Practical guides · Practical guide

Routine CSV support: organising external help and clarifying who decides

After implementation, updates, new roles and questions about earlier evidence continue. Who evaluates them when an external partner supports operations? A monthly allowance of hours leaves that open. Define tasks, deliverables and decisions your own team retains alongside the time available.

Table of contents

At a glance

How should ongoing CSV support be agreed?

Define the systems, tasks and internal owners covered by external CSV support. Agree how changes, incidents, reviews and evidence queries are reported, prioritised and handled. Assign external analysis, operational decision-making and technical execution separately. Include required client contributions and usable documentation in the agreement. An hours allowance describes available time. The client remains responsible for controlling and assessing outsourced work.

  • Operations require a controlled access mechanism to actual changes.
  • Changes, disturbances, and traceability questions may require different handling procedures.
  • Reaction time, processing, and decision-making are to be separated and defined.
  • The collaboration should ensure that knowledge remains accessible within the company.

Which systems and responsibilities are managed?

Monitor systems with respect to purpose, status, and relevant stakeholders. Describe the activities that should be continuously supported. For example, a service provider may prepare change evaluations, review documentation, or support appropriate review sessions. Technical administration and troubleshooting, on the other hand, may reside with other parties.

Address existing gaps and excluded activities. Ongoing support does not automatically constitute a requirement to recreate all historical documentation. If such work is required, it should be clearly assessed and appropriately planned. Otherwise, a one-time remediation project may go unnoticed and disappear from the operating budget.

Sources: [1] [2]

How to Ensure Changes and Inquiries Reach the Correct Point of Contact

Name where changes, faults and queries are recorded. The message should make the affected systems, the cause and known effects understandable. A tool is helpful for this, but not a process by itself. The decisive factor is who sees a message and decides on the further processing.

Capture supplier information and planned technical changes. Not every relevant change starts as an IT ticket. A new report, modified calculation logic, or additional interface may impact the intended use. The entry must appropriately account for these diverse pathways.

Sources: [1] [3]

Divide analysis, decision-making, and implementation clearly.

Assign responsibility for each operational task, specifying who analyzes, technically reviews, and makes the business decision. The matrix supports this alignment; actual authorities are determined within your organization.

The collaboration must not result in any inquiry being left solely with the external handler, while such handler lacks essential information or authorities. Conversely, an internal release step must be supported by a clear, professional rationale.

Divide analysis, decision-making, and implementation clearly.
TaskPossible external contributionTo be clarified internally
Change Prepare impact assessments and required evidence prior to implementation Purpose and responsible decision-making
Disruption Investigation and affected data paths supported Operational Measures and Responsible Parties
Review Integrate relevant information Assessment and further tasks
Test Prepare or execute appropriate cases Requirements, Access and Result Decision
Validation maintenance Update affected documentation Applicable version and approval process
Escalation Clearly communicate identified risks Decision and required resources

Sources: [1] [2] [3]

Evaluate changes based on their impact

A version change serves as a reason for review, but is not the only relevant modification. Roles, master data, formulas, reports, and interfaces can also alter the current usage. Assess which requirements and controls are affected and identify any additional insights needed.

From this assessment, appropriate validation evidence and documentation adjustments follow. A blanket full validation after each update is no more justified than the assumption that a minor update would have no impact. Common interdependencies should be clearly considered and documented. The external service provider must have a sufficient understanding of the actual change to support it.

Sources: [1] [4]

Further reading:

Associate data and process trails with deviations

A technically resolved failure does not automatically clarify which records were affected. For relevant cases, investigate the time period, data pathway, and potential impact on decisions. Separate verified facts from unconfirmed assumptions.

A finding may require immediate operational action, further investigation or a change. Annex 11 section 13 calls for reporting and assessment of all incidents, including system failures and data errors. For critical incidents the root cause should be identified and form the basis for corrective and preventive actions. Classify the event under the applicable procedure, record the rationale and account for relevant relationships.

Sources: [1] [5]

Further reading:

Planning example: A new report filter alters the statement

Suppose a report starts using a different filter after a change. The support team checks what the report is expected to show and whether previously used outputs may be affected. Suitable reference data helps compare the expected and actual results.

In the illustrated planning example, a limited correction may be acceptable upon targeted evaluation. If the investigation reveals a more extensive effect, additional data assessment and a reassessment will be required. The actual work needed will depend on the nature of the change and its consequences.

Sources: [1] [4]

Agree on clear allocation, priority, and response procedures

An agreed reaction time initially indicates that a notification has been acknowledged. It is not automatically a commitment to a final resolution, particularly where technical input or subject-matter expertise is lacking. Document what has been confirmed, investigated, and provided, as well as any dependencies that exist.

A possible structure for the operating agreement is shown in the list at the end of this section.

Plan recurring tasks and additional change requirements separately. An available monthly buffer is not equivalent to full coverage of all activities. Agree on how major events shall be defined and how additional work shall be decided. This ensures clarity regarding workload and remaining internal involvement.

