Case Studies
Where I have landed on each technology.
Each case study summarizes how I think about a specific clinical technology, what trade-offs it forces, and whether I am using it, evaluating it, or declining it. These pages are revised when my thinking changes; the change history will eventually live in git.
Not using
Ambient AI scribes
I do not use ambient AI scribes at the bedside. The microphone introduces consent, accountability, and cognitive-authorship problems that the current generation of products does not solve.
Using
Clinical decision support
Decision-support tools that have been externally validated on populations resembling mine, with a clear workflow for who owns the decision when the model is wrong, get used. Everything else waits.
Evaluating
Generative AI in clinical communication
Generative AI for patient-facing communication — discharge instructions, medication summaries, condition explanations — is a real opportunity. It is also being deployed to replace the cognitive work of being a doctor in ways that the published evaluation data does not yet support.
Watching
Hospital at home
Acute hospital-at-home is real care, not telehealth dressed up. It works for a specific patient profile, requires real infrastructure, and is being mis-sold as a cost-savings program rather than as the care-model evolution it actually is.
Watching
Patient data
Patient data has become a market. Most patients do not know this, and most consent forms do not meaningfully disclose it. My professional obligation is to remain useful to the patient inside that market, which sometimes means slowing down processes that have been optimized for the data flow rather than the care.
Watching
Value-based care
Value-based contracts that align clinician incentives with measurable patient outcomes are good medicine. Value-based contracts that align with administrative throughput metrics dressed up as outcomes are not the same thing, and conflating the two is the central rhetorical move of the industry.
Watching
Workforce shifts
The clinical workforce is restructuring — APP expansion, scribe-and-AI augmentation, agency staffing, telehospitalist coverage, hospital-at-home teams. Some of these shifts are improvements. Some are cost-shifting dressed as innovation. Patients are usually the last to be told which is which.