Evaluating the accuracy of data entry in a regional colorectal cancer database: implications for national audit.

OBJECTIVE

To determine the accuracy and completeness of data entry in a regional colorectal cancer database.

METHOD

Records were obtained for 50 patients selected at random from the database. A sample of 18 key data points (preoperative and operative) were selected from the Association of Coloproctology of Great Britain and Ireland colorectal cancer minimum dataset. Data were extracted from the case notes by a clinician and compared with database entries.

RESULTS

Overall 83% of data elements were present in the patient records. There was agreement between patient records and the database for 61% of data points overall, and for 70% of preoperative data and 47% of operative data.

CONCLUSION

The regional colorectal cancer database is inaccurate, particularly regarding operative data. Insufficient information in patient records accounts for a proportion of missing data, and lack of agreement between records and the database may be due to misinterpretation of records by nonclinical staff. The accuracy of regional data may be improved by weekly review of database entries and/or entry of clinical data by a clinician.