Evaluating a transfer gradient assumption in a fomite-mediated microbial transmission model using an experimental and Bayesian approach

• Published in: Journal of the Royal Society interface 2020-06, Vol.17 (167), p.20200121
• Main Authors: Wilson, Amanda M., King, Marco-Felipe, López-García, Martín, Weir, Mark H., Sexton, Jonathan D., Canales, Robert A., Kostov, Georgiana E., Julian, Timothy R., Noakes, Catherine J., Reynolds, Kelly A.
• Format: Article
• Language: English
• Subjects: Life Sciences–Mathematics interface
• ISSN: 1742-5689; 1742-5662
• DOI: 10.1098/rsif.2020.0121

Current microbial exposure models assume that microbial exchange follows a concentration gradient during hand-to-surface contacts. Our objectives were to evaluate this assumption using transfer efficiency experiments and to evaluate a model's ability to explain concentration changes using approximate Bayesian computation (ABC) on these experimental data. Experiments were conducted with two phages (MS2, Φ X174) simultaneously to study bidirectional transfer. Concentrations on the fingertip and surface were quantified before and after fingertip-to-surface contacts. Prior distributions for surface and fingertip swabbing efficiencies and transfer efficiency were used to estimate concentrations on the fingertip and surface post contact. To inform posterior distributions, Euclidean distances were calculated for predicted detectable concentrations (log 10 PFU cm −2 ) on the fingertip and surface post contact in comparison with experimental values. To demonstrate the usefulness of posterior distributions in calibrated model applications, posterior transfer efficiencies were used to estimate rotavirus infection risks for a fingertip-to-surface and subsequent fingertip-to-mouth contact. Experimental findings supported the transfer gradient assumption. Through ABC, the model explained concentration changes more consistently when concentrations on the fingertip and surface were similar. Future studies evaluating microbial transfer should consider accounting for differing fingertip-to-surface and surface-to-fingertip transfer efficiencies and extend this work for other microbial types.