Study management and data quality
- Concept
What makes data trustworthy: a procedures manual, clear forms, trained staff, monitoring and quality control.
What it is
Study management turns the protocol into daily tasks: a manual of procedures, the case report form (CRF), staff training and certification, a database with validation checks and an audit trail. A pilot study tests the whole circuit before the start. During the study, monitoring and quality control (double data entry, source data verification, review of outliers) catch errors while they can still be corrected.
Key facts
- Error rate = fields with errors / fields checked.
- Source data verification: the CRF is compared with the clinical record or another original document.
Worked example
- Data: 500 case report forms with 40 fields each: 500 × 40 = 20 000 fields.
- Independent double data entry: the two versions disagree in 60 fields: 60 / 20 000 = 0.3 %.
- Each discrepancy is resolved against the source document (the clinical record) and the change is kept in the audit trail.
- Monitoring: 10 % of forms are checked against source, 50 × 40 = 2 000 fields; 6 errors are found: 6 / 2 000 = 0.3 %.
- The manual set a 0.5 % error limit; 0.3 % is below it, so not every form needs checking.
- Reading: quality is planned before the start (manual, training, pilot) and measured during the study, not at the end.
Management tools and what they are for
| Tool | What it is for |
|---|---|
| Manual of procedures | Everyone does each step the same way |
| Case report form (CRF) | Collecting only what is needed, in coded form |
| Training and certification | Measurements comparable between staff |
| Database with validation checks | Blocking impossible values at entry |
| Audit trail | Knowing who changed which value and when |
| Monitoring and source verification | Catching errors and deviations in time |
| Pilot study | Testing the whole circuit on a small scale |
Markers
- CRF
- GCP
- SDV
In practice
- Good clinical practice requires that every CRF value can be traced back to its source document.
- In multi-site studies, regular meetings and recruitment reports quickly reveal the site that is falling behind.
Common pitfalls
- Collecting dozens of variables ‘just in case’ that nobody will analyse.
- Correcting values in the database without leaving a record.
- Checking quality only after the study has ended.
Test yourself
Think of the answer, then tap the question to see it.
What is an audit trail?
The automatic record of who created or changed each value, when and why.
3 000 fields are checked and 15 errors are found. What is the error rate?
15 / 3 000 = 0.5 %.
What is a pilot study for?
To test recruitment, instruments, timing and the database on a small scale before the main study.
What is risk-based monitoring?
Focusing review on the processes and data that most affect participant safety and the primary outcome.
Elsewhere in DurgiPedia
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: E6(R3) Good Clinical Practice (GCP) (U.S. Food and Drug Administration, 2025); Collecting Data (Centers for Disease Control and Prevention (The CDC Field Epidemiology Manual), 2024).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.