Keywords: coding, electronic medical records, surveillance, general practitioners, classification of diseases

Background:

The Infection Barometer surveillance system for general practices in Belgium, successor to the Covid-19 Barometer, automatically extracts infection-related data from Electronic Medical Records (EMR) on a daily basis and in real time. To optimise its functioning, General Practitioners (GPs) must encode diagnoses adequately to ensure the quality of extracted data. Currently, we observe that coding is not always systematic and accurate, though causes behind coding discrepancy remain relatively unexplored in the literature. The aim of this project is therefore to map the factors influencing GPs’ coding behaviour and understand how they can be levers to increase data quality.

Research questions:

What intrinsic and extrinsic factors do GPs believe influence the coding of diagnoses in the EMR?

Method:

In the first phase, a questionnaire was sent to GPs of the Sentinel GP Network and the Covid-19 Barometer to explore their coding behaviour. In the second phase, a questionnaire will be sent to GPs' software vendors to gain a better understanding of how coding is integrated into their system. In the third phase, semi-structured interview guides will be drafted based on the results of phase 1 and 2. Interviews will be conducted to gain in-depth understanding of GPs’ coding behaviour. In the fourth phase, focus groups will be set up with GPs, software developers, medical classification specialists and other stakeholders to cross-check the expectations, needs and recommendations.

Results:

The questionnaire was sent to GPs in May-June 2024. The software questionnaire will be sent in September 2025. The interviews and focus groups are planned for Spring 2026.

Conclusions:

This study will explore the intrinsic and extrinsic factors influencing GPs' coding and will enable relevant levers for action to improve and optimize the data extraction system in the GP Infection Barometer surveillance system.

Points for discussion: