Laboratory data integrity is the property a regulator or an accreditor tests when they pick one result and ask the lab to show where it came from. Lab data integrity is usually summarized as attributable, legible, contemporaneous, original and accurate: the result names its analyst, can be read, was recorded when it happened, is the first record of itself and matches what the instrument produced. A small lab meets that with a record that keeps the result with its method, lot, instrument, analyst and reviewer and logs every change with who and when. This page is about how the record answers each ask, and the free results review sheet on this site works the review load that integrity depends on from the lab's own numbers with no account.
Attributable and contemporaneous
The result names the analyst who produced it and the reviewer who released it, with the time of each, because the record wrote the name and the time when the state changed rather than when someone remembered. A record that lets a result be entered under a shared login has failed the first ask before the second is reached.
Original and accurate
The instrument's result lands on the test without a retype, and every change since is a new line with the old value kept; that is the audit trail Part 11 describes for regulated records and ISO 17025 asks of a testing lab's management system. Overwriting is the failure; the lab's own quality manager reads the rule.
The review that makes it legible
A result is released by a named reviewer who read it, and the review load is where integrity is lost when the queue backs up: 400 results a week at three minutes is 20 hours against 24 available on the worked example, from the results review sheet on this site. Limslane Pro keeps the attribution, the trail and the review at one flat price.
Questions people ask about laboratory data integrity
What is laboratory data integrity?
Every result attributable, legible, contemporaneous, original and accurate, with an audit trail of every change; tested by picking one result and asking where it came from.
How does a small lab's record answer it?
By naming the analyst and the reviewer with times, landing the instrument's result without a retype, and keeping every change as a new line with the old value.
Where is integrity usually lost?
In the review queue and in shared logins; the free results review sheet works the queue's load.