K
The Benefits of 21 CFR Part 11
Compliance — Even if the FDA Never Asks for Your Records
Introduction:
With its release in 1997 of 21 CFR Part 11, the FDA specified its requirements for accepting electronic records in lieu of paper records. If you are in the process of automating a drug discovery process, you should be aware of the requirements of the Part 11. Most companies have focused their Part 11 compliance efforts on manufacturing, on clinical trials, and on development, because that has been the focus of FDA auditing efforts. But it also makes good sense to design a discovery automation system to comply with Part 11, even though you may consider it unlikely that the FDA will every demand these records:
• The immediate goal of drug discovery is a patent filing. Compliance with Part
11 will ensure that you have full support for patents. You will be able to supply exact records supporting any filing or later patent dispute. The principal investigator, time of discovery, and methodology will be identified beyond any doubt.
• The requirements of Part 11 are all based on common sense and good
workflow. They are really just a codification of good laboratory practice (GLP). Part 11 may be a regulatory process, but it is also a set of guidelines
for improving the security and efficiency of your discovery-stage workflow.
• If you have the opportunity of considering Part 11 compliance during the system design process, the implementation requirements are not that onerous. It is much easier to build the system with Part 11 in mind than to deal with Part 11 remediation later.
• The dividing lines between the discovery and development stages are not
clear-cut, and drug candidates often cycle between the two stages. If your
discovery process is Part 11-compliant, the transition to Part 11- compliant development will be simpler.
• The FDA may decide in the future to further codify the requirements for
drug discovery records. If it does so, it is likely that the requirements will follow the same standards as Part 11. In general, it is a good idea to design any new automation system to comply with the regulations of Part 11. As this paper will show, it is not that difficult to design your system with Part 11 compliance in mind. This paper describes an approach to designing
a drug discovery system that is Part 11-compliant. The ideas described
have been implemented in Coffey Analysis’s Automation Explorer™. You may find this paper useful in reviewing your current systems, in designing a new system, or in evaluating the approaches suggested by software and hardware vendors. Part 11 describes the FDA’s requirements for acceptable electronic records but does not require a particular implementation. In addition, the FDA objects to describing software as “complying with Part 11”. The FDA views the company and its principal investigators as complying (or not) with Part 11. While this paper refers to software as “complying with Part 11” , please view that as shorthand for “software that assists the company and its
investigators in complying with 21 CFR Part 11”, and please view the ideas on implementation as those of Coffey Analysis rather than the requirements of the FDA.
Goals behind Part 11
When electronic records are submitted, the FDA wants to be able to ensure the following:
• That the information is complete – that each record can be tracked to its
source and that related records are connected.
• That the time of information entry and of any modification is noted.
• That the information has not been altered in a manner that obscures the
original information (much as is done with laboratory notebooks, where corrections are signed and where incorrect information is never deleted,
merely amended).
• That only authorized personnel can access the system, and the person
creating, modifying, or reviewing any information is identified and personally
attests to the validity of the process.
• That all experimental protocols are clearly documented and identified.
• That all personnel are trained in the protocols and that documentation is
available.
• That the system has been validated to operate correctly.
• That information can be viewed in either electronic or human-readable format and that records be available “in a timely fashion”.
• That the system be available for inspection,including both hardware and
software. Is there anything in this list that you would not want in your automation system?
A Typical Automation System
A typical discovery-stage automation system will include
• Manual processes (pipetting, weighing, etc.) with protocols defined and
documented for the processes.
• Automated processes that employ laboratory instruments, with protocols
defined and documented for the processes.
• Software that controls the individual instruments and generates data and
reports. This is usually commercial software supplied by the instruments’
manufacturers.
• Decisions and assessments made by scientists based on information generated by the processes, often based upon external third-party software
(e.g., protein sequencing, computer modeling).
• A central relational database that is used to store the information generated in the processes.
• Automation software, often called LIMS (Laboratory Information Management Software) that pulls all of this together.
