Bob Wachter
@bobwachter.bsky.social
Chair, UCSF Dept of Medicine. Career: What happens when a poli sci major becomes an academic physician. Author: "The Digital Doctor" and upcoming "A Giant Leap: How AI is Transforming Healthcare – and What it Means for Our Future." Hubby/Dad/Grandpa/Golfer
Precision Medicine – making clinical decisions based on nuanced analyses of patients' personal characteristics, including genetics – has been "just around the corner" for >30 years. In my new Substack, I describe why AI is poised to deliver on the promise.
robertwachter.substack.com/p/can-ai-fin...
robertwachter.substack.com/p/can-ai-fin...
Can AI Finally Deliver on Precision Medicine’s Promise?
Yes, But Only If We Nail Both the Biology and Decision Support
robertwachter.substack.com
November 6, 2025 at 4:27 PM
Precision Medicine – making clinical decisions based on nuanced analyses of patients' personal characteristics, including genetics – has been "just around the corner" for >30 years. In my new Substack, I describe why AI is poised to deliver on the promise.
robertwachter.substack.com/p/can-ai-fin...
robertwachter.substack.com/p/can-ai-fin...
Signing off from another great UCSF Management of the Hospitalized Patient conference. #UCSFMHP2025 Thanks to wonderful speakers and 500 very engaged participants!
See you next year in SF, October 15-17. 2026.
See you next year in SF, October 15-17. 2026.
November 1, 2025 at 7:31 PM
Signing off from another great UCSF Management of the Hospitalized Patient conference. #UCSFMHP2025 Thanks to wonderful speakers and 500 very engaged participants!
See you next year in SF, October 15-17. 2026.
See you next year in SF, October 15-17. 2026.
"Hospital Medicine Updates" #UCSFMHP2025, by "The Brads" (Sharpe/Monash):
1) Pyelo (even w/ bacteremia): 5-7d Rx equal to 10-14
2) Rigors: Good predictor of bacteremia
3) Acute back pain: Muscle relaxants work
4) Low Na: fast correction (~10meq/d) better
5) Pts w/ CHF & low-Fe anemia: IV iron works
1) Pyelo (even w/ bacteremia): 5-7d Rx equal to 10-14
2) Rigors: Good predictor of bacteremia
3) Acute back pain: Muscle relaxants work
4) Low Na: fast correction (~10meq/d) better
5) Pts w/ CHF & low-Fe anemia: IV iron works
November 1, 2025 at 7:20 PM
"Hospital Medicine Updates" #UCSFMHP2025, by "The Brads" (Sharpe/Monash):
1) Pyelo (even w/ bacteremia): 5-7d Rx equal to 10-14
2) Rigors: Good predictor of bacteremia
3) Acute back pain: Muscle relaxants work
4) Low Na: fast correction (~10meq/d) better
5) Pts w/ CHF & low-Fe anemia: IV iron works
1) Pyelo (even w/ bacteremia): 5-7d Rx equal to 10-14
2) Rigors: Good predictor of bacteremia
3) Acute back pain: Muscle relaxants work
4) Low Na: fast correction (~10meq/d) better
5) Pts w/ CHF & low-Fe anemia: IV iron works
#UCSFMHP2025: @gregorymmarcus on hospital arrhythmias. A Fib pearls:
1) New AF in hospital (PNA, post-op, "stress"): generally should still anticoagulate
2) DOAC lowers risk of dementia in AF
3) In AF: exercise good, EtOH bad, coffee is OK
4) Ablation works, but continue DOAC, at least for 1 year
1) New AF in hospital (PNA, post-op, "stress"): generally should still anticoagulate
2) DOAC lowers risk of dementia in AF
3) In AF: exercise good, EtOH bad, coffee is OK
4) Ablation works, but continue DOAC, at least for 1 year
November 1, 2025 at 6:43 PM
#UCSFMHP2025: @gregorymmarcus on hospital arrhythmias. A Fib pearls:
1) New AF in hospital (PNA, post-op, "stress"): generally should still anticoagulate
2) DOAC lowers risk of dementia in AF
3) In AF: exercise good, EtOH bad, coffee is OK
4) Ablation works, but continue DOAC, at least for 1 year
1) New AF in hospital (PNA, post-op, "stress"): generally should still anticoagulate
2) DOAC lowers risk of dementia in AF
3) In AF: exercise good, EtOH bad, coffee is OK
4) Ablation works, but continue DOAC, at least for 1 year
