Process

Objective

To ensure that all data items critical to primary and secondary uses are subject to data quality validation and results outputted through reports (sourced both internally and externally) that are made available to those responsible for taking corrective action.

Benefit

Having a data quality reporting process in place is essential for providing active review of data quality issues and timely resolution of the source data.  This is particularly relevant where validation cannot be applied at the point of entry thereby requiring a secondary level of validation to assure the completeness and validity of the wider clinical record.  Frequent reporting outlines performance in relation to explicit quality expectations and provides assurance of adherence to internal and external data quality standards. It provides recognition of all primary and secondary uses for which the data is required and offers transparency in relation to quality levels which may inform how the data is used. Data quality reporting also exposes process improvement opportunities.

Best Practice

Data quality reporting should be integrated into the organisation’s wider information management processes rather than being undertaken in isolation, as part of the national reporting cycle. This reporting consistency when reinforced with trends comparison against previous or subsequent reporting periods, supports reliability and continuity in information flows. Reporting control requires both human and technical resources, backed by training and enough time to produce high quality data verification. Investment in incorporating data quality checks into all information systems is essential.

The above should be supported by regular publication of statistical information detailing the quality and quantity of an organisation's health data. This should be derived from all available data quality reports, including those related to data quality incidents (see Part 2 of the framework), to facilitate analysis of the data’s strengths and weaknesses. This information should also be used to support selection and production of internal metrics (see Part 2 of the framework).