184 Data Quality In Gp Electronic Health Records And How To Improve It.

Mark Porcheret
Aims To investigate the impact of a package of repeated assessments, feedback and training on the quality of coded clinical data in a GP research network in North Staffordshire.
The network was established to collect coded consultation data for epidemiological research and health planning. The study presents the work carried out in seven of the practices.
Method Baseline assessments were made of:
1. Percentage of consultations that were coded,
2. Percentage of patients on a specific drug [e.g. tamoxifen] that had the relevant morbidity [e.g. breast cancer] coded,
3. Annual period prevalence of patients consulting with a number of conditions. A comparison was made to rates in the 4th National Study of Morbidity Statistics from General Practice [MSGP4].
The assessments were presented to the practices and training arranged.
The cycle of assessment, feedback and training continued over three years.
Results The first two measures showed variations between practices at baseline and on repeat assessments but over time all practices improved or maintained their level of coding.
The period prevalence was also variable but over time rates increased to levels comparable to, or above, MSGP4 rates.
The feedback was well received by the practices and staff were happy to undergo training.
Even highly motivated practices have variable levels of consultation coding and completeness of disease registers but can be helped to improve. The three elements of the package are essential and the assessment team needs to have the trust of the practices.
184 Data Quality in GP Electronic Health Records and How to Improve it.