7 key requirements for electronic batch records for GMP processes
What is a batch record?
A Batch Record is a document that provides the complete manufacturing history of a pharmaceutical product. It aims to assure the safety and quality of the manufactured product by:
- Providing processing instructions to the operator during the execution of a manufacturing processes
- Documenting the manufacturing process exactly as executed.
A Batch Record, sometimes also called a Batch Manufacturing Record (BMR) or a Batch Production Record (BPR) is a batch specific copy of a Master Batch Record. The Master Batch Record can be seen as the blueprint of the process, while the Batch Record contains the documentation of one single execution of the process. It explains exactly how a lot of a product is manufactured and provides info on who, when, where and how the different processes steps are executed. It also contains all the in-process and release tests that assure the quality of the lot and documents any deviation from the standard process.
1. Data integity
In 2018, 49% of the warning letters sent out by the FDA included a data integrity component [ref]. As the Batch Record is the crown witness of the quality of your product, it is no surprise that the data integrity and the traceability of every component in the document is of the utmost importance. An often used set of principles that define the integrity of a record is “ALCOA” (see table).
There are obvious differences in how paper-based records manage these principles compared to electronic systems. Paper-based batch records might for example require you to manually copy data such as lot numbers and expiry dates, or they might require on the spot calculations for a dilution or a cell count. On a digital record, these tedious and error prone tasks can be replaced by a barcode scanner, picture proofs and automated calculations.
|Paper-based batch records||Electronic batch records|
It must be clear who (or which device), created the data entry.
|An operator signs-off a data entry with his initials and a written timestamp.|
Device date must be copied manually.
|An electronic signature is linked to the operator or device identity and includes a precisely auto-generated timestamp.|
The entered data must be accessible and readable. Additionally it might have to be accessible for a long time.
|Readability depends on handwriting, paper quality and storage conditions.|
Accessibility of physical records is limited and generating backups is tedious.
|Data can be easily read, printed copies can be created. Data can be securely stored and easily backed up on multiple locations.|
The data creation and data logging must happen simultaneously.
|Often requires two operators; one who executes the proces, while another one makes notes||Often requires two operators; one who executes the proces, while another one makes notes
Similar, but data generated by a device can be linked to the record automatically.
The record must contain the data that is collected in its original form. It cannot be a copy of the original data.
|The original record is the paper on which the data was first written on. If the initial data was corrected later, the original (corrected) data must remain available.||Any data representation must be verified to be an exact copy of the data.
Real-time availability of data avoids the need for copies
The record must reflect the reality and must be complete.
|Records need to be verified after the facts. Every manual copy introduces a chance on mistakes.||Idem, but the digital system is able to check user input in real-time which reduces the chance of incomplete or wrongly collected data.|
An electronic Batch Record does not only save a lot of time for the operators, also the number of mistakes are reduced dramatically. In a 150 page batch record it’s not unlikely that an operator misses a signature, or forgets to write something down. A digital batch record will remind the user to fill it out completely. It can also log every keystroke of a user (if needed) in a live updating audit trail. Cryptographic digital signatures ensure you that all captured data cannot be tampered with.
- For electronic records, compliance to 21 CFR 11 or EU GMP annex 11 is critical.
- Make sure to test the business continuity plan in case the software fails.
- Choose a cloud solution if you don’t have an in-house IT team to avoid downtime and data losses. Managing on-premise installed software can be hard and expensive.
2. Data integration
Digital batch records help to streamline data centralisation and data integration, since data from different sources can be linked more easily. Operators in the cleanroom or the lab interact more and more with instruments and devices that require user input or produce relevant process data (e.g. QC readouts, sensor data,…). Currently the data these devices gather is either disconnected from the batch record (e.g. as a paper printout in addendum) or considerable manual user input is required to copy the data into the (paper) batch record itself. In contrast, modern batch recording software communicates directly with these instruments, reducing manual errors and increasing data availability.
This direct connection provides fully traceable data, from work instruction to operator action to raw process data. For example, the operator can automatically verify if the medium he’s using is cleared for use by scanning the barcode on the bottle. Meanwhile, the ERP inventory of the medium is updated in real-time once the operator has used the medium in the production process.
- Think about dataflows and connectivity strategies between LIMS, ERP and batch records well in advance! Once committed to a system with poor 3rd party integration options, it might be difficult to further expand your data management strategy beyond that tool.
- Verify whether your instruments allow communication with a batch record software. This might require a tailor-made integration with your software provider.
3. Flexibility and Scalability
Flexibility and scalability are qualities that are often underestimated for an electronic batch record. Many batch record systems on the market originate in the classical pharma or food industry and have a hard time adapting to the cell & gene market. Given the inherent biological variability in cell & gene manufacturing processes, you require an integrated data management framework that puts you in full control of the process execution, data collection, QC/QA workflows, reporting and analytics.
