SOP: Data Safety Monitoring Board Charter and Unblinded Data Firewall

This is a ready-to-use SOP for standing up a Data Safety Monitoring Board (DSMB), also called a Data Monitoring Committee, and keeping the firewall between the unblinded data it reviews and the sponsor’s blinded study team intact for the life of the trial. Replace every <<FILL: ...>> placeholder with your own specifics, set your document numbers and dates, and route it through your normal document control, review, and approval. A worked filled specimen follows. Verify each cited regulation against the current source before you rely on it. This is general guidance to adapt and verify, not legal or regulatory advice.

Document control header

1. Purpose

This procedure defines how <<FILL: COMPANY NAME>> establishes a Data Safety Monitoring Board for a blinded trial, charters its access to unblinded interim data, and maintains the firewall that prevents unblinded results from reaching the sponsor’s blinded study team, so that ongoing trial conduct and data cannot be influenced, consciously or not, by knowledge of an interim treatment effect.

2. Scope

This procedure applies to any blinded interventional trial for which a DSMB or equivalent independent data monitoring committee is planned, whether the committee is standing (covers a program) or study-specific. It covers charter development, membership and independence criteria, the unblinded statistical reporting chain, meeting conduct, and the boundary of what may be communicated back to the sponsor. It does not cover emergency or accidental unblinding of an individual subject for safety reasons, which is governed by <<FILL: SOP-ID for emergency unblinding>>.

3. Responsibilities

4. Definitions

• DSMB / DMC: an independent group, typically clinicians and a statistician with no other role on the trial, that periodically reviews unblinded interim data against a pre-specified plan and issues the sponsor a recommendation on whether the trial should continue, continue with modification, or stop.
• Firewall: the set of people, systems, and procedural controls that prevent unblinded data or any information that would allow the treatment effect to be inferred from reaching anyone outside the DSMB’s unblinded reporting chain.
• Independent unblinded statistician: a statistician, not part of the sponsor’s blinded study team, who has access to treatment assignment and produces the unblinded reports the DSMB reviews.
• Interim analysis: a pre-specified analysis of accumulating data conducted before the planned end of the trial, most commonly for safety and, where the design allows, for early evidence of efficacy or futility.

5. Procedure

5.1 Charter the board before enrollment reaches the first planned look

1. Define DSMB membership, requiring independence from the sponsor’s operational study team, disclosed conflicts of interest, and relevant clinical/statistical expertise.
2. Specify the interim analysis schedule (by calendar time, by event count, or by enrollment fraction) and, where the design uses one, the pre-specified statistical stopping boundary.
3. Name the independent unblinded statistician or statistical data analysis center and confirm they are organizationally separate from the sponsor’s blinded reporting chain.
4. Define exactly what the DSMB may communicate to the sponsor: a recommendation (continue / continue with modification / stop), and any high-level information the charter explicitly permits without revealing the treatment effect.
5. Route the charter for sponsor approval and DSMB member acknowledgment before the first unblinded look; a charter signed after the fact does not protect the trial’s remaining data.

5.2 Produce and review unblinded interim data

1. The independent unblinded statistician generates the interim tables, listings, and figures per the statistical analysis plan.
2. The DSMB meets in a closed session; sponsor blinded study-team members, including the blinded project statistician and the medical monitor managing ongoing site issues, do not attend the closed portion.
3. The DSMB reaches a collective recommendation and records it in minutes that describe the recommendation, not the underlying unblinded data.
4. Underlying unblinded data and minutes detail are retained in the DSMB’s own record, held separately from the sponsor’s blinded trial master file.

5.3 Communicate the recommendation across the firewall

1. The DSMB chair communicates only the recommendation, and any charter-permitted high-level information, to the sponsor’s designated recipient (typically a senior sponsor representative who is also outside the blinded operational team, or a defined limited-access role).
2. The sponsor implements the recommendation, for example a protocol amendment or a stopping action, without learning the unblinded basis for it.
3. If the recommendation is to stop or unblind, a controlled unblinding follows <<FILL: SOP-ID for study-level unblinding>>, executed as a deliberate, documented action, not an incidental leak.

5.4 Maintain and verify the firewall

1. Confirm, at each interim look, that no blinded study-team member has obtained system access to the unblinded statistician’s environment or outputs.
2. Where a sponsor also runs a separately chartered unblinded safety desk for expedited individual case safety reports, confirm its access is scoped to case-level safety data and remains firewalled from efficacy data and from the DSMB’s aggregate results.
3. Where a mid-study decision (protocol amendment, site management action) is made near an interim look, document, contemporaneously, that the decision-makers did not have access to the interim result.

5.5 Handle a suspected firewall breach

1. Any suspected leak of unblinded information to the blinded team is reported to Clinical QA the same working day.
2. Clinical QA and the sponsor’s unblinded statistician assess who received what, and whether the study’s remaining blinded conduct is compromised.
3. Escalate per <<FILL: SOP-ID for deviations>>; assess impact on the analysis population, the SAP, and whether regulatory notification is warranted.

6. Acceptance criteria

• The DSMB charter is signed by the sponsor and acknowledged by DSMB members before the first unblinded interim look.
• Every DSMB member’s independence from the sponsor’s operational study team is documented, including disclosed conflicts.
• The unblinded statistician is named, organizationally separate from the blinded reporting chain, and this separation is evidenced, not just asserted.
• Meeting records capture the recommendation given to the sponsor; underlying unblinded data is retained in the DSMB’s own record, not the sponsor’s blinded file.
• No instance exists where unblinded data, rather than a recommendation, crossed to the blinded sponsor team without a documented, deliberate unblinding action.

7. References

ICH E6(R2)/E6(R3), general principles on trial oversight and participant safety. ICH E9, Statistical Principles for Clinical Trials. FDA Guidance for Clinical Trial Sponsors, Establishment and Operation of Clinical Trial Data Monitoring Committees (2006). 21 CFR 312 (IND recordkeeping, where DSMB records intersect with sponsor obligations).

Confirm the current version of each reference before issue.

8. Record generated: DSMB meeting and firewall confirmation record

9. Revision history

10. Approvals

Filled specimen

The following shows the record completed for an example Phase 3 cardiovascular outcomes trial’s first interim look, so you can see the level of detail an inspector expects. The company, study, and numbers are illustrative.

In this example, the blinded project statistician authored a concurrent protocol amendment for a secondary-endpoint clarification that same month and provided a written statement confirming the amendment rationale did not depend on knowledge of the interim result. That written confirmation, not just the DSMB minutes, is what closes the loop on the firewall for this look.

Common inspection findings this SOP prevents

• A DSMB charter that is undated, unsigned, or was finalized after the first interim look already occurred.
• Meeting minutes that record underlying unblinded data or effect sizes instead of just the recommendation.
• A “blinded” project statistician who, on inspection, had informal system access to the unblinded analyst’s environment or outputs.
• No documented evidence that a mid-study protocol or site-management decision was made independently of an interim result.
• An unblinded safety desk whose access scope was never formally separated from DSMB efficacy data.

How to adapt this SOP

1. Set your document number, owner, and effective date in the header.
2. Attach or reference the actual DSMB charter and statistical analysis plan for interim looks; this SOP governs the process, the charter governs the specific trial’s statistical detail.
3. Name your real independent statistical data analysis center or unblinded statistician arrangement in section 5.1.
4. Point the cross-references in sections 2 and 5.3 to your real emergency-unblinding and deviation procedures.
5. Confirm every regulation in section 7 against the current published version before issue.