Routine assessment of gastrointestinal symptom using a validated questionnaire in the clinical setting to assess the probability of organic or functional gastrointestinal diseases.

Abstract

Patients presenting with gastrointestinal symptoms can be challenging in terms of determining etiology and management strategies. Identifying likely organic pathology is important since it can be treated and may result in further, long-term harm to the patient if not treated. Currently, organic pathology is often identified via invasive procedures such as endoscopy or referral to a medical imaging service. We report on an approach that offers a first step at identifying patients with an organic gastrointestinal disease based on the SAGIS, a validated symptom questionnaire. 8,922 patients referred to a tertiary care hospital were classified as having either functional gastrointestinal disease or an organic gastrointestinal disease. A model was developed to distinguish organic from functional symptoms on one random split half of the sample and validated on the other half. The incremental benefit of including psychological conditions and extra-gastrointestinal conditions was also evaluated. Functional gastrointestinal patients scored higher on average than organic patients on all dimensions of the SAGIS and reported higher rates of psychological and extra-gastrointestinal conditions. All five dimensions of the SAGIS provided statistically independent discrimination of organic from functional diagnoses with good overall discrimination (AUC=0.75). However, there was no noticeable incremental benefit of adding either psychological or extra-gastrointestinal conditions. Model performance was highly reproducible. The proposed algorithm for identifying likely organic gastrointestinal disease applied to symptoms as recorded in the SAGIS questionnaire provides a useful tool for the clinician in deciding what or if further diagnostic testing is required.