Implementing error rate checks to improve the data quality in the Victorian Audit of Surgical Mortality.

Objective: The Victorian Audit of Surgical Mortality (VASM) is part of the Australian and New Zealand Audit of Surgical Mortality, aiming to identify factors associated with surgical mortality. A quality assurance method was utilised to rectify system issues through error rate checks of VASM data points. This method ensured data collected and entered in the audit database were accurate.

Method: Clinical data was collected using case record forms (CRFs), entered from paper forms by VASM staff or by the surgeon into an online interface. Closed audited cases from 1 January 2012 to 31 December 2016 were analysed. From 5528 closed cases, 485 (8.8%) were reviewed, containing a total of 1117 CRFs and 169,789 data fields. Each CRF entered was visually inspected in the database against the source document by an independent audit staff member. The error rate for each period was calculated and considered acceptable below 10 per 10,000 fields.

Results: Text errors made up the majority of data entry errors regardless of how the case was submitted.

Conclusion: Application of error rate checks is beneficial to maintain good clinical data. This activity improved and streamlined the data collection process to reduce errors associated with data entry. Once the entry system was stabilised, a reduction in error rate was observed showing potential for further improvements. We acknowledge that errors cannot be entirely eliminated and it is unrealistic. By investigating the rigour of the data management processes based on research guidelines, the findings can contribute to improve quality of clinical audits.