Evaluating hypothetical interventions effects on hospital-acquired infection outcomes with stacked probability...

Hospital-acquired infections (HAIs), contribute to increased morbidity, prolonged hospital stays, and higher healthcare costs. Evaluating intervention effects in dynamic clinical settings requires advanced modeling techniques. This study presents HAISim and StaViC, two interactive R Shiny apps designed to support decision-making in Infection Prevention and Control.

A six-state extended illness-death multistate model was built with a time-constant transition hazard assumption. The multistate model maps patient trajectories involving healthcare-associated infections as an intermediate event, and mortality and discharge as absorbing events in order to describe the time-dependent dynamics within hospitals. Using two settings, we simulated the implementation of hypothetical treatments by modifying hazard rates: Setting 1 (improved treatment intervention only) and Setting 2 (combined enhanced treatment and infection prevention). These were used to create the interactive and user-friendly R Shiny Apps HAISim (HAIs Interventions Simulator) and StaViC (Stacked probAbility Visualization & Comparison). The Shiny Apps use inputs from literature or user data, such as transition-specific hazard rates and intervention-related parameters.

HAISim models the effects of hypothetically improved treatment and infection prevention on outcomes such as the number of lives saved and the number of patient days decreased by simulating a hypothetical scenario based on actual clinical data. StaViC makes it possible to compare potential interventions and their impacts before and after implementation by visualizing the stacked probabilities of patients across various health conditions.

These tools bridge methodological rigor and practical implementation, offering hospitals a flexible framework to prioritize cost-effective IPC strategies.