Identifying Interaction Types in Primary Care Data within the SAIL Databank

Abstract

Introduction: Understanding the interactions occurring in GP surgeries is crucial for understanding how General Practice functions, informing policy decisions and guiding healthcare provisioning. This work aims to identify the type of interaction that a particular patient is having at their registered general practice on a specific day.

Methods: Relevant Read codes were identified by keyword search in the Read code glossary. These were then categorised into 2 distinct categories of interaction. These types of interaction were further tagged with the associated Health Care practitioner. The Welsh Longitudinal General Practice Dataset (WLGP Dataset) from the SAIL databank was then cleaned to identify unique interactions, extracting annual counts for each interaction type.

Results: The analysis revealed that the number of face-to-face GP interactions has been relatively stable since 2002. The number of ‘non-face-to-face’ interactions has greatly increased since 2002, overtaking face-to-face interactions since 2016. A notable decline in face-to-face interactions was observed during the COVID-19 pandemic.

Conclusion: This work identified GP interaction types and associated healthcare practitioners, highlighting important trends. Future work will involve refining Read codes with healthcare professionals. In addition, further analysis is needed to categorise a portion of the cleaned WLGP data that was not labelled with an interaction type.