Data Management in Clinical Trials: Best Practices in 2026

From protocol to database lock: a discipline in its own right

Clinical data management is often seen as a secondary, technical step. In reality, it forms the backbone of any clinical trial: without clean, complete and traceable data, neither the statistical analysis nor the regulatory submission can be completed on time. In 2026, the EMA's and the FDA's requirements regarding data quality have never been so precise.

Here are the key stages of rigorous data management, from study start-up to database lock.

1. Drawing up a solid Data Management Plan

The Data Management Plan (DMP) is the reference document that describes all the data management processes for a given study. It specifies the structure of the CDMS, the validation rules, the monitoring strategy, the medical coding procedures and the database lock criteria.

A well-drafted DMP prepared in advance avoids ambiguities during the study and makes audits easier. It should be finalised before the first site is opened and updated with every substantial protocol amendment.

2. CDMS configuration and validation rules

CDMS configuration should be driven by the protocol and the statistical analysis plan, not the other way round. Each data-entry field should correspond to a defined analytical variable. The validation rules — range checks, inter-visit consistency checks and mandatory missing-data checks — should be programmed before the first patient.

A well-configured CDMS reduces the volume of queries generated during the study by 30 to 50%. At Axiodis, we configure Exagis from the protocol specifications and the SAP in under five days for studies of standard size.

3. MedDRA and WHO-Drug coding

The coding of free-text medical terms is a critical step for data comparability. MedDRA applies to adverse events, medical history and diagnoses. WHO-Drug covers concomitant treatments.

In 2026, the EMA recommends using the most recent MedDRA version available at the database lock date. At Axiodis, coding is carried out by certified coders, with a double review of ambiguous terms and the use of a shared local dictionary to ensure intra-study consistency.

4. Centralised monitoring and query management

Centralised monitoring, governed by the ICH E6(R3) regulation, makes it possible to identify risk signals ahead of on-site visits. The progress dashboards — completion rates, query resolution times and statistical outliers — should be reviewed continuously by the lead Data Manager.

Queries should be worded precisely, documented and resolved within the timeframes defined by the DMP. A backlog of unresolved queries as the database lock approaches is the main cause of delay on studies.

5. CDISC compliance and database lock

The CDISC standards — CDASH for collection and SDTM for submission — are now mandatory for any FDA submission and strongly recommended by the EMA. The preparation of SDTM datasets should be anticipated from the CDMS configuration stage onwards, not treated as a last-minute conversion.

The database lock takes place once all queries have been resolved, all coding validated and all final quality controls passed. It is formalised by an electronic signature in the system and triggers the freezing of the database for the biostatistics teams.

Structured, well-equipped and documented data management is not a luxury: it is what ensures your study delivers usable data, on time and by the book.