Practical guides · Practical guide

What does computer system validation involve in the pharmaceutical industry?

The software is ordered and the implementation partner has a test plan. QA now needs to know how the intended use will be validated. What must the system reliably do in your operation, what could go wrong and how will you demonstrate suitability? A pharmaceutical project puts these questions and the terminology behind them into practice.

Table of contents

At a glance

What is computer system validation (CSV)?

Computer system validation (CSV) is documented evidence that a computerised system is fit for its intended use. Assess the application and configuration in operation, its data flows, the people involved and controlled procedures. In EU GMP, the application should be validated and the IT infrastructure qualified. A justified risk assessment determines scope. A supplier certificate or successful software update does not replace that assessment.

  • Start with the intended use and the relevant decisions.
  • Define responsibilities and verifiable outcomes prior to execution.
  • Plan the operation with: release, changes, and subsequent replacement being interconnected.

First, clarify the intended use, then determine the testing requirements

Computer System Validation, Computerized System Validation, and the spelling "Computer System Validation" denote the same activity herein; CSV is not used to refer to a file format in this article. A specific software vendor is not required. A laboratory, ERP, or document management system may have different implications for product quality, patient safety, and data integrity, depending on its application.

This guide concerns computerised systems in EU GMP manufacturing of human medicines. Relevant uses may occur in pharmaceuticals, biotechnology and testing laboratories. Determine scope separately for medical devices, clinical trials and other regulated activities. Industry membership alone cannot establish a common checklist.

Sources: [1]

Further reading:

Example: Correctly using laboratory test results

A pharmaceutical company has implemented a LIMS. The laboratory records test results, evaluates them against the applicable specification, and issues a test report for further batch assessment. An interface transmits the test status to the ERP. Therefore, it is not merely about whether an input form functions.

The team begins with the following statement: For both sample and batch, it must be clearly traceable which verified result is based on which specification and who has authorized it. Thereafter, the data pathway from sample request to received status is mapped. Raw data from a separate analytics system is identified as a dependency; while such data does not appear in the LIMS, it remains part of the evaluation.

The first usable project status is thus manageable: a description of the use, a system boundary, named data paths and a list of open decisions. It is still open, for example, whether historical results are migrated. This question remains visible instead of being silently considered to be contained or excluded.

From a laboratory task, requirements and tests are derived.

In the example, the requirement "The system manages specifications" is insufficient. The laboratory defines which specification version applies to an order and how a subsequent change is handled. QA and the technical department review the rule, while IT clarifies the technical implementation. Only then can meaningful testing be conducted.

Three potential failure scenarios aid in the selection process: a result is incorrectly assigned to the wrong sample. A report displays an outdated version. Or the ERP system receives a status different from the professionally approved one. The team justifies, for each case, the controls and evidence required. The following table is an excerpt from this evaluation; it is not a complete test matrix.

LIMS Example: From Decision to Evidence
Process questionExample of an inspectionExpected result
Does the value belong to the correct sample? Test similar sample IDs and an invalid assignment Assignment and error response are traceable
Which version applies? Change the result and generate the report again Current status and changes remain identifiable
What happens during a disrupted transfer? Interrupt status transfer and consider repetition. Recipient status, recognition and processing have been clarified

Sources: [1]

Further reading:

Who is responsible for which tasks within the project?

CSV becomes actionable when a specific decision is defined at each transition point. In the LIMS example, the laboratory defines the required workflow. The system owner and IT are responsible for setting up the environment, permissions, and technical operations. The supplier provides details on configuration and functionalities. Quality responsibility assesses the evidence during the agreed release process.

An external CSV partner can support or take over planning, requirement clarification, risk assessment, validation evidence preparation, and result review. Responsibilities will be agreed upon with the operator. An external signature does not transform an ambiguous technical "should" into a viable requirement.

Possible project structure with tangible outcomes
SectionTo be clarified jointlyUsable result
Project start Usage, Limitations, Responsibilities and Existing Documentation Aligned scope and open decisions
Planning Requirements, risks, supplier documentation, and procedures Verifiable requirements and justified validation plan
Implementation System status, test data, execution, and findings Associated results with evaluated deviations
Usage decision Operational readiness and evaluation of results Documented decision in accordance with the own quality process
Operation and Retirement Changes, review, data access, and subsequent archiving Responsibilities and applicable operating procedures

Sources: [1] [2]

Further reading:

Which documentation actually provides meaningful support?

A requirement specifies what is needed. The evidence demonstrates what was actually identified in the specified system state. A clear link must remain between the two. Whether your organization uses individual documents or appropriately structured shared documentation is determined by its own regulatory framework.

For the example, the team addresses the specification status by aligning the requirement, the relevant configuration, and the test result. A finding refers to the affected test. After correction, only the changed status is evaluated. Thus, a reviewer can follow the decision without having to work through a collection of unassigned screenshots.

Existing supplier documentation will be assessed for suitability. If it covers our own usage and the specific system status, it can contribute effectively. Gaps in local configuration, interface, or operational procedures will be specifically addressed. This avoids unnecessary re-evaluation; it prevents redundant and untargeted repetitions.