There is a right balance between flexibility and rigidity for every phase of the product life cycle. Too much flexibility might hamper data integrity, but flexibility is needed at least to a level that it allows planned deviations to the process. Especially in the early phases, a batch record software that enforces a too rigid workflow will result in a tsunami of incident reports and a horde of frustrated cleanroom operators. You don’t want to lose agility due to burdensome IT overhead in the early phases, but other hand, when you’re chasing the next phase towards commercialisation, you don’t want to be stopped in your tracks by a software that is not adapted to the needs of a commercial manufacturer. The software platform should be easily moldable to suit your business needs, today and tomorrow.
- Make sure to select a software provider that understands your processes and is able to guide you with the implementation of your workflows.
- Look for a software framework that combines a plug and play installation for early phase companies, with the option for highly customisable workflows and integrations once you scale. This allows a step-by-step implementation that always focuses on the most critical issue at hand.
- Verify the software update strategy of the supplier and make sure it’s compatible with your change control strategy.
- Verify that incidents are handled gracefully and that they can efficiently integrate with your CAPA processes if needed.
4. Analytics & Reporting
Your highly trained operators are wasting precious time when they copy data from paper-based records to an excel file in order to generate a trend line (e.g.daily CFU counts on contact plates for environmental monitoring). Apart from the lack of added value for this type of tasks, you risk introducing errors during these tedious data manipulations. Whereas in a digital batch record, once the operator has entered the data once, this can automatically generate data visualisations, analytics and reports with a single click of a button.
- Transforming data into real knowledge with the help of data analytics requires collaboration with a software provider who really understands your process.
- Make sure the batch record software supports configurable reporting options. Standard reports are good to get you started quickly, but your reporting needs evolve over time.
- Make sure efficient data export options are available, or even better, a read-only copy of the database is available on which you can run your own BI tools. That way you are sure to fulfill your current and future highly custom reporting needs
A couple of years ago, user-friendliness might not have made the list of requirements for a digital batch record. It should however not be underestimated how critical the user experience of a batch record software is. We are all used to user-friendly apps outside the work environment. Wherever we want, we have all our information at the tip of our fingers. Way too often, when we enter the work environment, suddenly we’re back in the proverbial stone age.
Especially for processes in the cleanroom, care should be taken that the user interface is adapted to its environment (e.g. gloves, disinfectants,…), the work instructions should be crystal clear, and that data input is efficient, intuitive and fail-safe. You want more than a “paper on glass” approach where you basically fill in a static pdf document on a tablet. Also 1 dimensional questionnaire style workflows are often not sufficient to capture complex manufacturing processes that require conditional branches and parallel production steps. These approaches provide little added value over paper-based records. On top of improving your data traceability and solving your reporting bottlenecks, the software should actively improve the workflow of the operators and assist them in their day to day tasks. While an adaptation period has to be foreseen in any case, a bad user experience will result in a higher error rate and frustrated operators.
- Chose for software that can cope with the inherent variability of bio-manufacturing. User-friendliness is closely related to the flexibility of the workflows in this specific industry. Bio-manufacturing processes are inherently variable and a software that is not able to cope with that will force you to file one incident report after the other.
- In order to enhance your operators’ daily workflow, look for a software that allows interactive work instructions, real-time user input verification, autocomplete options, automated calculations, review by exception mechanisms, and (semi-)automated triggering of conditional and parallel workflows.
6. Standardisation & Knowledge management
Improved digital work instructions can be more interactive than their paper counterparts, for example by adding video instructions for the most difficult steps. Therefore a digital implementation will help you instruct your operators more efficiently. Since batch records are digitally distributed to all operators, you can more easily verify that everyone is working on the exact same version, even when managing multiple facilities. Additionally, user input is actively guided by the software and all consecutive analytics or reporting tasks are automated. Such a digital workflow will contribute significantly to achieving ‘Standard Work’ and ‘Right First Time’, even when employee turnover is high.
With standardised data collection, you are preparing yourself to leverage powerful data analytics from an early stage. Data analytics are bound to revolutionise the biomanufacturing field, however large amounts of high quality data are required for that. Setting up an electronic batch record system that produces highly structured data as early as possible allows you to leverage this body of knowledge throughout the rest of the product life cycle.
- Start today with standardising your processes. It’s never too early if it can be done in an agile way.
7. Efficiency & time gains
Paper-based systems suffer from slow handovers, gaps in the chain of custody, data being manually copied from one file to the other, delayed data analytics,… We calculated together with an international biomanufacturing team that a digital batch record system would save 22% of the total man hours spent on a batch. Maybe even more important, since digital systems make handovers and procedures more efficient, a 5% capacity increase of the production facility was predicted.
- Start your digitalisation with processes that allow for quick wins.
- Prime suspects are processes that require highly repetitive manual data logging where the administrative overhead is large (e.g. in-process QC tests).
- Alternatively focus on steps where the reporting is a time limiting factor (e.g. just digitising the reporting for sterility testing, very often the last step in the release workflow, could shorten your ‘vein to vein time’ with 1 or 2 days).