Records, documentation and data integrity

A useful record describes the activity closely enough that the next person does not need to rely on memory. The good documentation practice overview identifies contemporaneous documentation as a basic expectation. In practical terms, the record belongs beside the task. A distant form that is filled out later creates a gap between observation and evidence, even when the worker intends to be accurate.

Consider a container transfer. The meaningful record connects the material identity, its status, the equipment used and the relevant observations. Writing the result after several unrelated transfers increases the chance of mixing the details. An approved procedure can identify when entries are made and who checks them. A completed signature line supports accountability only when it represents a real activity. Familiar handwriting is not a substitute for a clear connection between the entry and the task.

Good form design helps. Space for observations, a readable sequence and instructions for unusual conditions reduce pressure to invent workarounds. A record that forces the operator to choose an inaccurate answer needs attention through the document system.

An honest correction improves a record. Concealing the earlier entry removes information that a reviewer may need. The documentation overview describes signed and dated handwritten modifications, reasons for changes where appropriate, and preservation of the altered text. The facility’s approved correction procedure determines how these principles are applied to its records.

For example, a miswritten equipment identifier can be corrected so that both the mistake and the intended identifier remain understandable. The explanation should describe the actual reason for the change. It should not pretend that a check occurred earlier than it did. If the error raises doubt about which equipment was used, correcting the written identifier alone does not answer that operational question.

Electronic systems make information easier to search and combine, but they can also hide a weak control behind a polished display. Part 11 on electronic records and signatures addresses controls including authorized access, protection of records and electronic audit trails. The audit-trail provision states that changes must not obscure previously recorded information. The trail is part of the evidence needed to understand an electronic record.

A shared account weakens the connection between an action and the individual who performed it. A change history that exists but cannot be understood during review provides limited practical help. Teams commonly evaluate whether records can be retrieved, whether earlier values remain visible and whether permissions match assigned responsibilities. Those questions should include routine users, administrators and reviewers.

Electronic signatures also have a meaning. Approval, review and authorship are different activities. The system and the procedure need to make the intended meaning understandable rather than treating every electronic confirmation as interchangeable. A screen confirmation cannot repair an incomplete underlying record.

A reviewer needs the information that explains a conclusion, not simply a selected result. A laboratory value without its preparation details or relevant observations may be difficult to assess. An unexplained discrepancy between a written record and an electronic history may point to an entry error, a process problem or a system-control weakness. The investigation determines what happened; the appearance of the record does not.

This is why review has to extend beyond checking whether fields contain text. Related evidence should tell a coherent account of the activity. A record might identify equipment that was unavailable, show an activity outside the approved sequence, or contain an unexplained repeated test. Each inconsistency raises a question that a signature alone cannot resolve.

Teams can make review more focused by defining the evidence required for each decision. The expectations for a routine material check differ from those for a laboratory investigation. Both need an accountable reviewer and a way to document unresolved questions. A clear connection between the supporting information and the decision makes later review more dependable.

The account of Ranbaxy Laboratories describes poor record keeping identified during an audit and subsequent findings about falsified application data. It states that FDA halted reviews of the company’s drug applications in February 2009 after finding frequent data falsification at a manufacturing plant. The account also records a guilty plea in May 2013 involving adulterated drugs and misrepresentation of clinical generic drug data.

These are specific facts from that account. They do not establish the status of any other manufacturer. The manufacturing lesson is narrower: evidence that appears complete can still be unreliable if the underlying data or the description of its origin has been changed. Review systems need to examine how information is created and handled, not only whether a final document has the expected appearance.

A similar principle applies to an ordinary plant-floor discrepancy without implying misconduct. If records conflict, the investigation should resolve the conflict from evidence. Pressure to close a review is a reason to strengthen the process, not to remove an inconvenient observation.

Reliable documentation depends on the whole work system. Operators need accessible current instructions, practical entry methods and enough time to record what matters. Reviewers need related information in a usable form. Electronic system users need permissions and training that fit their duties. Supervisors need a clear route for responding when the record cannot express what happened.

A useful improvement discussion follows a record through its life: creation, checking, use in a decision, storage and retrieval. Each handoff is an opportunity for information to become detached from its context. Clear links between supporting material and the main record reduce that risk. Controlled revisions help prevent an obsolete form from quietly returning to use.