Data mining for correlations between diet and Crohn's disease activity

Abstract

Crohn's disease is a debilitating condition that affects the entire gastrointestinal tract and often requires aggressive and invasive therapies. Several studies have suggested dietary triggers for disease activity. We have created a web-based tool to allow participants to record both daily food intake and wellness (i.e., disease-specific quality of life). We seek to determine if measurable correlations exist between these events in patients with Crohn's disease. Advanced data mining techniques are employed to find such correlations and the efficacies of chosen techniques are assessed. We tested our web-based system in a pilot study involving 7 participants, and we found that traditional statistical techniques identified diet and disease activity correlations in short-term data sets.