Hybrid Laboratory Records: Paper, Electronic Data and Responsibilities

An analysis may begin with a weighing entered on paper, continue in the CDS and finish with a signed printout. Its record does not necessarily coincide with any one medium. A hybrid system works when each activity’s evidence can be identified and its components remain linked over time.

Paper does not automatically become the overriding original because it bears a signature. Scanning a sheet does not automatically complete the electronic record either. Define content, origin, versions, controls and responsibilities before deciding what to approve and retain.

Describe the complete record and its components

Start with the sample pathway: receipt, preparation, measurement, processing, review and decision. At each step identify where data originate, where they are copied and what is needed to interpret them. Distinguish original recording, controlled instruction, working copy, report and approval: similar media may serve different purposes.

A weighing first recorded on a controlled form and an electronically acquired signal have different origins. A report may summarise a result without including necessary native data, processing parameters or audit trails. Calling a printout “official” does not remove other required evidence.

Describe the whole system, including manual and software interfaces. PIC/S highlights hybrid complexity and encourages replacement where possible; meanwhile, each component needs effective controls. Digitising without clarifying the process may simply relocate the same gaps.

Link identifiers, versions and responsibilities

Use references supporting movement from paper to electronic data and back: sample identifier, sequence or dataset, method version, processing version and relevant report. A folder named only by date or a link to the “current result” may become ambiguous after reprocessing. The relationship must identify the version actually reviewed.

This matrix proposes an illustrative analytical record; assign roles and retention according to the organisation and applicable obligations. It specifies no universal retention period.

Control transcription, printing and scanning

Identify critical manual steps: copying a mass, transcribing a result or selecting a file may change value, unit or sample association. Define suitable additional verification, such as a second operator or a validated electronic check in context. Record what was compared against which source, not just a generic “checked” signature.

For printouts, ensure sample, version, units, pages and exclusions are identifiable. A static printout does not automatically reproduce dynamic data functionality. Do not use it to avoid reviewing necessary electronic information.

For scans, check completeness, order, legibility, annotations and linkage to the original. OCR can alter numbers or symbols: recognised text should not be treated as verified data without checking. Use of a verified copy requires a documented process preserving necessary content and meaning; creating a PDF alone does not authorise original disposal.

Define what is reviewed and approved

The procedure must explain how the reviewer obtains the full set and correlates its components. Relevant electronic data and significant events must be reviewable alongside the paper report. Result review should relate to available, identifiable evidence, not information inaccessible to the signer.

Make signature meaning and scope explicit. Without a controlled link, a handwritten signature on the last page does not demonstrate approval of every file in the system. Where different people check preparation, processing and release, assign activities and handovers without gaps between responsibilities.

For a discrepancy, retain both representations, identify its origin and assess effects on decisions. Corrections must preserve readable previous content and the relevant reason; electronic history remains available as required. Never silently replace a signed printout to match the latest result.

Simulated case: the signed printout predates reprocessing

Report P01 contains the result from CDS processing v1 for sample S84 and is signed after review. Later an analyst reprocesses the same dataset, creating v2 with a different parameter. The approved printout still documents v1: its signature does not automatically transfer to v2.

The laboratory retains P01, v1, v2 and the audit trail, documents the reason and authorisation for reprocessing and assesses the difference. The reviewer examines the relevant set, identifies accepted processing and records a new decision according to procedure. If an earlier decision relied on v1, its impact is assessed; a new printout does not erase history.

The file clearly distinguishes both states and supports retrieval. Keeping both versions is not an error to conceal: it is needed to reconstruct the sequence. This case is simulated, not a test performed on a real CDS.

Retain the set and test retrieval

The data archiving plan should cover paper, electronic components and their relationship index. A useful test starts from the sample identifier, retrieves the full file and then checks the reverse route. Changes to location, software or responsibility must not break the relationships.

EU GMP Chapter 4, January 2011, addresses hybrid relationships and record management. PIC/S PI 041-1, 1 July 2021, §9.10, addresses controls and combined review. FDA Part 11 Scope and Application, final 2003, and the Data Integrity guidance, final December 2018, direct attention to actual use and record content: calling records “hybrid” does not automatically exclude electronic components from applicable requirements.

Sources consulted 2 October 2026. Control is convincing when another authorised reviewer can reconstruct activities, versions and decisions without relying on the analyst’s memory.