Regional variation in Helicobacter pylori resistance patterns is a significant contributing factor for the ineffectiveness of traditional treatments. To improve treatment outcomes, we sought to create an individualized, susceptibility-driven therapeutic approach among our patient population, which is one of the poorest in the nation. It is medically underserved, minority-predominant and has high incidence of H pylori infection. We compiled various factors involved in the antibiotic resistance of H pylori from literature. We then created a predictive model to customize therapies based on analyzed data from 2,014 H pylori patients with respect to several of these factors. The predictions of the model were further tested with analysis of patient stool samples. A clear pattern of H pylori prevalence and antibiotic resistance was observed in our patients. We observed that majority of H pylori patients were women (62%) and over the age of 40years (80%). 30% and 36% of the H pylori patients were African American and Hispanic, respectively. A median household income of less than $54,000, past H pylori infection, previous use of certain antibiotics for any infection decreased the chance of eradication. Results of the stool testing were consistent with model predictions (90% accuracy). This model demonstrates the predictive accuracy of H pylori infection and antibiotic resistance based on patient attributes and previous treatment history. It will be useful to formulate customized treatments with predicted outcomes to minimize failures. Our community attributes may contribute toward broad applicability of model for other similar communities.
Read full abstract