The optimum granularity for coding diagnostic data in primary care: report of a workshop of the EFMI Primary Care Informatics Working Group at MIE 2005

Abstract

Introduction: The EFMI Primary Care Informatics Working Group held a workshop to explore interventions used across Europe to improve the data quality in primary care computerised medical records.

Method: A plenary session reviewed the UK literature about improving data quality and then the session split into three small groups. Fifteen delegates from nine countries contributed to the workshop. These groups reported back at the end of the session.

Results: The groups defined what they meant by data quality. The principal requirement was that data must be 'fit for purpose'. The participants felt this was particularly important for diagnostic data, while recognising that the purpose might not be known at the point of data recording. They also described the barriers to recording structured and coded data. The most important were an inappropriate interface with the coding system and inappropriate granularity of codes. There was a wide range of suggestions as to how to overcome these barriers, including providing feedback, links to expert systems, education and training, use of the data for care elsewhere in the health system and mandation of electronic data recording.

Conclusions: The workshop developed a new characteristic of data quality: 'fit for purpose'. This is different from definitions that focus on completeness, accuracy, currency, or its positive predictive value and sensitivity. The group also highlighted the importance of data quality of diagnoses, as these data are important throughout the health system as well as acting as a prompt for other interventions within the individual consultation. More research is needed into appropriate levels of granularity for diagnostic recording in primary care.