Process validation is commonly considered a technical activity involving qualification of instruments, key process characteristics, inspection plans, statistical assessment, etc. However, many defects in the validation processes are caused by much simpler reasons, including human mistakes.
A properly designed process may yield erroneous results of the validation due to misunderstanding of instructions by an operator, improper registration of information, incorrect input of variable values, and lack of analysis of deviations. Regulatory auditors are aware of this fact, which prompts them to evaluate not only the validation protocol itself but also the level of training of personnel, documentation practices, and decision-making during the study.
Knowledge of how human errors affect process validation can help Quality Assurance, Production, Quality Control, and Validation departments avoid unnecessary failures of the validation process.
In a batch record validation process, the information regarding mixing time might have been entered incorrectly. But the investigation should not stop there. The investigation should also look for the reasons why the error occurred in the first place.
The most common human errors occur when:
Common mistakes in the execution of the validation protocol include - sampling locations that were not pre-defined, samples taken outside of the time frames specified, incorrect sampling instruments being utilized, missing environmental variables, etc.
All these mistakes may not be enough to invalidate the whole validation process but will require scientific review. What is most important is whether the error has compromised the goal of the validation.
Stating that "the result is approved" might not be sufficient.
Unfortunately, significant mistakes happen when information is not available for each department.
To illustrate, engineering may replace a sensor prior to the validation run and fails to tell validation that the sensor was calibrated. In addition, production may deal with an unexpected process interruption, but the QA is not notified.
A good validation program relies on timely communication in addition to decent paperwork.
Some of the typical documentation errors are:
A more relevant approach consists in assessing whether the employees understand what validation activity they carry out.
For instance, an operator making blend uniformity samples should know the significance of the location where the samples are taken. A technician should understand the influence of the location of the sensors when conducting temperature mapping.
Essentially effective validation training is more about knowing why this activity is performed rather than remembering the steps.
Picture a validation team that awaits a successful result because its previous tests of the engineering design produced satisfying results. During the validation phase, the data that are contrary to the expectation may be interpreted hastily as either analytical randomness or sample inadequacy.
When I perform an investigation, I seek proof that alternative hypotheses have been sufficiently considered.
Ideally, a validation must be based on the data regardless of whether the latter are inconvenient.
Here are a few examples:
Possible improvements can be independent verification of crucial inputs, adoption of standardized forms for data collection, applying electronic data captures when possible, using well-known points of sampling, organizing walk-throughs of the protocol before the task is undertaken, consideration of the team briefing and organization of its follow-up review.
An orderly review is likely to detect human-factor problems before they become inspection comments.
ICH Q9(R1) stresses the use of quality risk management for risk identification and control throughout the product life cycle, including risks stemming from the execution of the process and from people's performance.
Successful programs in validation understand that people are included in the process. By improving execution of the protocol, taking care of the data integrity, promoting communication across functional departments, and analyzing the errors from a scientific standpoint, pharmaceutical companies are able to produce valid and sound data for regulatory inspections.
Knowledge of how human errors affect process validation can help Quality Assurance, Production, Quality Control, and Validation departments avoid unnecessary failures of the validation process.
Human Error Is Usually a System Problem
I have found that one of the most common mistakes is to blame the operator for every error related to validation. While it is true the operator could be at fault for making the wrong move, it does not mean that the reason for the mistake has to be an operator error.In a batch record validation process, the information regarding mixing time might have been entered incorrectly. But the investigation should not stop there. The investigation should also look for the reasons why the error occurred in the first place.
- Was the batch record not clear enough?
- Have the parameters been confirmed independent of the operator?
- Did the HMI equipment accept the incorrect information?
- Did the operator have enough training?
- Did the operator have to cope with too many tasks while running the process?
Incorrect Recording of Critical Process Parameters
The Critical Process Parameters (CPPs) are the operational parameters that affect the quality of the product. During the process validation the CPP values are most important part of the evidence of the validity of the process.The most common human errors occur when:
- The parameters are being recorded after the activity rather than at the same time.
- Decimal points are entered incorrectly.
- Units are mixed up.
- Manual values read do not correspond to electronic data.
- The corrections are made but not documented.
Poor Execution of the Approved Validation Protocol
A validation protocol is a plan for an experiment or study to be performed. The validation activity should be executed as set in the validation protocol unless there is an approved deviation documented.Common mistakes in the execution of the validation protocol include - sampling locations that were not pre-defined, samples taken outside of the time frames specified, incorrect sampling instruments being utilized, missing environmental variables, etc.