Next #UCSFMHP2025, @LucasZierMD on cardiac ischemia. Points:
1) Serial troponins are more useful to detect ischemia than EKG or single troponin (latter not sensitive enough). See algorithm below
2) Use new terminology (ACS with or without ST Elevation, vs Demand Ischemia (MI Type 2). Graphic below
1) Serial troponins are more useful to detect ischemia than EKG or single troponin (latter not sensitive enough). See algorithm below
2) Use new terminology (ACS with or without ST Elevation, vs Demand Ischemia (MI Type 2). Graphic below
November 1, 2025 at 5:09 PM
Next #UCSFMHP2025, @LucasZierMD on cardiac ischemia. Points:
1) Serial troponins are more useful to detect ischemia than EKG or single troponin (latter not sensitive enough). See algorithm below
2) Use new terminology (ACS with or without ST Elevation, vs Demand Ischemia (MI Type 2). Graphic below
1) Serial troponins are more useful to detect ischemia than EKG or single troponin (latter not sensitive enough). See algorithm below
2) Use new terminology (ACS with or without ST Elevation, vs Demand Ischemia (MI Type 2). Graphic below
Next #UCSFMHP2025: Jen Babik on viral infections in the hospital:
1) Flu: hard to diagnosis, as any older pts don't have temp; vaccine is only 50% effective. So need to test all during flu season. See table for meds.
2) RSV: Rx with ribavirin (oral, not aerosol) for severe PNA
3) Covid: see table
1) Flu: hard to diagnosis, as any older pts don't have temp; vaccine is only 50% effective. So need to test all during flu season. See table for meds.
2) RSV: Rx with ribavirin (oral, not aerosol) for severe PNA
3) Covid: see table
November 1, 2025 at 4:31 PM
Next #UCSFMHP2025: Jen Babik on viral infections in the hospital:
1) Flu: hard to diagnosis, as any older pts don't have temp; vaccine is only 50% effective. So need to test all during flu season. See table for meds.
2) RSV: Rx with ribavirin (oral, not aerosol) for severe PNA
3) Covid: see table
1) Flu: hard to diagnosis, as any older pts don't have temp; vaccine is only 50% effective. So need to test all during flu season. See table for meds.
2) RSV: Rx with ribavirin (oral, not aerosol) for severe PNA
3) Covid: see table
Last day #UCSFMHP2025. Brad Sharpe on electrolytes:
1) K: Clinical risk high only if K<2.5, >6.5. Replete goal: 4.2. "Kayexalate is fine" (& 1/30th cost of Lokelma)
2) Mag: few sxs unless <1.0. "PO Mg is OK" (1/300 cost of IV)
3) HyperCA: Ddx mostly cancer or hyperpara. Don't "adjust for albumin"
1) K: Clinical risk high only if K<2.5, >6.5. Replete goal: 4.2. "Kayexalate is fine" (& 1/30th cost of Lokelma)
2) Mag: few sxs unless <1.0. "PO Mg is OK" (1/300 cost of IV)
3) HyperCA: Ddx mostly cancer or hyperpara. Don't "adjust for albumin"
November 1, 2025 at 4:30 PM
Last day #UCSFMHP2025. Brad Sharpe on electrolytes:
1) K: Clinical risk high only if K<2.5, >6.5. Replete goal: 4.2. "Kayexalate is fine" (& 1/30th cost of Lokelma)
2) Mag: few sxs unless <1.0. "PO Mg is OK" (1/300 cost of IV)
3) HyperCA: Ddx mostly cancer or hyperpara. Don't "adjust for albumin"
1) K: Clinical risk high only if K<2.5, >6.5. Replete goal: 4.2. "Kayexalate is fine" (& 1/30th cost of Lokelma)
2) Mag: few sxs unless <1.0. "PO Mg is OK" (1/300 cost of IV)
3) HyperCA: Ddx mostly cancer or hyperpara. Don't "adjust for albumin"
Some pearls on IBD inpatient management by @SaraLewinMD #UCSFMHP2025:
1) NPO only if the patient can't tolerate po
2) Steroids (40mg IV) for most
3) Consider rescue Rx (usually Infliximab) if not better day 3-5
4) Consider surgery if no better by day 7; but consider JAK1 inhibitor first
1) NPO only if the patient can't tolerate po
2) Steroids (40mg IV) for most
3) Consider rescue Rx (usually Infliximab) if not better day 3-5
4) Consider surgery if no better by day 7; but consider JAK1 inhibitor first
October 31, 2025 at 6:52 PM
