Clinical Trajectory-Based Modeling of C. difficile to Link Individual Risk Profiles to Healthcare Burden

PROJECT SUMMARY C. difficile infection (CDI) remains one of the most burdensome healthcare-associated infections, causing sub- stantial morbidity, mortality, and healthcare costs. Most analytic frameworks treat CDI as a single event that consists only of symptomatic infection. This fails to capture the full clinical progression including asymptomatic colonization, heterogeneity in severity, persistent carriage after treatment, and relapse. By oversimplifying CDI’s multistage natural history, all-or-none models can bias estimates of factors associated with disease progression, and obscure opportunities to identify and rigorously evaluate targeted infection-control strategies. To address these gaps, we propose a Hidden Markov modeling framework that explicitly integrates latent and observable CDI states using patient-level electronic health data. Our approach combines latent-state inference with regression-based transition rates. This enables modeling of individual clinical trajectories and response to control interventions. • Aim 1: Quantify patient-level drivers of CDI progression by developing and parameterizing a multistage Hidden Markov model to determine how clinical characteristics such as antibiotic exposures influence transitions across disease stages. • Aim 2: Evaluate how interventions (e.g., antibiotic stewardship, prophylaxis, and contact precautions) affect individual incidence, severity, relapse, persistent colonization. • Aim 3: Project the population-level impact of interventions by incorporating stages of CDI progression into a stochastic, covariate-dependent compartmental model. Indivdual- and population-level effect will be compared by simulating clinical trajectories, transmission dynamics and system-wide burden under different prevention strategies. Expected outcome and impact: This project will improve our understanding of the natural history of the disease, CDI risk stratification, and intervention effectiveness. This will enable data-driven clinical and policy decisions, supporting more precise antimicrobial stewardship, screening, prophylaxis, and infection-control measures to reduce CDI incidence, recurrence, and healthcare costs.