The OOS Flowchart
Phase I: Laboratory Investigation
(Did the Lab make a mistake?)
[ START: OOS Result Detected ]
|
v
[ Notify Supervisor Immediately ]
|
v
[ CHECK: Is there an obvious Lab Error? ]
(Calculation error, power failure, spilled sample, wrong instrument)
|
+———————–+
| |
[ YES ] [ NO ]
| |
v v
[ Invalidate Original Result ] [ No Lab Error Found ]
| |
v |
[ Repeat Test Correctly ] |
| |
v v
[ Result Passes? ] [ PROCEED TO PHASE II ]
| | (Full Scale Investigation)
YES NO
| |
[PASS] [OOS Confirmed]
Phase II: Full-Scale Investigation
(Did Production make a mistake? or is the product bad?)
[ START: Phase II ]
|
v
[ CHECK: Manufacturing/Production Review ]
(Did a machine break? Was the wrong raw material used?)
|
+—————————–+
| |
[ YES: Error Found ] [ NO: No Error Found ]
| |
v v
[ Determine Root Cause ] [ LABORATORY RETESTING ]
| (Testing the sample again with
| multiple distinct analysts)
v |
[ BATCH REJECTION ] v
(Usually) [ Do results pass? ]
|
+————-+————-+
| |
[ YES ] [ NO ]
| |
v v
[ BATCH RELEASE ] [ BATCH REJECT ]
(If scientific data (OOS is confirmed)
proves the batch is
actually good)
Explanation of Steps (In Easy Language)
1. Phase I: Laboratory Investigation
Before blaming the product, you must check if the analyst or equipment made a mistake.
Obvious Error: Did the analyst use the wrong chemical? Did the calculator fail? Did they spill the sample?
If YES: You document the error, mark the first result as “void” (invalid), and run the test again.
If NO: The lab did everything right, but the result is still bad. Now you have to look at how the product was made.
2. Phase II: Production Review
Check the manufacturing logbooks.
Production Error: Did the mixer stop working? Did someone add too much water?
If YES: The batch is likely bad because it was made incorrectly. The batch is usually rejected.
3. Phase II: Retesting
If the lab made no mistakes, and production looks fine, you must test the sample again to be absolutely sure.
The Rule: You cannot just test until you get a passing result (“Testing into compliance”). You usually need to run the test 3 to 6 times (often with a different analyst) to prove the first bad result was a fluke.
The Decision:
If retests are Passing: The first result might be an anomaly. The Quality Unit may release the batch.
If retests are Failing: The product is definitely bad. The batch is rejected.
Summary