Some pearls on IBD inpatient management by @SaraLewinMD #UCSFMHP2025:
1) NPO only if the patient can't tolerate po
2) Steroids (40mg IV) for most
3) Consider rescue Rx (usually Infliximab) if not better day 3-5
4) Consider surgery if no better by day 7; but consider JAK1 inhibitor first
1) NPO only if the patient can't tolerate po
2) Steroids (40mg IV) for most
3) Consider rescue Rx (usually Infliximab) if not better day 3-5
4) Consider surgery if no better by day 7; but consider JAK1 inhibitor first
Palliative care pearls from Steve Pantilat at #UCSFMHP2025:
1) A way to frame palliative care: "To help you live as well as possible for as long as possible"
2) Prognosis: big bands ("weeks to months"). But add: "I hope you do better" and "things can happen suddenly"
3) Empathy improves outcomes
1) A way to frame palliative care: "To help you live as well as possible for as long as possible"
2) Prognosis: big bands ("weeks to months"). But add: "I hope you do better" and "things can happen suddenly"
3) Empathy improves outcomes
October 31, 2025 at 5:58 PM
Palliative care pearls from Steve Pantilat at #UCSFMHP2025:
1) A way to frame palliative care: "To help you live as well as possible for as long as possible"
2) Prognosis: big bands ("weeks to months"). But add: "I hope you do better" and "things can happen suddenly"
3) Empathy improves outcomes
1) A way to frame palliative care: "To help you live as well as possible for as long as possible"
2) Prognosis: big bands ("weeks to months"). But add: "I hope you do better" and "things can happen suddenly"
3) Empathy improves outcomes
A. Josephson's neuro emergencies talk at #UCSFMHP2025:
1) In both Guillain-Barré & myasthenia, do measure lung strength
2) Short peak time to max pain: main way to tell SAH from other headaches
3) Slurred speech: main way to distinguish central vertigo
4) Empiric Rx of meningitis: include steroids
1) In both Guillain-Barré & myasthenia, do measure lung strength
2) Short peak time to max pain: main way to tell SAH from other headaches
3) Slurred speech: main way to distinguish central vertigo
4) Empiric Rx of meningitis: include steroids
October 31, 2025 at 5:49 PM
A. Josephson's neuro emergencies talk at #UCSFMHP2025:
1) In both Guillain-Barré & myasthenia, do measure lung strength
2) Short peak time to max pain: main way to tell SAH from other headaches
3) Slurred speech: main way to distinguish central vertigo
4) Empiric Rx of meningitis: include steroids
1) In both Guillain-Barré & myasthenia, do measure lung strength
2) Short peak time to max pain: main way to tell SAH from other headaches
3) Slurred speech: main way to distinguish central vertigo
4) Empiric Rx of meningitis: include steroids
Last talk #UCSFMHP2025 Day1: Erik Levinsohn, consult psychiatrist @UCSF. Key points:
Delirium most common cause of agitation in the hospital
Deprescribe before new meds: anticholinergic most common culprit
Haldol most useful Rx; its toxicity is overrated
Remove restraints ASAP & push mobilization
Delirium most common cause of agitation in the hospital
Deprescribe before new meds: anticholinergic most common culprit
Haldol most useful Rx; its toxicity is overrated
Remove restraints ASAP & push mobilization
October 30, 2025 at 11:54 PM
Last talk #UCSFMHP2025 Day1: Erik Levinsohn, consult psychiatrist @UCSF. Key points:
Delirium most common cause of agitation in the hospital
Deprescribe before new meds: anticholinergic most common culprit
Haldol most useful Rx; its toxicity is overrated
Remove restraints ASAP & push mobilization
Delirium most common cause of agitation in the hospital
Deprescribe before new meds: anticholinergic most common culprit
Haldol most useful Rx; its toxicity is overrated
Remove restraints ASAP & push mobilization
Andy Josephson on stroke update at #UCSFMHP2025. Key updates:
1) Most hospitals are switching from TPA to TNK (easier-to-use)