Sources: [1]

Further reading:

With the first batch release, the work is not complete.

After system startup, the laboratory receives a new reporting requirement. The team assesses which existing statements are affected, implements the necessary changes, and supplements the relevant documentation. In the case of an incident, it must be clearly defined who evaluates the data and the affected decision. Data restoration is also considered during this process.

For future decommissioning, the company will require continued access to the necessary records. These issues are easier to resolve as long as the supplier, data format, and responsible parties remain identifiable. Therefore, the validation plan should align with the intended operation and not conclude merely with the completion of an installation acceptance.

Sources: [1]

Further reading:

Why the product name still does not permit cost estimation

Two LIMS projects can require vastly different levels of effort. In one case, the laboratory process is well-aligned, the configuration is stable, and appropriate documentation is available. In the other, requirements are changing, migration plans are unclear, and test data are missing. The number of users does not adequately explain this difference.

For planning, separate external work, internal contributions, supplier services, and ongoing operations. Request clarification on the assumptions underlying any estimate: which interfaces, review cycles, and iterations are included? Who provides technical decision-making? A schedule only becomes actionable when these inputs are available. This foundational article therefore does not provide a one-size-fits-all project duration.

Further reading:

What to include in your enquiry

Do you need to turn a proposed system into a defined project? We discuss the first decision and the information your team needs to make it.

When Daniel Herrmann Consulting fits your project

You need someone to lead the technical work on a CSV project and work directly with the laboratory or process owners, IT, QA and your implementation partner. We assess what you already have, help carry out the work and transfer the knowledge to your team.

  • You want to assess existing evidence and use it to define the next work packages.
  • You need a technical contact for system boundaries, responsibilities and decisions that are still open.
  • Your team needs to be able to use the agreed processes once the project ends.

Tell us what prompted your project, what evidence you have and where you need support. Software procurement, supplier implementation and final approval for use each have their own owner. We agree these responsibilities before work starts.

Discuss CSV scope and first step

We discuss why you are getting in touch, where your system stands and which decision comes next.

What to prepare with your team before the call

If you want to prepare with your team first, request the worksheets on the linked guides. Enter your name, company, work email and telephone number. After a short call, we will email the relevant documents to you.

Project experience and collaboration

Agree the intended use with your team

At the start of a CSV project, we follow the actual workflow with process owners, IT and QA. We record the decisions the system supports, the evidence already available and the questions still open. This defines the assignment, deliverables and contributions needed from your team. The laboratory example in this article shows how the approach works without describing a particular client engagement.

Questions regarding entry into computerized system validation

Must every IT system in a pharmaceutical company be validated?

Not solely due to its location within the company. Clarify the intended use, the relevant regulated activities, and the applicable requirements. This article addresses the EU-GMP implications; other activities require individual assessment.

Sources: [1]

Is software validated by the supplier already sufficient?

Supplier documentation can contribute. However, the regulated user evaluates whether it meets their requirements and intended use. Local configurations, data pathways, and operating procedures may require additional validation evidence.

Sources: [1]

Is GAMP 5 a law?

GAMP 5 is an ISPE guideline for a risk-based lifecycle approach. It supports the methodical implementation. The applicable regulatory requirements are determined by the specific application and relevant regulatory frameworks.

Sources: [2]

Does 21 CFR Part 11 apply to every pharmaceutical IT project?

No. Determine applicability from the electronic records and electronic signatures covered by § 11.1. Using software or belonging to a particular industry does not, by itself, bring a process within scope.

Sources: [3]

What should we prepare for our first meeting?

A brief description of the intended use, the planned system configuration, the rationale for the initiative, and an overview of existing documentation. Identify open questions. Unprotected patient, personnel, or production data are not required for the initial exchange.

What is the eight-week plan intended for?

He supports planning an already scoped initiative. First assess prerequisites, stakeholder involvement, and evidence requirements for your project. The timeline does not constitute a general performance guarantee for any system.

Sources and review date

The technical entry pertains to EU GMP in the manufacture of human medicinal products. The Commission continues to reference Annex 11 as of the 2011 revision. Project structuring, LIMS example, and procurement guidelines are covered in separate methodological explanations. GAMP 5 is a guidance document. The applicability of US requirements is assessed separately.

  1. European Commission: EU GMP Annex 11, Revision 2011Principle; No. 1–4, 7, 10–13 and 17: Use, Risk, Roles, Evidence, and Operation.
  2. ISPE: GAMP 5, Second Edition (2022)Lifecycle and risk-based approach; no legal obligation to submit.
  3. eCFR: 21 CFR Part 11, specifically § 11.1Scope of electronic records and electronic signatures; to be evaluated separately.

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Do you need a clear CSV task derived from the system concept?

Discuss with Daniel Herrmann Consulting regarding usage, existing evidence, and the next pending decision. Whether a physical inventory check, a limited scope task package, or team support is appropriate is determined by the specific project.

  • 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 →