Writing:
Perspectives, parts of the guide, Clinical Meaning in the Age of Health Data & AI, free written exercises for clinicians, sample reviews and CIRI™ method, are all published here
Perspectives
From the guide
Clinical Meaning in the Age of Health Data & AI. Two of the guide's four parts are published here in full.
Part I - Seeing the Data Layer
What documentation selects, and the nine objects one save creates.
Chapter 10 - When clinical data becomes AI data
What a model inherits, and why it cannot repair it.
Chapter 11 - When the data producer is no longer human
Ambient documentation, omission, and what a signature now means.
Chapter 12 - When AI sounds right but means wrong
Why fluency was a valid proxy, and why it no longer is.
For clinicians
Five written exercises. Each one puts a record in front of you and asks what somebody else could establish from it. Twenty minutes each, on paper, alone..
Reading a system-drafted note
A twenty-minute written exercise. One clinical note produced by an ambient documentation system, four things wrong with it, and the five questions that find them. Free, and it asks for nothing.
Reading a system-drafted note: paediatrics
A paediatric note with a missing age, a diagnosis the clinician did not reach, a parent's condition in the child's record, and a prescription nothing in the note supports.
Reading a system-drafted note: paediatrics II
An infant admitted for the right reasons, and a note that no longer shows them: a negation stretched too far, a saturation without its conditions, a glucose nobody measured, and a prematurity that disappeared.
What a coder can establish
The only one with no AI in it. Two notes for the same patient, the same care and the same outcome - and two records that describe different people.
العربيةقراءة ملاحظة كتبها نظام: حين يتكلم المريض بلهجته
The first exercise in Arabic. A consultation held in dialect, a note written in English, and four errors made in the distance between what a word means in the patient's dialect and what it means in the dictionary.
Sample reviews
One worked sample for each review under Clinical Meaning Integrity. The cases are simulated; no patient data is involved.
Sample Clinical Claim Review
Five claims made for Arabic-capable ambient AI scribes, tested against the published evidence.
Sample Clinical AI Evaluation
A simulated paediatric consultation in Levantine Arabic, and the ambient scribe note made from it, evaluated element by element.
Sample Terminology Mapping Review
A simulated outpatient problem list, in English and colloquial Arabic, mapped to SNOMED CT.