Journal
Articles.
Short pieces on hospital operations read as computational problems. Unlike the case study audits, these aren’t bound to a fixed structure or one of the four domains.
When the Machine Points at Nothing, Doctors See Something Anyway
How a single false-positive AI suggestion measurably shifts even experienced clinicians’ judgment.
The Algorithm That Cried Sepsis
Inside the study that showed a sepsis-prediction tool used at hundreds of hospitals badly missed the mark.
One Million Alarms a Week, and the One That Mattered Got Missed
A hospital unit can generate an average of one million alarms in a single week.
The Health System That Can’t Talk to Itself
A pediatrician wrote about losing her grandfather, not to his illness alone, but to a hospital system that couldn’t assemble his own medical history.
When AI Gets It Wrong, the Law Still Blames the Doctor
AI is now woven into triage software, radiology reads, and clinical decision-support tools across hospitals nationwide.
Why a Hospital’s Own Data Can Be Its Biggest Blind Spot
An AI model can look excellent where it was trained and still quietly get worse the moment it’s deployed somewhere else.
The Vital Sign That’s Already 42 Minutes Old by the Time Anyone Sees It
“Real-time” gets used loosely in health tech marketing. In a real hospital, it usually isn’t.
“FDA Cleared” Sounds Like Proof. It Isn’t.
About 90% of physicians assume that when an AI medical device reaches their hospital, it’s already been through something like a randomized clinical trial.
Most Hospital Dashboards Tell You What Already Happened
It looks like real-time intelligence. Often, it’s closer to a live rearview mirror.
Nine Billion Fax Pages a Year, and Counting
A meaningful share of the actual paperwork behind a patient’s care is still moving through a phone line built for the 1980s.
The Data Gap Nobody’s Rural Hospital Can Afford to Close
Every case study and article on this site so far has assumed something quietly: that a hospital has the budget, staff, and infrastructure to actually build a predictive system.
40% of a Nursing Shift Is Spent Typing, Not Treating
Nurse burnout usually gets framed as a staffing crisis — not enough nurses, too many patients per shift.
Half the Physical Exam Just Doesn’t Happen on a Video Call
It also has a structural gap that hasn’t gone away: a screen can’t touch a patient.
49 Out of 50 Telehealth Platforms Shared Your Data With Advertisers
Most patients assume that typing symptoms into an intake form is a private conversation between them and a doctor.