Using read codes to identify patients with irritable bowel syndrome in general practice: a database study

Background: Estimates of the prevalence of irritable bowel syndrome (IBS) vary widely, and a large proportion of patients report having consulted their general practitioner (GP). In patients with new onset gastrointestinal symptoms in primary care it might be possible to predict those at risk of persistent symptoms. However, one of the difficulties is identifying patients within primary care. GPs use a variety of Read Codes to describe patients presenting with IBS. Furthermore, in a qualitative study, exploring GPs' attitudes and approaches to defining patients with IBS, GPs appeared reluctant to add the IBS Read Code to the patient record until more serious conditions were ruled out. Consequently, symptom codes such as 'abdominal pain', 'diarrhoea' or 'constipation' are used. The aim of the current study was to investigate the prevalence of recorded consultations for IBS and to explore the symptom profile of patients with IBS using data from the Salford Integrated Record (SIR). Methods: This was a database study using the SIR, a local patient sharing record system integrating primary, community and secondary care information. Records were obtained for a cohort of patients with gastrointestinal disorders from January 2002 to December 2011. Prevalence rates, symptom recording, medication prescribing and referral patterns were compared for three patient groups (IBS, abdominal pain (AP) and Inflammatory Bowel Disease (IBD)). Results: The prevalence of IBS (age standardised rate: 616 per year per 100,000 population) was much lower than expected compared with that reported in the literature. The majority of patients (69%) had no gastrointestinal symptoms recorded in the year prior to their IBS. However a proportion of these (22%) were likely to have been prescribed NICE guideline recommended medications for IBS in that year. The findings for AP and IBD were similar. Conclusions: Using Read Codes to identify patients with IBS may lead to a large underestimate of the community prevalence. The IBS diagnostic Read Code was rarely applied in practice. There are similarities with many other medically unexplained symptoms which are typically difficult to diagnose in clinical practice.