Hemolyzed Specimens
Chapter 5 Bedside Questions · 11 questions · about 11 minutes
Executives read numbers that have passed through six to eight unowned transformations as if they were calibrated lab values. No one owns the data proposition end to end, so artifacts become decisions, and the cost shows up disguised as denials, manual abstraction, low case-mix index and burnout.
A hemolyzed specimen: a lab value that looks real but is artifact; corrupt data mislead every dashboard, payment calculation and AI model just as a hemolyzed potassium misleads the physician.
Answer as your executive team would, with the evidence in hand. Each question has a factual answer in a system you already own.
Which of the five data quality dimensions has any dashboard been formally tested against, and can leadership name them?
Add a note: where is the evidence, or who owns it?
For each consequential metric, is its intended use written down?
Add a note: where is the evidence, or who owns it?
Who owns the master patient index and terminology services?
Add a note: where is the evidence, or who owns it?
Has data lineage for the top five quality metrics been audited end to end in the last 12 months?
Add a note: where is the evidence, or who owns it?
Are dashboard definitions stable across departments, or does each leader read a different number?
Add a note: where is the evidence, or who owns it?
Which measurements perform differently across patient subgroups, and who is responsible for knowing?
Add a note: where is the evidence, or who owns it?
Do you run manual abstraction in parallel because you do not trust electronic measurement, and what does it cost annually?
Add a note: where is the evidence, or who owns it?
Have AI deployments been locally validated with subgroup performance monitored over time?
Add a note: where is the evidence, or who owns it?
Who catches definitional drift between systems, and what is the escalation path?
Add a note: where is the evidence, or who owns it?
Has the executive team walked a metric from definition to dashboard with an analyst in the last year?
Add a note: where is the evidence, or who owns it?
When the dashboard disagrees with the bedside, which wins, and who decides?
Add a note: where is the evidence, or who owns it?
A reflection tool from the book, not a validated diagnostic test or benchmark. Your answers and notes stay in this browser unless you choose to send them. Please don't enter patient information.
We are paying physicians to produce data, at the highest labor rate in the building, and much of what they produce cannot be used.
The Diagnostician, Chapter 5