Introduction Irritable bowel syndrome (IBS) is often a diagnosis of exclusion, with poor diagnosis coding in primary care. This makes identification of eligible research participants challenging.

We present the methodology development of a multi-centre, observational, retrospective research study ongoing in primary care, designed to overcome the challenges of IBS patient identification.

Methods Study feasibility was conducted by pH Associates (research consultancy; study coordinators) for Almirall UK Ltd (Sponsor) using medical opinion, clinical coding searches and NIHR Clinical Research Network expertise. FARSITE, a software tool for identification of research participants in primary care developed by the Greater Manchester Comprehensive Local Research Network and North West e-Health, was used to screen anonymised primary care records for potential eligible patients. Search criteria: patients aged 18–60; combination READ code symptoms indicative of IBS and prescription of IBS drugs 01/01/2009–31/12/2011. GP practices with eligible patients were invited to participate, with GPs reviewing clinical records of the FARSITE-generated list of patients to apply full eligibility criteria for final patient selection.

Inclusion criteria: medical diagnosis of IBS or meeting ROME III criteria; provision of consent. Exclusion Criteria: diagnosis excluding IBS; IBS symptoms secondary to other condition; IBS medications for non-GI symptoms. The study is ongoing in 8 GP practices in Salford and Greater Manchester (Ethical approval 13/LO/0692).

Results FARSITE feasibility search using READ code for IBS identified 50 (0.02%) patients. Combining READ codes with symptom and prescriptions criteria selected 4714 (1.9%) From these, 3 GP practices each screened 10 random patient records for eligibility and 12/30 (40%) were found eligible. Eligibility READ codes were revised following feasibility.

Following study approvals, FARSITE identified 1089 potential eligible patients at the 8 participating practices, of which 297 (27.3%) were eligible and approached for consent for participation. Main reasons for non-eligibility were symptom characteristics not meeting ROME III criteria or not confirmed as IBS by medical opinion.

Conclusion Identification of patients with IBS using READ code is sub-optimal in primary care. A combination search of READ codes with symptom and prescription data via FARSITE has enabled potential participants to be identified with a reasonable screening failure rate. FARSITE is a valuable research tool aiding study feasibility by reducing the need for manual patient identification.

Disclosure of Interest I. Caldwell: None Declared, J. Collins: None Declared, M. Rance Employee of: Almirall UK Limited, R. Dew Conflict with: Commissioned by Almirall UK to provide research design, conduct analysis and scientific editorial services.