Clinical data management: from a protocol question to usable data

Clinical data management is the organized work of making study information complete, consistent, traceable, and suitable for its intended analysis. At site level, it is also a daily operating problem: staff need to know what to record, where to record it, and how to resolve a question without guessing.

A clean-looking database is not enough. The important question is whether the information accurately represents the study activities and can be understood with its supporting context. The protocol, data management plan, system instructions, and assigned responsibilities shape the work.

Define the information before collecting it

Start with the study question and required assessments. Agree on what each field means, the permitted values, relevant units, timing, and handling of missing information. Differences that seem small in conversation can become substantial when data from many visits or sites are combined.

For example, an incomplete assessment, an assessment not required at a visit, and an assessment that could not be performed are different situations. Staff should follow the study's conventions rather than choosing whichever option makes a form appear complete.

Identify the original record and the destination

Map where information is first captured and where it must later appear. Some information may be recorded directly in an electronic system; other information may originate in a separate source record. Do not assume every study uses the same arrangement.

Document who enters or transfers information, who reviews it, and how discrepancies are corrected. Avoid creating undocumented copies that become competing versions of the same fact. If a correction is needed, preserve the history and follow the study's authorized process.

Build data work into the visit workflow

A visit is not operationally complete merely because the participant has left. The team may still need to finalize documentation, enter required data, check receipt of external results, and route items for clinical review.

Separate these tasks from urgent medical responsibilities. A routine data-entry queue must not become the place where a time-sensitive safety concern waits for attention. The study's clinical escalation instructions take priority.

Assign coverage for absences and use a queue that makes overdue work visible. A task without an owner will not become more reliable because it appears on a dashboard.

Resolve queries with evidence

A data query asks for clarification or correction. Read the question, consult the relevant record, and respond through the appropriate workflow. Do not change a value simply to satisfy an edit check if the original information is accurate.

When a query cannot be resolved from available evidence, escalate it to the right person. Record what is known, what is uncertain, and what additional clarification is needed. Repeated queries about the same field can indicate unclear instructions or a process that needs training.

Reconcile information across systems

Laboratory results, visit records, study intervention records, and other datasets may need reconciliation. Identify the matching fields, expected timing, and handling of discrepancies. A missing record can reflect a delayed transfer, a mismatched identifier, or an activity that never occurred; these require different actions.

Use synthetic examples to test integrations and exports. Confirm behavior when information is corrected after transfer. A system that imports new records correctly may still mishandle later updates.

Review readiness before database milestones

Before a planned data review or lock milestone, agree on the remaining work: incomplete forms, unresolved queries, external data reconciliation, required investigator review, and documented exceptions. The sponsor's process determines the actual milestone requirements.

Use a short issue list with owners and dates. A percentage-complete metric is helpful only when the team understands which unresolved items matter and what evidence closes them.

FDA's electronic systems guidance and E6(R3) provide relevant principles for trustworthy records and study oversight. They do not replace an assessment of the specific system and study. The weekly management guide explains how to make data work visible alongside staffing, visits, and finances.