All these mistakes may not be enough to invalidate the whole validation process but will require scientific review. What is most important is whether the error has compromised the goal of the validation.
Stating that "the result is approved" might not be sufficient.
Poor Interdepartmental Communication
Validation typically involves multiple departments. For example, production makes the batch, quality control runs tests, engineering provides assistance, validation coordinates the project, and the QA checks the documentation.Unfortunately, significant mistakes happen when information is not available for each department.
To illustrate, engineering may replace a sensor prior to the validation run and fails to tell validation that the sensor was calibrated. In addition, production may deal with an unexpected process interruption, but the QA is not notified.
A good validation program relies on timely communication in addition to decent paperwork.
Documentation Errors Could Hurt Validation Evidence
Validation records are evidence in terms of regulations. Poor documentation could weaken the validity, even if the process appears to be effective.Some of the typical documentation errors are:
- Absence of signatures
- Not fully filled dates and times
- Not giving the reasons behind corrections
- Unclear entries
- Absence of attachment references
- Wrong batch code
- No control over worksheets
- Not attaching the materials
When Training Becomes the Root Cause
The records of previous training may indicate the personnel members went through GMP training, but mere attendance does not imply that they are competent.A more relevant approach consists in assessing whether the employees understand what validation activity they carry out.
For instance, an operator making blend uniformity samples should know the significance of the location where the samples are taken. A technician should understand the influence of the location of the sensors when conducting temperature mapping.
Essentially effective validation training is more about knowing why this activity is performed rather than remembering the steps.
The Risk of Confirmation Bias During Validation
Errors committed by humans sometimes do not stem from accident but from confirmation bias.Picture a validation team that awaits a successful result because its previous tests of the engineering design produced satisfying results. During the validation phase, the data that are contrary to the expectation may be interpreted hastily as either analytical randomness or sample inadequacy.
When I perform an investigation, I seek proof that alternative hypotheses have been sufficiently considered.
Ideally, a validation must be based on the data regardless of whether the latter are inconvenient.
What Inspectors Commonly Ask
Regulatory inspectors tend to use straightforward questions in order to check whether the validation team has a good understanding of the study.Here are a few examples:
- What makes you think that this parameter is important?
- Why did you choose this specific sampling site?
- What was happening in the course of the interruption?
- Can you show me the raw data as collected?
- Who checked the values recorded?
- What made you think that this deviation was okay?
- How do you know that the validated status has been preserved?
Reducing Human Errors Through Better Process Design
The leading organizations do not depend only upon the advice of being cautious. They create verification processes in order to minimize the emergence of the opportunities for human errors.Possible improvements can be independent verification of crucial inputs, adoption of standardized forms for data collection, applying electronic data captures when possible, using well-known points of sampling, organizing walk-throughs of the protocol before the task is undertaken, consideration of the team briefing and organization of its follow-up review.
Checklist for Effective QA Review
Before giving a green signal to a process validation document the QA department must consider:| Review Area | Key Question |
| Protocol implementation | Was every essential step taken according to the protocol? |
| CPP documentation | Do the manual and electronic data match? |
| Sampling | Were all steps and times defined before followed? |
| Non-conformance | Was the influence on validation studied scientifically? |
| Primary data | Is every data point traceable back to the initial document? |
| Training | Are the people involved qualified to perform the required tasks? |
| Data integrity | Is any change made visible and accounted for? |
| Final conclusion | Does the information collected enable the right validation decision to be made? |
An orderly review is likely to detect human-factor problems before they become inspection comments.
Regulatory Perspective
FDA describes the validation process as a life cycle activity that involves scientific evidence, written procedures, and ongoing verification of procedures. Human aspects such as the accuracy and quality of the documentation, as well as the level of adherence and investigation can affect the quality of the evidence.ICH Q9(R1) stresses the use of quality risk management for risk identification and control throughout the product life cycle, including risks stemming from the execution of the process and from people's performance.
Successful programs in validation understand that people are included in the process. By improving execution of the protocol, taking care of the data integrity, promoting communication across functional departments, and analyzing the errors from a scientific standpoint, pharmaceutical companies are able to produce valid and sound data for regulatory inspections.
Get ready to use editable Validation Protocols in MS-Word FormatView List
Get documents for Audit preparation in MS-Word FormatView List
No comments:
Post a Comment
Please don't spam. Comments having links would not be published.