Data preparation for an adaptive trial interim analysis is not a pre-interim sprint — it is a continuous operational discipline that must begin at study startup. The five requirements that must be true at the moment of interim analysis are: all key efficacy and safety variables entered and queries resolved, no outstanding queries on primary endpoint data, biomarker data confirmed accurate, RTSM dispensing records reconciled with EDC treatment records, and protocol deviations in the analysis population assessed and documented. The ROBIN project (BMC Medicine, 2025) found that these standards are only achievable on the timeline adaptive designs require when data management operates as a continuous process — not a batch activity triggered by proximity to the interim window.
—
Why Is Data Readiness the Adaptive Trial’s Efficiency Condition?
The efficiency advantage of an adaptive trial design is realized at interim analysis. If the data is not ready — if queries are unresolved, key variables are missing, or the database has not been cleaned to the standard required for a reliable decision — the interim analysis delivers its result late, on uncertain data, to a committee that cannot act with confidence. The design efficiency is gone.
Data readiness for interim analysis is an operational problem, not a statistical one. It is solved in the months before the interim window, not in the weeks before.
—
What Does “Data Readiness” Actually Mean at the Interim Window?
At the point an interim analysis is conducted, the following must be true for the analysis dataset to support a reliable adaptive decision:
- All key efficacy and safety variables for patients in the interim analysis population are entered, queried, and resolved
- No outstanding queries on primary endpoint data
- Biomarker data (where eligibility- or analysis-defining) is confirmed accurate and complete
- The RTSM dispensing records are reconciled with the EDC dose/treatment records
- Protocol deviations affecting the analysis population are assessed and documented
This is not the state that most oncology trials are in on a rolling basis. It is the state that must be achieved specifically for the interim window — which means it must be built toward continuously, not created on demand.
Kyle Hanson, Director of Clinical Operations, Sitero:
“It means several things have to be simultaneously true, and that’s the hard part. The database has to be locked or at minimum query-resolved to a pre-specified threshold for the analysis population. All primary endpoint data for patients who have reached the analysis landmark must be entered and verified. Any data management issues flagged in the preceding weeks need to be resolved, not deferred. And critically, the unblinded statistician and the DSMB charter have to be aligned on exactly which data cut standard applies — because ‘data readiness’ is only meaningful relative to a pre-defined rule. The most overlooked element is usually outcome ascertainment lag — patients who should be evaluable for the interim are technically enrolled but haven’t had a response assessment yet because of scheduling delays. If that number is too high, the DSMB can’t make a reliable decision, and you either wait or you make a decision with an underpowered dataset.”
—
How Do You Build Toward Interim Readiness Continuously?
The ROBIN project’s most actionable recommendation for data readiness is treating the interim analysis as a routine operational milestone that the entire data management process builds toward — not as a one-off event (BMC Medicine, 2025).
Four continuous data management practices that support interim readiness:
- Define key variables upfront. At study startup, identify the variables required for the interim analysis dataset and document prioritized entry and cleaning expectations for them specifically.
- Set site-level entry expectations for priority variables. For primary endpoint data, a 48-hour data entry expectation is the operational standard in ROBIN project case studies, compared to end-of-month batch entry in conventional programs.
- Run automated validation checks daily. Batch querying at the end of a data cycle is incompatible with interim readiness. Daily automated checks on key variables identify issues while they are still easy to resolve.
- Document the data lock process for the interim dataset. Assign clear roles and timelines. The data lock for an interim is a managed operational event with a clear owner, not an ad hoc response to the statistician’s request.
—
Why Should Every Adaptive Program Do a Dry Run?
The ROBIN project recommends conducting a dry run of the interim analysis process using blinded trial data before the first real interim (BMC Medicine, 2025). This is not a standard practice in most CRO operations. It should be.
A dry run validates that:
- The data extraction process works as intended on the actual database structure
- The statistical programs run cleanly on the real data
- The team can complete the full interim analysis process within the expected timeframe
- Roles, handoffs, and communication protocols function as designed
Problems discovered in a dry run are fixable. Problems discovered during the actual interim are not. The cost of a dry run is small relative to the cost of a delayed or invalidated interim analysis decision.