2) Endovascular therapy is indicated even after successful thrombolytics
3) Importance of permissive hypertension: don't lower BP after acute stroke
1) Most hospitals are switching from TPA to TNK (easier-to-use)
2) Endovascular therapy is indicated even after successful thrombolytics
3) Importance of permissive hypertension: don't lower BP after acute stroke
October 30, 2025 at 6:47 PM
Andy Josephson on stroke update at #UCSFMHP2025. Key updates:
1) Most hospitals are switching from TPA to TNK (easier-to-use)
2) Endovascular therapy is indicated even after successful thrombolytics
3) Importance of permissive hypertension: don't lower BP after acute stroke
1) Most hospitals are switching from TPA to TNK (easier-to-use)
2) Endovascular therapy is indicated even after successful thrombolytics
3) Importance of permissive hypertension: don't lower BP after acute stroke
Tracy Minichiello on anticoagulation at #UCSFMHP2025. Some take-homes below. Note trends away from dual therapy (ASA+DOAC), and increasing use of reduced-dose DOACs. Also, in many patients (such as post-DVT or AFib), the benefits of anticoag exceed risks even in patients with history of falls.
October 30, 2025 at 6:19 PM
Tracy Minichiello on anticoagulation at #UCSFMHP2025. Some take-homes below. Note trends away from dual therapy (ASA+DOAC), and increasing use of reduced-dose DOACs. Also, in many patients (such as post-DVT or AFib), the benefits of anticoag exceed risks even in patients with history of falls.
At #UCSFMHP2025: Jonathan Davis on inpatient management of heart failure. Key messages are that aggressive hospital management, with Goal-Directed Medical Therapy, is key and has an impressive mortality benefit.
October 30, 2025 at 4:34 PM
At #UCSFMHP2025: Jonathan Davis on inpatient management of heart failure. Key messages are that aggressive hospital management, with Goal-Directed Medical Therapy, is key and has an impressive mortality benefit.
Key points from Lekshmi Santhosh sepsis update at #UCSFMHP2025. Remember: give fluids early in shock but de-escalate quickly once hemodynamically stable.
Hot off the presses: @NEJM study showing that many ICU sepsis patients don't benefit from arterial lines nejm.org/doi/full/10....
Hot off the presses: @NEJM study showing that many ICU sepsis patients don't benefit from arterial lines nejm.org/doi/full/10....
October 30, 2025 at 3:41 PM
Key points from Lekshmi Santhosh sepsis update at #UCSFMHP2025. Remember: give fluids early in shock but de-escalate quickly once hemodynamically stable.
Hot off the presses: @NEJM study showing that many ICU sepsis patients don't benefit from arterial lines nejm.org/doi/full/10....
Hot off the presses: @NEJM study showing that many ICU sepsis patients don't benefit from arterial lines nejm.org/doi/full/10....
It's always a thrill to welcome > 500 hospitalists to my annual
UCSF Management of the Hospitalized Patient conference in SF. I'll tweet 1-2 take-home points from talks over the next few days. Below: Lekshmi Santhosh on sepsis updates (incl. data hot off presses)
#UCSFMHP2025
UCSF Management of the Hospitalized Patient conference in SF. I'll tweet 1-2 take-home points from talks over the next few days. Below: Lekshmi Santhosh on sepsis updates (incl. data hot off presses)
#UCSFMHP2025
October 30, 2025 at 3:39 PM
It's always a thrill to welcome > 500 hospitalists to my annual
UCSF Management of the Hospitalized Patient conference in SF. I'll tweet 1-2 take-home points from talks over the next few days. Below: Lekshmi Santhosh on sepsis updates (incl. data hot off presses)
#UCSFMHP2025
UCSF Management of the Hospitalized Patient conference in SF. I'll tweet 1-2 take-home points from talks over the next few days. Below: Lekshmi Santhosh on sepsis updates (incl. data hot off presses)
#UCSFMHP2025
My new Substack: Will AI Rescue Primary Care?
