In short: in a plant making medicines, cosmetics and supplements, the batch record was printed, filled in by hand and checked line by line by quality assurance. With DataFlows Pharma, the operator runs the procedure on a tablet, scans the materials and signs every step, while QA reviews only the exceptions. The system was validated together with the quality team, based on a validation package delivered with the application.
The context
Batch production, with recipes and master procedures approved by QA, raw materials received in two stages (quantitative and per batch), a quality control laboratory and a warehouse that already worked with stock by location. GMP requirements, periodic inspections and customers who run their own audits.
The problem
- Every transcription in the batch record was a chance for error, and releasing a batch sometimes waited for days because of a missing signature or an illegible figure.
- Deviations, corrective actions and changes sat in separate spreadsheets, with no link to the batches they affected.
- In a potential recall, the list of customers who had received a batch was reconstructed by hand, from several sources.
- Raw materials, laboratory results and quality decisions were in different places.
What we built
- Electronic batch record on a tablet. The master procedure becomes a record for every batch, with targets calculated for its size. Materials are added by scanning the label on the container, with a quantity tolerance check. Every step is signed, and critical steps also have a witness.
- QA review by exception. Deviations from tolerances, uncleaned or uncalibrated equipment are flagged automatically. The review starts from them.
- Audit trail on every field. Old value, new value, author, time and reason. The log cannot be edited, not even by an administrator.
- The quality system in the same application. Deviations, CAPA, change control, OOS and OOT, stability studies and the equipment register, linked to the batches they affect.
- Container-level traceability. Every drum, bag or box has its own label and history. Starting from a batch, you immediately see which customers received it.
- Integrated warehouse. Quarantine, batch statuses and batch allocation are part of the same system, with DataFlows Depo.
Validation
Software isn’t certified, it is validated at the user’s site. We delivered the validation package: user requirements with a traceability matrix, the risk analysis and the position against Annex 11 and 21 CFR Part 11, the data integrity statement based on the ALCOA+ principles, the IQ, OQ and PQ templates and the operating procedures. The system is treated as configured and custom-developed software (GAMP 5, category 5). Qualification was carried out together with the plant’s quality team.
On the development side, the solution has over 800 automated tests written together with the code, and the regulated requirements are linked to tests through the traceability matrix.
How the implementation went
- Process analysis, starting from the existing procedures and batch records, and writing the user requirements together with QA.
- Configuring recipes, master procedures, specifications and approval workflows by the roles in the plant.
- Importing items, stock and partners from the ERP, in a test run.
- IQ, OQ and PQ qualification based on the validation package.
- Training operators, the laboratory and QA, then real batches with us alongside in the first weeks.
Results
- Data is recorded once, at the source; transcription errors disappeared along with the paper.
- Deviations are visible during the batch, not at the review afterwards, and release no longer waits for missing signatures.
- In a recall, the list of affected orders and customers is generated on the spot, grouped by customer.
- At inspection, the batch record with all signatures, the audit trail, deviations and laboratory results are in the same place.
What we learned
Starting in stages mattered: first the warehouse and quality control, then the electronic batch record. The team got used to scanning and electronic signatures on simple flows before reaching manufacturing. And, just as important, the existing batch records stayed the starting point: the steps, parameters and controls remained the plant’s own, only the place where they are filled in changed.
Do you make medicines, cosmetics or supplements and still work on paper? See how DataFlows Pharma works or request a demo.
