuring how many "gaps in care" we were able to close. Yes, some are important but guess what...that became mechanical and non-personal. It became "more important" (because we could measure it) to make sure a patient got her mammogram rather than dive into the home situation that made her COPD get worse (ie a cat was nipping at the O2 tubing). I transitioned shortly thereafter.
Remarkable example, Archelle. That's a stronger example of anything I came up with. Thanks for passing that along. I'll add that many of these payers are supported by really smart, caring people who are forced into this kind of measurement and, as we see, it changes everything. Thanks for the example.
Goodhart’s Law: “When a measure becomes a target, it ceases to be a good measure.”
I'm glad you brought up the A1c. I have many older adult patients with A1cs that = "pre-dabetes". Yet, I have not seen a single 70- or 80-something with one of these A1cs go on to frank diabetes, needing medication for high blood sugars, etc. These folks are worrying for nothing IMHO!
I fear this is where AI is taking us. “I’m concerned that we’ve lost the language to defend things that matter but can’t be measured.”
THIS is the difference between AI medical care and human medical care. Thank you, Bryan, for articulating it so well.
Thank you, Tricia. The efficiencies of mechanical care are hard for alot of folks to ignore :(
We probably ought not call it “care”, more like mechanical diagnosis and treatment. Hard to see the future income stream that is sustainable for human doctors. A partnership, AI and human, may work. Some people will always want the human interface. There’s one thing AI will never be able to do, that is be human. They can talk the talk, but never walk the walk.
Satisfaction scores depend on getting all the patients to fill out the satisfaction score questionaire. When the care is bad, giving a bad number does not provide information on why the care was bad. If you give a bad rating, no one calls to find out why you gave a bad score.
A few years ago, I underwent an endoscopy of my left ureter to work up gross bleeding.
The urologist tore my ureter with the scope and had to place a stent. In the recovery room I
had incredible pain. It felt like someone was hitting my testicle with a hammer very hard./ The pain was so bad I was biting the bed rail . screaming and vomiting. The recovery room nurse said "it is time to go home now". No pain meds were given and the surgeon was not called. Several days later. I received a satisfaction questionaire. I thought what possible good could come of this.? This is a joke.
Does Homeland Security collect satisfaction scores on people they abuse ( for protesting !)?
What good would it do?
And thanks for this important post. I agree.
Absolutely love and appreciate this post. When I was the Chief Medical Officer at UHC, we eliminated prior auth and replaced it with coordinating care--REALLY coordinating care at a personal level. It worked. Readmissions went down 42% However, employer customers wanted data on our value proposition and the financial impact of what we prevented (since we couldn't measure denials--since we weren't doing that). So we transitioned to measuring how many "gaps in care" we were able to close. Yes, some are important but guess what...that became mechanical and non-personal. It became "more important" (because we could measure it) to make sure a patient got her mammogram rather than dive into the home situation that made her COPD get worse (ie a cat was nipping at the O2 tubing). I transitioned shortly thereafter.
Remarkable example, Archelle. That's a stronger example of anything I came up with. Thanks for passing that along. I'll add that many of these payers are supported by really smart, caring people who are forced into this kind of measurement and, as we see, it changes everything. Thanks for the example.