Why AI-enabled solutions have mostly failed to solve the primary care crisis, and my hope/prediction that the endgame will likely be a new two-tiered system that increasingly relies on AI for basic care needs.
robertwachter.substack.com/p/will-ai-re...
Why AI-enabled solutions have mostly failed to solve the primary care crisis, and my hope/prediction that the endgame will likely be a new two-tiered system that increasingly relies on AI for basic care needs.
robertwachter.substack.com/p/will-ai-re...
Will AI Rescue Primary Care?
Eventually, But Not Yet
robertwachter.substack.com
October 8, 2025 at 2:10 PM
My new Substack: Will AI Rescue Primary Care?
Why AI-enabled solutions have mostly failed to solve the primary care crisis, and my hope/prediction that the endgame will likely be a new two-tiered system that increasingly relies on AI for basic care needs.
robertwachter.substack.com/p/will-ai-re...
Why AI-enabled solutions have mostly failed to solve the primary care crisis, and my hope/prediction that the endgame will likely be a new two-tiered system that increasingly relies on AI for basic care needs.
robertwachter.substack.com/p/will-ai-re...
Just 1 month until my 29th (!) annual Management of the Hospitalized Patient CME course in SF, Oct 30-Nov 1. Great topics (mostly clinical, for hospitalists & others who care for inpatients), awesome faculty. Co-sponsored by SHM.
Hope you can make it!
managementhospitalizedpatient.ucsf.edu
Hope you can make it!
managementhospitalizedpatient.ucsf.edu
Management of the Hospitalized Patient
managementhospitalizedpatient.ucsf.edu
October 4, 2025 at 1:10 AM
Just 1 month until my 29th (!) annual Management of the Hospitalized Patient CME course in SF, Oct 30-Nov 1. Great topics (mostly clinical, for hospitalists & others who care for inpatients), awesome faculty. Co-sponsored by SHM.
Hope you can make it!
managementhospitalizedpatient.ucsf.edu
Hope you can make it!
managementhospitalizedpatient.ucsf.edu
My latest Substack: As OpenEvidence sprinted past UpToDate to become medicine's best AI-enabled point-of-care tool, I've been waiting for UTD's response. Now it's here, raising fascinating questions about which knowledge base should be our source of truth.
robertwachter.substack.com/p/medicines-...
robertwachter.substack.com/p/medicines-...
Medicine’s AI Knowledge War Heats Up
The Battlegrounds May Surprise You
robertwachter.substack.com
October 1, 2025 at 5:53 PM
My latest Substack: As OpenEvidence sprinted past UpToDate to become medicine's best AI-enabled point-of-care tool, I've been waiting for UTD's response. Now it's here, raising fascinating questions about which knowledge base should be our source of truth.
robertwachter.substack.com/p/medicines-...
robertwachter.substack.com/p/medicines-...
Everyone hates prior auth, so I was surprised last month when Medicare rolled out a new prior auth program. AI can solve the prior auth mess–but it means replacing our fax-based system with AI-powered chart reviews. My new Substack on how this could work.
open.substack.com/pub/robertwa...
open.substack.com/pub/robertwa...
Can AI Rescue Us from the Mess of Prior Auth?
Yes, But It’ll Take More Than Putting Lipstick on Today’s Pig of a System
open.substack.com
September 10, 2025 at 12:36 PM
Everyone hates prior auth, so I was surprised last month when Medicare rolled out a new prior auth program. AI can solve the prior auth mess–but it means replacing our fax-based system with AI-powered chart reviews. My new Substack on how this could work.
open.substack.com/pub/robertwa...
open.substack.com/pub/robertwa...
To me, understanding the "Productivity Paradox" is vital for appreciating why digitizing healthcare has been so hard, why the EHR has been so fraught, and why AI in medicine will be bumpy (but positive in the end).
