Data Integrity in Pharma: ALCOA+, Regulators, and the 483 Failures

Data integrity is the degree to which data are complete, consistent, accurate, trustworthy, and reliable throughout their entire lifecycle, from the moment they are generated, through processing, review, reporting, and retention, to eventual disposal. In a pharmaceutical manufacturing context it is the assurance that the record of what happened on the floor faithfully represents what actually happened, and that it stays that way for as long as the record is kept. When a regulator questions data integrity, they are questioning whether the evidence of quality can be believed at all, which is why it has become the dominant theme in FDA inspection findings and warning letters.

The reason the topic feels abstract is that integrity is not a single control you can install; it is a property that emerges from many controls working together. The industry compresses those expectations into a memorable framework, ALCOA, extended to ALCOA+, and the regulators (FDA, MHRA, WHO, PIC/S, and in India, the revised Schedule M) each express the same expectations in their own guidance. This guide walks the principles, the regulatory expectations, the failures that recur in 483s, and how electronic records turn the principles from aspiration into enforcement.

The FDA sets its expectations in the guidance Data Integrity and Compliance With Drug CGMP: Questions and Answers (2018), framed around the existing cGMP requirements in 21 CFR Parts 210, 211, and 212. It defines data integrity through the ALCOA attributes, expects audit trails for GMP-relevant electronic records to be reviewed with the same rigour as the records themselves, expects access controls that prevent unauthorised changes, and treats failures such as shared logins, disabled audit trails, and uncontrolled spreadsheets as serious cGMP violations. The controls in 21 CFR Part 11 are the mechanism through which these electronic-record expectations are met.

The MHRA publishes the GxP Data Integrity Guidance and Definitions (2018), which is among the most cited statements on the subject. It defines ALCOA and the ALCOA+ extensions explicitly, stresses that effort and control should be proportionate to the risk the data carry to patient safety and product quality, introduces the concept of the data lifecycle and of designing systems so the right behaviour is the easy behaviour, and is pointed about the limits of relying on people rather than system design. The WHO guidance, Annex 3 of Technical Report Series No. 1019, “Guideline on good data and record management practices,” covers the same ground for the WHO prequalification and member-state context, and PIC/S PI 041 provides the harmonised inspectorate guidance for participating authorities. Read across all four, the message is identical: integrity is a lifecycle property, controls should be risk-based, and the system should make the trustworthy path the path of least resistance.

Reading across published FDA 483s and warning letters, the recurring data-integrity failures cluster into a short list: shared or generic logins that break attributability; audit trails that were disabled, not reviewed, or did not capture deletions; original electronic data deleted or overwritten after acquisition; “testing into compliance,” where failing results are discarded and only passing re-tests reported; back-dated or fabricated records; and uncontrolled spreadsheets standing in for validated systems. Our own inspection write-ups show the pattern concretely: the data-integrity findings at Tyche Industries, the data-integrity observations at Immacule Lifesciences, the environmental-monitoring data-integrity findings at Natco, and the QC data-integrity observations at Catalent each trace to one or more of these root causes.

What links them is that paper and hybrid systems cannot enforce the principles; they can only request them. Electronic records, properly controlled, enforce them by design: a unique login makes every entry attributable; capture at the instrument makes it contemporaneous and original; a non-defeatable audit trail makes it consistent and complete; and managed retention keeps it enduring and available. This is why the FDA’s data-integrity guidance and Part 11 are so tightly linked: the controls that satisfy Part 11 are the same controls that make ALCOA+ real on the floor.

Schedule M sits within India’s Drugs and Cosmetics Rules and defines the Good Manufacturing Practices and requirements for premises, plant, and equipment for pharmaceutical products, administered by the Central Drugs Standard Control Organisation (CDSCO). Its revised version modernises Indian GMP substantially: it introduces a formal Pharmaceutical Quality System, quality risk management, and a product quality review, and, most relevant here, it makes data integrity and the control of computerised systems explicit requirements rather than implicit ones, aligning the domestic standard far more closely with WHO-GMP and the expectations in EU GMP and PIC/S. For a manufacturer, the practical effect is that the ALCOA+ principles and computerised-system controls described above are now domestic regulatory expectations, not just the price of exporting to the US or EU.

The implementation has been phased by manufacturer size, with different timelines for larger manufacturers versus small and medium enterprises, and CDSCO has revised those timelines over the course of the rollout. Because the exact notification date and the current compliance deadlines have shifted during implementation, this guide describes the substance of the revised requirements rather than asserting specific dates; the precise effective dates applicable to a given manufacturer should be confirmed against the current CDSCO notification before they are relied upon. The direction, however, is settled: revised Schedule M brings explicit data-integrity, quality-system, and computerised-system requirements into Indian GMP, so a site that has built genuine ALCOA+ controls for FDA or EU purposes is already most of the way to meeting it.

The durable way to satisfy data integrity is to remove the opportunities for it to fail: to capture data at the source, attribute every entry to a unique identity, record every change in an audit trail that cannot be switched off, and retain records in a controlled, retrievable form for their full life. That is exactly what a purpose-built manufacturing execution system does across batch execution, electronic logbooks, and QA review: guided steps that must be performed in sequence, instrument data captured without transcription, electronic signatures linked to records, and a complete 21 CFR Part 11 audit trail behind every action. The result is that ALCOA+ stops being a training topic and becomes the way the floor already works.

Getting there does not require a single high-risk cutover. The pragmatic route is the phased one described in our MES implementation strategy: start with the highest-risk records (batch records and logbooks), prove the controls, then expand, so data-integrity risk falls from the first phase rather than waiting for a complete deployment. For the underlying batch record itself, the electronic batch records guide describes how the record is built as the batch runs, which is where attributability, contemporaneousness, and originality are won or lost.

Data integrity and cleaning validation are the two areas where the gap between a written procedure and the enforced reality is most visible to an inspector, which is why the electronic batch records and cleaning validation guidelines pillars sit alongside this one. The common thread is that integrity is won at the point of capture: a record that is attributable, contemporaneous, and original when it is created rarely becomes a finding later, while one reconstructed from memory at end of shift almost always can. Build the controls into how the work is done, and ALCOA+ stops being an audit anxiety and becomes a property of the record itself.