  • Supported systems, included functions, and excluded purification projects.
  • Incoming channel and priority rationale based on process or data impact.
  • Reaction, analysis, technical processing and operational decision-making with respective responsible roles.
  • Recurrent tasks, particularly periodic evaluation in accordance with Annex 11, Section 11 and change control as specified in Section 10.
  • Escalation, reporting and records available to the contract giver under Chapter 7, paragraph 7.16.
  • Evaluation and approval prior to sub-assignment in accordance with Chapter 7, Section 7.11, as well as knowledge transfer and regulated contract termination.

Sources: [2] [1]

So that evidence and knowledge remain within the organization

The handling should be traceable and understandable within the organization. A clear and logical workflow should link the rationale, assessment, decisions, and outcomes. Critical information must not be retained solely in personal messages or with external personnel.

Regularly review open tasks and the alignment of collaboration. Your team should be able to clearly explain the current system status and the actions taken in the event of a relevant change. If internal expertise develops, external support can be more strategically directed toward specialized topics or review activities. Appropriate oversight ensures clarity and enables continuous improvement.

  • Clearly specify the inventory and listed tasks.
  • Reliably capture changes from the technical and IT departments.
  • Assign analysis, operational decision, and technical correction.
  • Plan effort based on actual triggers and dependencies.
  • Document results such that the company can continue working with them.

Sources: [1] [3] [2]

Further reading:

Project experience and collaboration

Agree who handles day-to-day CSV work and decisions

Before arranging ongoing support, we review the systems in scope and typical changes or incidents. Using one case, we agree which analysis the external specialist will do and who will make decisions or carry out technical work internally. This also lets the team practise the handover. A single system needs different arrangements from a varied portfolio involving several teams. The allocation of specialist time should follow the agreed tasks.

Frequent Questions on External CSV Support in Operations

Does an ongoing CSV agreement assume our entire responsibility?

No. It can clearly distribute tasks and contributions. Chapter 7 Nos 7.4 and 7.8 leaves responsibility for the management of outsourced activities and the evaluation of results with the client. Internal responsibilities, expert decisions and appropriate monitoring remain to be organized. The contract should describe this division of work.

Sources: [1] [2]

Must a complete revalidation be performed after each update?

First, assess what changes result from the update and the potential impact on your usage. The appropriate testing and additional documentation depend on these changes. Incorporate joint technical and functional dependencies to ensure that the evaluation neither overlooks relevant implications nor redundantly repeats all considerations.

Sources: [1]

Is every deviation automatically a CAPA?

No. According to Annex 11, Section 13, all incidents must be reported and evaluated. For critical incidents, the root cause must be identified and used as the basis for CAPA. A technical solution alone does not address the impact on existing data or the required follow-up actions.

Sources: [5] [1]

Is a guaranteed response time equivalent to a guaranteed solution?

No. A solution may require additional technical effort, information, or decisions. Independently define what reaction and handling entail. Escalation paths and missing prerequisites should also be clearly established.

Can we utilize support without a fixed monthly quota?

Single assignment or a limited allocation may be appropriate depending on the scope of responsibilities. Assess availability, recurring work, and significant change events. The suitable approach depends on your requirements and agreed performance.

How do we ensure that knowledge does not remain with the service provider?

Arrange accessible, clear documentation and joint review of key decisions. Enable internal personnel to take on appropriate responsibilities and explain results. A subsequent handover should remain feasible with existing documentation.

Sources: [3] [2]

Sources and review date

Primary source date: 17 October 2026. The article addresses ongoing support of computerized systems within the human medicinal product GMP environment. Task matrix, capacity assessment, and reporting example serve as internal organizational tools. No universal contract terms, fixed response times, or blanket guidelines for revalidation are asserted.

  1. European Commission: EU GMP Annex 11, Computerised Systems, January 2011 revisionHuman medicinal products – GMP; in particular, §§1–4, 7, 10–13, 16–17; specific application for each section.
  2. European Commission: EU GMP Chapter 7, Outsourced ActivitiesOutsourced GMP-related activities, suitability, responsibilities, agreements, and monitoring.
  3. European Commission: EU GMP Chapter 2, Personnel, effective from 16.02.2014Personal, role-specific training and assessment of practical effectiveness; official EudraLex entry with original chapter.
  4. ICH Q9(R1): Quality Risk Management, official text from the European CommissionRisk-driven decisions, appropriate formalities, and handling of subjectivity; own matrices are implementation proposals.
  5. European Commission: EU GMP Chapter 1, Pharmaceutical Quality SystemQuality management, accountability, root cause analysis, CAPA and change control; no project budget allocation.

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Identify three typical triggers for operational support

Bring a change, a deviation, and an open validation question from your system inventory. This enables clarification of where external support can specifically assist and which internal contributions remain necessary.

  • 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 →