INTRODUCTION: Accurate well-maintained registers are a prerequisite to co-ordinated care of patients with chronic diseases. Their effectiveness in enabling improved management is dependent on the quality of the information captured. This paper provides an overview into the methodology and data quality of the electronic Practice Based Research Network. METHODS: Clinical records with no identifying information are routinely extracted from four general practices. The data are linked in the data warehouse. Data quality is assessed for completeness, correctness and consistency. Reports on data quality are given back to practices and semi-structured interviews provide information to interpret the results and discuss how data quality could be improved. FINDINGS: Data quality is mostly complete for sex and date of birth but indigenous status, smoking and weight were incomplete. There are generally high levels of correctness and internal consistency. Completeness of records in assisting the management of diabetes patients using the annual cycle of care was poor. GPs often use the progress notes to enter information during the consultation and coding diagnoses was considered onerous. DISCUSSION: The routine capture of electronic clinical health records from primary health care and health services can be used to monitor performance and improve the quality of clinical records. There is a need for accurate and comprehensive clinical records to ensure the safety and quality of clinical practice. Understanding the true reasons for poor data quality is complex. Having a community-based research network may assist in answering some of these questions. CONCLUSION: Electronic health records are increasingly being used for secondary research and evaluation, beyond the primary purpose of supporting clinical care. The data must be of sufficient quality to support these purposes.
INTRODUCTION: Accurate well-maintained registers are a prerequisite to co-ordinated care of patients with chronic diseases. Their effectiveness in enabling improved management is dependent on the quality of the information captured. This paper provides an overview into the methodology and data quality of the electronic Practice Based Research Network. METHODS: Clinical records with no identifying information are routinely extracted from four general practices. The data are linked in the data warehouse. Data quality is assessed for completeness, correctness and consistency. Reports on data quality are given back to practices and semi-structured interviews provide information to interpret the results and discuss how data quality could be improved. FINDINGS: Data quality is mostly complete for sex and date of birth but indigenous status, smoking and weight were incomplete. There are generally high levels of correctness and internal consistency. Completeness of records in assisting the management of diabetespatients using the annual cycle of care was poor. GPs often use the progress notes to enter information during the consultation and coding diagnoses was considered onerous. DISCUSSION: The routine capture of electronic clinical health records from primary health care and health services can be used to monitor performance and improve the quality of clinical records. There is a need for accurate and comprehensive clinical records to ensure the safety and quality of clinical practice. Understanding the true reasons for poor data quality is complex. Having a community-based research network may assist in answering some of these questions. CONCLUSION: Electronic health records are increasingly being used for secondary research and evaluation, beyond the primary purpose of supporting clinical care. The data must be of sufficient quality to support these purposes.