Chemistry Reference
In-Depth Information
UCL
Target
LCL
2
4
6
8
10
12
14
16
Run Number
FIgure 6.3
Typical control chart illustrating OOS results. Charts such as this can be
used to monitor for OOS results. By setting an upper control limit (UCL) and a lower con-
trol limit (LCL) around a target or average value according to specifications, OOS samples,
injections, batches, etc., can be easily observed.
outside the specification or acceptance criteria established in new drug applications
(NDAs), official compendia, or by the manufacturer” [15]. Control charts, such as
the example shown in Figure 6.3, can be used to easily inspect for OOS results.
OOS results can also observed, for example, in chromatography, from system suit-
ability results. Once an OOS result is obtained, an investigation must be launched
to determine the cause, and each step of the investigation must be documented. The
first phase of the investigation should include an initial assessment of the accuracy
of the data, before test solutions are discarded. It is the responsibility of the analyst
to review the data for compliance with specifications, and in cases where unexpected
results are obtained and no obvious explanation exists, retain test solutions, and
inform the supervisor. The supervisor's assessment should be objective and timely,
and include the following steps:
1. Discussing the test method with the analyst to confirm that the proper pro-
cedure was performed.
2. Examine the raw data, to identify potentially anomalous or suspect information.
3. Confirmation of instrument performance by reviewing qualification and
system suitability data.
4. Verify that proper reference standards, solvents, reagents, and other solu-
tions were used, and that they meet quality control specifications.
5. Compare the test method performance to ensure that it is performing to the
standard expected based on method validation data.
6. Documented evidence of the assessment.
Examining retained samples promptly is important to facilitate assigning a cause
to OOS results. For example, reinjection where a transient instrument malfunction
is suspected can provide strong evidence to rule out sample or sample preparation
anomalies. However, laboratory error should be relatively rare. Frequent errors sug-
gests inadequate training, poorly maintained or calibrated instruments, or careless
work. Whenever laboratory error is identified, corrective action should be taken to
prevent the problem from reoccurring.
