18th March 2014, 7:15pm
Testing Mathematical Antibiotic Resistance Mitigation Strategies in Clinical Trials: are models just too coarse to capture truly complex, evolving, living systems?, Professor Robert Beardmore
Abstract
The world around us is complex: it has many interdependent components. Mathematics is the key to understanding, predicting, controlling and designing complex systems. I will describe how mathematics gives power to the popular concepts of emergence, tipping points and nudging. I will lay out a manifesto for how the IMA will contribute to the future of the nation.
It has been suggested (several times) that mathematical models predict that the optimal way of mitigating antibiotic resistance in hospital settings is to prescribe drugs to patients as if “at random”. The idea, that maximising the heterogeneity of prescription is the optimal behavioural strategy in the clinic, has been tested in several trials and the jury is still out. Moreover, analyses of clinical trial data that have examined millions of patient days of treatment have found no correlation between the heterogeneity of prescription and the drug resistance patterns that result. But did those mathematical models really say this?
The purpose of this talk is to re-examine the models and their analyses to ask if the conclusions that were drawn really are found robustly in models of antibiotic use in the clinic. Or, as I contend, does the lack of concrete data to rule out unreliable models mean that they can be judiciously selected (including parameter sets within models) to support almost any view on antibiotic use? So when the clinician asks, “Which is the optimal hospital policy, antibiotic cycling or maximal antibiotic heterogeneity?” I believe the mathematical modeller should say something like, “At this level of generality, there is no answer to that question”.
We end with a discussion of antibiotic-related phenomena that mathematical models can help understand, including genomic duplications of bacterial efflux pumps, two-drug collateral sensitivities and the evolutionary stability of two-antibiotic interactions.
No charge is made to attend meetings and non-members are welcome.