—
Conventional vs. Adaptive Trial: Data Management Comparison
| Data Management Practice | Conventional Trial Standard | Adaptive Trial Requirement |
|---|---|---|
| Data entry expectation | End-of-month cycle | 48-hour entry SLA for key variables |
| Query generation cadence | Weekly or biweekly batch | Daily automated validation on priority variables |
| Data lock process | Defined at end of study | Pre-defined for each interim dataset with clear role assignment |
| Interim dataset specification | Not applicable | Defined at study startup; key variables identified and prioritized |
| Statistical team access | Single blinded team throughout | Separate blinded/unblinded teams; data firewall documented |
| Dry run | Not standard | Recommended before first interim (BMC Medicine, 2025) |
| Closeout data management | Standard timeline | Ongoing high-performance discipline transfers directly to faster closeout |
—
What Is the Data Management Team’s Expanded Role in Adaptive Programs?
In adaptive trials, the data management team’s role is more operationally demanding than in conventional studies. They must:
- Maintain clean data on a continuous basis, not in cleanup sprints
- Coordinate with the biostatistics team on interim dataset specifications
- Manage the firewall between blinded and unblinded data when separate statistical teams are involved
- Support rapid database updates following adaptation decisions
This is a resourcing and scoping requirement that must be built into the CRO contract, not assumed to be covered by standard data management rates. Sitero’s oncology programs have achieved a 60% decrease in data management closeout timelines (Sitero oncology program data) — an outcome that reflects the continuous data review discipline directly transferable to interim analysis readiness in adaptive programs.
Learn how Sitero supports adaptive oncology trial design: Adaptive Oncology Trial Design
—
Frequently Asked Questions
Q: How far in advance of an interim analysis should data entry and cleaning reach the required standard?
The ROBIN project’s framing is useful here: the standard should be maintained continuously, not achieved in advance of specific dates. If data entry is running at a 48-hour standard and daily validation is running throughout, the data is always at or near interim-ready state. If the target is to achieve readiness by a specific date, the standard compresses the timeline and introduces the risk of a data quality sprint that misses issues.
Q: How do you handle blinded vs. unblinded data when the same CRO manages both the study and the interim analysis?
Separate statistical teams with documented firewalls are the standard approach. The blinded team manages the study database; the unblinded team accesses only the interim analysis dataset through a formal, documented data transfer process. The firewall documentation must be auditable — regulators have reviewed CRO processes for blinding integrity in adaptive programs.
Q: What happens to the data management closeout timeline when a trial has run multiple adaptive interims?
The closeout timeline is typically shorter, not longer, because continuous data management discipline means the database is never significantly behind. Sitero’s 60% reduction in data management closeout timelines (Sitero oncology program data) reflects this effect — programs that maintain real-time data standards have less remediation work at closeout regardless of the number of interims.
—
Planning an oncology trial with adaptive interim analysis requirements?
Sitero has supported 200+ oncology studies across 67+ countries. Talk to an oncology trial expert to discuss your protocol.
—
References
- Sitero. Oncology Program Operational Data. Internal dataset. 200+ oncology studies across 67+ countries. sitero.com/oncology/
- Hanson K. Director of Clinical Operations, Sitero. Expert interview conducted for this article. April 2026.
- Mossop H, Walmsley Z, Wilson N, et al. Practical guidance for conducting high-quality and rapid interim analyses in adaptive clinical trials (ROBIN project). BMC Medicine. 2025;23. doi:10.1186/s12916-025-04362-x
Drug supply management in an umbrella trial requires RTSM configuration that explicitly models arm expansion and arm closure scenarios — not supply planning for the initial design with manual adjustments [...]
Running a basket trial operationally requires infrastructure that treats each tumor type cohort as a distinct operational unit within a shared governance framework: cohort-specific site selection (not study-level), a central [...]
