Regulations and compliance · Practical guide

GMP instrument printouts and raw data: when is a paper record sufficient?

A GMP instrument printout may contain raw data or show only part of an electronically stored event. QA and laboratory management first need to establish what the instrument generates and retains. A printed value may need additional context. Even a detailed printout may leave essential parts of the electronic record behind.

Table of contents

At a glance

Which data do we need in addition to the instrument printout?

Compare the instrument’s functions, stored dataset and printout with the purpose of the record. For a simple instrument without electronic storage, the direct printout may be the original record, provided it has the required person, sample and time context. Where relevant electronic data or processing steps are stored, an abbreviated printout may be insufficient. Assess the actual instrument and process.

  • Not the paper format, but the actual data generation determines the required evidence.
  • promptly and traceablely supplement missing sample and personnel accountability.
  • Do not reject electronic source data solely because a printed copy exists.

1. Identify the data function first

This guide concerns GMP for human medicinal products. In section 8.9, PIC/S distinguishes simple instruments that produce paper records directly and store no data from more complex electronic systems. Check the data functions of the actual device. Balances with different designs or configurations can have different capabilities.

Review supplier documentation that includes specific configuration and a traceable test. The following examples describe general device types and do not assume any specific functions of a real product. A visible printer port neither proves nor disproves the presence or absence of electronic logging.

Sources: [1] [2]

2. Display, storage, and printing overlaid

A display may show a current value, while the storage contains additional values, units, or processing steps. A report can only present a selected final value. For use in the process, it must be clarified which information is actually required and where it remains available.

Our own comparison matrix helps provide a concrete answer to this question. It is not a requirement to retain every technical equipment alert independently of its purpose. What is significant are the required data and the contextual information relevant for their interpretation.

Own comparison matrix for a specific measurement procedure
LevelContent to be reviewedOpen evidence question
Device display Value, unit and, if applicable, status Is the display already the used result?
Electronic storage Measurement values, assignment, and relevant processing Which required information is missing from the printout?
Paper printout Actually printed fields Is the probe, device, person, and time point traceable?
Labor record Approach, methodology, and intended use How is the expression linked to the process?
Retention Original or appropriate complete copy Can the required context still be retrieved and read?

Sources: [3] [1] [4]

3. A direct paper original is reasonably complementary

For simple systems without data storage, PIC/S specifies, in direct form, assignment to the responsible person with signature and date, along with the required traceability, for example through sample or batch. The record must be linked to the corresponding manufacturing or testing documentation.

Additions should be timely and unambiguous. Collecting loose notes at a later stage hampers reliable attribution. If multiple measurements are generated, the procedure must clearly indicate which ones were performed and how they are applied. A subsequent handwritten annotation does not automatically transform an unclear value into a reliable measurement record.

Sources: [1] [3]

4. Example: Three weighings, a derived mean value

In this example, Device A is assumed to be a simple device without electronic data storage. It generates three direct printouts weighing 10.02 g, 10.00 g, and 9.98 g. For a specified testing purpose, a mean value is calculated from these. The total sum is 30.00 g, resulting in a mean of 10.00 g.

A later printout showing only “10.00” would establish neither the three individual measurements nor their unit and identity. In a second example, instrument B stores the complete measurements electronically and prints only the mean. The evidence also needs to cover the underlying electronic records and their context. This calculation specifies no actual test method.

Own completed assay matrix for the three weighings
RecordExpected informationPresent in the exampleSupplement or evaluation
Expression A1 10.02 g, measurement reference Value and Unit Assign each probe, person, and process unambiguously.
Expression A2 10.00 g, measurement reference Value and Unit Associate with the same defined test series
Output A3 9.98 g, measurement reference Value and Unit Not rejectable based on the mean value
Calculation 30.00 g ÷ 3 = 10.00 g Controlled calculation logging Individual records reference
Report from Device B Only mean value visible 10.00 g Verify electronic records separately

Sources: [1] [5] [4]

5. Secure legibility over the required time

Paper may fade, become damaged, or detach from its associated sheet. Therefore, evaluate the material, fastening, and storage conditions for the required duration. In particular, for thermal printing, initial readability alone is not sufficient as a long-term evidence of legibility.

An appropriate copy or digitalization may be part of the procedure. Its suitability depends on whether the content and required context are fully and accurately preserved. The mere existence of a PDF file does not automatically permit the destruction of the original nor does it substantiate future retrievability.

Sources: [3] [1] [4]

6. Control the path from print to calculation

When values are manually transferred to another system, an additional source of error is introduced. Specifically verify unit, decimal places, and correct assignment to the appropriate process. The transfer documentation must be commensurate with the significance of the data; the proper functioning of a device does not prevent subsequent transfer errors.

An automated transfer can support this work but requires appropriate controls. Compare representative source values with target records and actual usage. A technically successful transfer does not prove that all required fields were correctly interpreted.

Sources: [2] [5] [1]

Further reading:

7. Handle unassigned assignments promptly

When an old printout can no longer be reliably linked to a specific sample, a plausible entry must not be used to substitute for missing evidence. Review existing independent records and document the actual information they support. A reminder may assist in the investigation, but it is not a fabricated original.

The significance depends on the intended use of the measurement value. The absence of context in a quality-related decision may be more critical than the same deficiency in an unused technical test report. The significance must be evaluated specifically, without automatically classifying any finding as such.