Great explainer by @johncassidysays.bsky.social www.newyorker.com/news/the-fin...
Great explainer by @johncassidysays.bsky.social www.newyorker.com/news/the-fin...
The A.I.-Profits Drought and the Lessons of History
Like the steam engine, electricity, and computers, generative artificial intelligence could take longer than expected to transform the economy.
www.newyorker.com
September 9, 2025 at 3:18 PM
To me, understanding the "Productivity Paradox" is vital for appreciating why digitizing healthcare has been so hard, why the EHR has been so fraught, and why AI in medicine will be bumpy (but positive in the end).
Great explainer by @johncassidysays.bsky.social www.newyorker.com/news/the-fin...
Great explainer by @johncassidysays.bsky.social www.newyorker.com/news/the-fin...
My book on healthcare AI, "A Giant Leap" (sites.prh.com/agiantleapbook), is still a few months out, but it's a thrill to receive endorsements from early readers.
Here are a couple of new ones, from two of my heroes: @zekeemanuel.bsky.social and Peter Lee from @microsoft.com
Here are a couple of new ones, from two of my heroes: @zekeemanuel.bsky.social and Peter Lee from @microsoft.com
September 2, 2025 at 1:27 PM
My book on healthcare AI, "A Giant Leap" (sites.prh.com/agiantleapbook), is still a few months out, but it's a thrill to receive endorsements from early readers.
Here are a couple of new ones, from two of my heroes: @zekeemanuel.bsky.social and Peter Lee from @microsoft.com
Here are a couple of new ones, from two of my heroes: @zekeemanuel.bsky.social and Peter Lee from @microsoft.com
I loved (and agree with) today's @axios piece on AI. The big thought:
"[Leaders] with the most insight, power & money share the AI obsession. The topic is daunting, sprawling, occasionally scary & often eye-popping. It's worthy of your time to start obsessing." www.axios.com/2025/09/02/a...
"[Leaders] with the most insight, power & money share the AI obsession. The topic is daunting, sprawling, occasionally scary & often eye-popping. It's worthy of your time to start obsessing." www.axios.com/2025/09/02/a...
Behind the Curtain: Why you should be AI-obsessed
Imagine knowing electricity or the internet were coming.
www.axios.com
September 2, 2025 at 1:05 PM
I loved (and agree with) today's @axios piece on AI. The big thought:
"[Leaders] with the most insight, power & money share the AI obsession. The topic is daunting, sprawling, occasionally scary & often eye-popping. It's worthy of your time to start obsessing." www.axios.com/2025/09/02/a...
"[Leaders] with the most insight, power & money share the AI obsession. The topic is daunting, sprawling, occasionally scary & often eye-popping. It's worthy of your time to start obsessing." www.axios.com/2025/09/02/a...
More on AI-induced deskilling – terrific piece in today's @nytimes.com by Teddy Rosenbluth, with thoughts by @adamrodmanmd.bsky.social, Chris Longhurst, and me on implications, and how to – and whether we need to – mitigate the loss of skills. www.nytimes.com/2025/08/28/w...
Are A.I. Tools Making Doctors Worse at Their Jobs?
www.nytimes.com
August 28, 2025 at 12:30 PM
More on AI-induced deskilling – terrific piece in today's @nytimes.com by Teddy Rosenbluth, with thoughts by @adamrodmanmd.bsky.social, Chris Longhurst, and me on implications, and how to – and whether we need to – mitigate the loss of skills. www.nytimes.com/2025/08/28/w...
My newest Substack post: in light of the Lancet study that found deskilling after brief use of AI-colonoscopy, I discuss AI-based deskilling in medicine, how other industries handle it, and why we're not asking the right questions. Don't forget to subscribe (free). open.substack.com/pub/robertwa...
Deskilling and Healthcare AI
Making Sure We're Asking the Right Questions
open.substack.com
August 21, 2025 at 2:51 PM
My newest Substack post: in light of the Lancet study that found deskilling after brief use of AI-colonoscopy, I discuss AI-based deskilling in medicine, how other industries handle it, and why we're not asking the right questions. Don't forget to subscribe (free). open.substack.com/pub/robertwa...