Related Experiment Videos
Is double data entry necessary? The CHART trials. CHART Steering Committee. Continuous, Hyperfractionated,
D Gibson1, A J Harvey, V Everett
1Medical Research Council, Cancer Trials Office, Cambridge, England.
Controlled Clinical Trials
|December 1, 1994
Summary
Double data entry in clinical trials identified fewer errors than initially entered data. This study suggests double data entry is unnecessary for the CHART trials due to low error rates.
Area of Science:
- Clinical research methodology
- Data management in healthcare
Background:
- The necessity of double data entry in clinical trials is debated.
- The effort involved in double data entry versus the errors detected is a key concern.
Purpose of the Study:
- To evaluate the frequency and types of errors in clinical trial data entry.
- To determine if double data entry is justified for the CHART trials.
Main Methods:
- A sample of 44 patients' data from the CHART trials was used.
- Data was entered twice by different data managers.
- The two entries were compared to identify discrepancies and error rates.
Main Results:
- The combined error rate for double data entry was 14 per 10,000 fields.
- The initial data entry error rate was 15 per 10,000 fields.
- The vital/important error rate was 2.5 per 10,000 fields.
Conclusions:
- The error rate for initial data entry was not significantly higher than for double data entry.
- The low incidence of vital/important errors does not justify the resources for double data entry.
- Double data entry is not being implemented for the CHART trials based on these findings.