Sources: [3] [1]

8. Device and Record Matching Checklist

Follow a real, representative process from start to end, including the archived documentation. This reveals whether the documented device specification aligns with actual daily operations. A sample form alone does not demonstrate this.

This standalone assistance tool supports the evaluation but does not replace device-specific testing. In particular, the question regarding the presence of electronic original data must not be answered based on a similar device.

  • Specific device configuration and data storage identified.
  • Display, storage, and printing at the same process step compared.
  • Attributable personnel, sample, equipment, and time references securely documented.
  • Original individual values combined with derived results.
  • Manual or automatic transfer appropriately verified.
  • Readability and retrievability assessed over the required period.
  • Unexplained historical gaps are treated as such.

Sources: [1] [2]

Project experience and collaboration

Trace the instrument reading through to the complete record

With the laboratory or production team, we trace a printout to the measurement. We establish what else the device stores and which identifier links the printout to its sample or batch. For a simple balance, the paper original may be central. If the device stores electronic data, those data and their context also matter. We agree how to file, copy and review the records. A legible printout should never obscure missing information.

Frequently asked questions about Instrument printouts and raw data

Is every device printout automatically an original record?

No. A direct output of a simple non-storage device may constitute the original record. For a storage-capable system, the same output can represent only a summary. Verify the actual data function and required content, rather than inferring the status solely from the paper format.

Sources: [1] [4]

Must each balance electronically store data?

No. For simple balances without data storage, the printed copy may serve as the original record. PIC/S requires a signature, date, and traceability, for example through sample or batch number (8.9.1). The FDA expects retention of the printout or a true copy (Question 10). If the balance electronically stores data, such data must instead be evaluated based on their attributes. A paper printout is therefore not automatically equivalent to the complete original record.

Sources: [1]

Which information may be manually added?

Required context such as sample or batch assignment can be reliably supplemented if it is not automatically captured. The timestamp, individual responsible, and link to the output must be reliable. Handwritten entries must not be used to invent missing original information solely based on plausibility.

Sources: [1] [3]

Is a photo of the printout sufficient?

The photograph must be fully complete, legible, and unambiguously linked to the intended use. Additionally, verify permanent availability and controlled storage. A standalone smartphone photograph without a reliable reference does not automatically substitute for the original or an appropriate controlled copy.

Sources: [3] [4]

Can individual values calculated after the mean be rejected?

No, if they form part of the required complete justification. An average value does not describe the underlying variability or data selection. Associate the original individual values with the calculation and their specific application. The method determines which result formation is technically appropriate.

Sources: [5] [4]

Does a simple paper printout require an electronic audit trail?

Not simply because it is on paper. Determine what the instrument actually processes and stores electronically. PIC/S discusses paper-based controls for simple instruments without storage. Assess relevant system controls as well when further electronic processing takes place.

Sources: [1] [2]

How do we assess deteriorating thermal labels?

Assess whether documentation remains legible during the required retention period. If appropriate copies are required, ensure timely arrangement, completion, and assignment of their creation, completeness verification, and allocation. Once copied information has been lost, it cannot be recovered.

Sources: [1] [3]

What do we do with an expression without a sample ID?

Seek reliable, timely evidence for assignment and evaluate its clarity. A subsequent entry must be treated as such. If the assignment remains unclear, this gap must be considered in the decision regarding the use of the result.

Sources: [3] [1]

Is a signed copy always complete?

No. The signature may support personal attribution but does not automatically impute missing measurement values, units, or processing steps. First, verify the required record scope. Formal confirmation and substantive completeness are distinct aspects of the same evidence.

Sources: [1] [3]

Who should perform the data comparison?

Laboratory or department know the use of the values; device managers and IT clarify technical storage and transmission. Together, the actual detection chain can be assessed. A pure visual inspection of the expression by QA does not replace the knowledge of the device function.

Sources: [2] [1]

Sources and review date

Source date: 7 October 2026. Human medicinal products – GMP; FDA comments specifically relate to Drug CGMP within their US scope. Tables and case studies are internal reference tools, not results of completed product testing.

  1. PIC/S: PI 041-1, Data Management and IntegrityPI 041-1, 1 July 2021. Sections 8.9.1–8.9.2 cover direct printouts from simple systems without storage, 8.10 covers retention, 9.1.5 the required data and 9.10 hybrid records.
  2. European Commission: EU GMP Annex 11, Computerised SystemsRevision January 2011; Principle; 4.3 System description; 5 Data exchange; 6 Manual entries; 7 Storage and 17 Archiving
  3. European Commission: EU GMP Chapter 4, DocumentationJanuary 2011; Definition of Records/Reports; 4.1 Media and Connections; 4.8–4.10 Timely Records, Modifications, and Availability
  4. FDA: Data Integrity and Compliance With Drug CGMP, Questions and AnswersFinal Guidance, December 2018; Q1(b)–(d) metadata and static/dynamic data; Q9–Q10 appropriate copies and electronic original records
  5. European Commission: EU GMP Chapter 6, Quality ControlRevision 2014; 6.16 critical evaluation of results/calculations and 6.17 inspection content of test records

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Trace device data within the actual workflow process

We help you assess paper and electronic data pathways, establish required context, and implement appropriate controls for your team.

  • Computer system validation →

    Bring your GxP systems into validated operation, from laboratory software, MES and eQMS to ERP. We define the scope, plan risk-based tests using GAMP 5 and hand the evidence to QA for release.

Send a project enquiry Customer testimonials →