Advancing and Addressing Uncertainties in Scenario-Specific Healthcare QMRAs with Multidisciplinary Approaches

Abstract

The quantitative microbial risk assessment (QMRA) framework continues to develop to address exposure routes beyond its original application in water quality contexts, especially in its use to address healthcare-associated infections. Exposure models used within these QMRAs must be advanced to incorporate multiple exposure routes and to account for not only the magnitude of microbial spread but also spatial patterns of this spread. In this dissertation, an agent-based model integrated with an exposure transfer model was developed to evaluate the contribution of wheelchairs to spatial contamination spread in a healthcare facility and exposures to subsequent patients riding the wheelchair. This integrated framework provided insights into emergent patterns of exposures for subsequent riders on contaminated wheelchairs and spread throughout the facility. Main findings include that disinfection of wheelchairs in between patients may protect future riders under low contamination conditions, and that the frequency of traveled paths is related to heterogeneity of fomite contamination throughout a facility. In analyzing emergent behaviors, the number of wheelchairs had a positive relationship with number of contaminated patches over a specific concentration threshold. Cleaning wheelchairs in between patients weakened this relationship. While the integration of multiple modeling approaches is a future direction of QMRA, uncertainty in mechanisms of microbial spread still need to be addressed in order to improve accuracy in integrated model frameworks. For example, in this work, the exposure transfer model used in this integrated model includes the assumption that transfer occurs according to a concentration gradient. This was evaluated experimentally with bacteriophage transfer efficiency studies and supported the hypothesis that transfer efficiency varies by a ratio of the concentrations on the fingertip and surface prior to contact. These experimental data were then used in an Approximate Bayesian Computation (ABC) analysis to compare two microbial transfer models in their ability to explain the experimental data and offer insights regarding swabbing efficiencies and transfer efficiencies that are challenging to measure experimentally. This analysis demonstrated that swabbing efficiencies and direction-specific transfer efficiencies “balance” exchanges between the fingertip and surface in predicting after-contact concentrations on the fingertip or surface. Future research involves furthering the integration of exposure model frameworks to account for complex transmission systems, especially those that incorporate spatial human behavior. This also necessitates further experimental and mathematical evaluation of these models to advance not only our ability to model complex built environment systems but also to improve our accuracy in estimated exposure and health outcomes.