Compliance — Even if the FDA Never Asks for Your Records
Introduction:
With its release in 1997 of 21 CFR Part 11, the FDA specified its requirements for accepting electronic records in lieu of paper records. If you are in the process of automating a drug discovery process, you should be aware of the requirements of the Part 11. Most companies have focused their Part 11 compliance efforts on manufacturing, on clinical trials, and on development, because that has been the focus of FDA auditing efforts. But it also makes good sense to design a discovery automation system to comply with Part 11, even though you may consider it unlikely that the FDA will every demand these records:
• The immediate goal of drug discovery is a patent filing. Compliance with Part
11 will ensure that you have full support for patents. You will be able to supply exact records supporting any filing or later patent dispute. The principal investigator, time of discovery, and methodology will be identified beyond any doubt.
• The requirements of Part 11 are all based on common sense and good
workflow. They are really just a codification of good laboratory practice (GLP). Part 11 may be a regulatory process, but it is also a set of guidelines
for improving the security and efficiency of your discovery-stage workflow.
• If you have the opportunity of considering Part 11 compliance during the system design process, the implementation requirements are not that onerous. It is much easier to build the system with Part 11 in mind than to deal with Part 11 remediation later.
• The dividing lines between the discovery and development stages are not
clear-cut, and drug candidates often cycle between the two stages. If your
discovery process is Part 11-compliant, the transition to Part 11- compliant development will be simpler.
• The FDA may decide in the future to further codify the requirements for
drug discovery records. If it does so, it is likely that the requirements will follow the same standards as Part 11. In general, it is a good idea to design any new automation system to comply with the regulations of Part 11. As this paper will show, it is not that difficult to design your system with Part 11 compliance in mind. This paper describes an approach to designing
a drug discovery system that is Part 11-compliant. The ideas described
have been implemented in Coffey Analysis’s Automation Explorer™. You may find this paper useful in reviewing your current systems, in designing a new system, or in evaluating the approaches suggested by software and hardware vendors. Part 11 describes the FDA’s requirements for acceptable electronic records but does not require a particular implementation. In addition, the FDA objects to describing software as “complying with Part 11”. The FDA views the company and its principal investigators as complying (or not) with Part 11. While this paper refers to software as “complying with Part 11” , please view that as shorthand for “software that assists the company and its
investigators in complying with 21 CFR Part 11”, and please view the ideas on implementation as those of Coffey Analysis rather than the requirements of the FDA.
Goals behind Part 11
When electronic records are submitted, the FDA wants to be able to ensure the following:
• That the information is complete – that each record can be tracked to its
source and that related records are connected.
• That the time of information entry and of any modification is noted.
• That the information has not been altered in a manner that obscures the
original information (much as is done with laboratory notebooks, where corrections are signed and where incorrect information is never deleted,
merely amended).
• That only authorized personnel can access the system, and the person
creating, modifying, or reviewing any information is identified and personally
attests to the validity of the process.
• That all experimental protocols are clearly documented and identified.
• That all personnel are trained in the protocols and that documentation is
available.
• That the system has been validated to operate correctly.
• That information can be viewed in either electronic or human-readable format and that records be available “in a timely fashion”.
• That the system be available for inspection,including both hardware and
software. Is there anything in this list that you would not want in your automation system?
A Typical Automation System
A typical discovery-stage automation system will include
• Manual processes (pipetting, weighing, etc.) with protocols defined and
documented for the processes.
• Automated processes that employ laboratory instruments, with protocols
defined and documented for the processes.
• Software that controls the individual instruments and generates data and
reports. This is usually commercial software supplied by the instruments’
manufacturers.
• Decisions and assessments made by scientists based on information generated by the processes, often based upon external third-party software
(e.g., protein sequencing, computer modeling).
• A central relational database that is used to store the information generated in the processes.
• Automation software, often called LIMS (Laboratory Information Management Software) that pulls all of this together.
