17.1 Data Management Processes

Clinical Data Management (CDM) is a critical phase in clinical research, ensuring that data collected during a clinical trial is accurate, secure, reliable, and ready for statistical analysis. The primary goal of CDM is to gather high-quality data that supports the study's conclusions and meets regulatory standards. This process involves the entire lifecycle of data, from the initial design of collection instruments to the final locking of the database.

The Role of Clinical Data Management

The integrity of a clinical trial relies entirely on the quality of its data. If data is flawed, incomplete, or inconsistent, the study's conclusions about a drug's safety and efficacy will be invalid. CDM processes are designed to mitigate these risks by establishing robust frameworks for data collection, entry, review, and cleaning.

Key Components of Data Management

Electronic Data Capture (EDC) Systems

In modern clinical trials, EDC systems are the standard for data collection. These platforms offer significant advantages over paper-based systems, including real-time data visibility, automated edit checks, and streamlined query management.

Common EDC systems used in the industry include Medidata Rave, Oracle Clinical, and IBM Clinical Development. While interfaces differ, all EDC systems share common regulatory requirements, particularly adherence to 21 CFR Part 11, which governs electronic records and electronic signatures.

Advantages of EDC Systems

• Real-Time Access: Sponsors and monitors can view data immediately after entry, allowing for faster decision-making and safety monitoring.
• Automated Edit Checks: The system can automatically flag illogical or out-of-range data at the time of entry (e.g., a resting heart rate of 250 bpm).
• Audit Trails: Every change made to the data is recorded, detailing who made the change, when it was made, and the reason for the change.
• Efficient Query Management: Queries can be issued, tracked, and resolved within the system, eliminating the need for separate paper logs.

The Data Entry Process

For Clinical Research Coordinators (CRCs), data entry is a significant daily responsibility. Data is typically transcribed from the source document (e.g., electronic medical record, lab report) into the EDC system.

Best Practices for Data Entry

To maintain data integrity and reduce the query burden, CRCs should follow these best practices:

1. Timeliness: Enter data as soon as possible after the subject visit, typically within 3 to 5 days, as specified by the sponsor's guidelines.
2. Accuracy: Ensure that the data entered exactly matches the source document. Do not round numbers, estimate values, or correct obvious medical errors in the EDC without first amending the source document.
3. Completeness: Complete all required fields. If data is missing or unknown, use sponsor-approved abbreviations (e.g., "ND" for Not Done, "UNK" for Unknown) rather than leaving the field blank.

Source Data Verification (SDV)

Source Data Verification is the process by which a Clinical Research Associate (CRA) or monitor compares the data entered into the CRF/EDC with the original source documents. SDV ensures that the data is accurate, complete, and verifiable.

The SDV Process

During a monitoring visit, the CRA will:

• Review the source documents (medical records, lab results, subject diaries).
• Compare the source data against the EDC entries.
• Verify that protocol-specific procedures were conducted within the required timeframes.
• Check that all adverse events and concomitant medications have been accurately recorded.

If the CRA identifies a discrepancy during SDV, they will issue a query in the EDC system for the site to resolve.

Real-World Example: Oncology Trial Data Flow

Consider a Phase III oncology trial evaluating a novel targeted therapy. The data flow typically follows this sequence:

1. Patient Visit: A subject attends Day 1 Cycle 1. Vital signs, blood work, and an ECG are completed. The nurse records this information in the hospital's Electronic Medical Record (EMR).
2. Data Entry: The CRC reviews the EMR (the source document) and transcribes the vital signs, lab results, and ECG findings into the trial's EDC system within 48 hours.
3. Automated Check: The EDC system's edit checks run in the background. If a lab value is out of the predefined range, an automated query fires immediately, prompting the CRC to verify the entry.
4. Monitoring: Two weeks later, the CRA logs into the EDC remotely (Risk-Based Monitoring) or visits the site to perform SDV, comparing the EDC data against the hospital EMR.
5. Data Cleaning: The Clinical Data Manager at the sponsor reviews the aggregated data for trends, missing pages, or logical inconsistencies across visits, issuing manual queries as needed.
6. Database Lock: Once all subjects have completed the trial, all queries are resolved, and all data is verified, the database is locked, and the data is handed over to the biostatisticians for analysis.

The Database Lock

The culmination of the data management process is the database lock. This is a critical milestone indicating that the data is considered final, clean, and ready for statistical analysis.

Before a database can be locked:

• All data must be entered into the EDC.
• All queries must be resolved and closed.
• All CRFs must be signed by the Principal Investigator (e-signature in the EDC).
• All external data (e.g., central lab results, ECG readings) must be reconciled with the EDC data.
• A final quality control review must be completed.

Once locked, write access to the database is revoked. Any changes required after a database lock are extremely rare and require a formal "database unlock" process, accompanied by rigorous documentation and justification, as it can significantly impact the statistical analysis and regulatory submission.