2016/05/22 by Elizabeth C. Lee, Michael R. Kelly, Lee, Elizabeth C. +11 · 1 citation
Biochemistry, Genetics and Molecular Biology · Medicine · Mathematics · #Vibrio bacteria research studies #Influenza Virus Research Studies #COVID-19 epidemiological studies
paper · pdf · doi:10.48550/arxiv.1605.06790
Mathematical models of cholera and waterborne disease vary widely in their\nstructures, in terms of transmission pathways, loss of immunity, and other\nfeatures. These differences may yield different predictions and parameter\nestimates from the same data. Given the increasing use of models to inform\npublic health decision-making, it is important to assess distinguishability\n(whether models can be distinguished based on fit to data) and inference\nrobustness (whether model inferences are robust to realistic variations in\nmodel structure). We examined the effects of uncertainty in model structure in\nepidemic cholera, testing a range of models based on known features of cholera\nepidemiology. We fit to simulated epidemic and long-term data, as well as data\nfrom the 2006 Angola epidemic. We evaluated model distinguishability based on\ndata fit, and whether parameter values and forecasts can accurately be inferred\nfrom incidence data. In general, all models were able to successfully fit to\nall data sets, even if misspecified. However, in the long-term data, the best\nmodel fits were achieved when the loss of immunity form matched those of the\nmodel that simulated the data. Two transmission and reporting parameters were\naccurately estimated across all models, while the remaining showed broad\nvariation across the different models and data sets. Forecasting efforts were\nnot successful early, but once the epidemic peak had been achieved, most models\nshowed similar accuracy. Our results suggest that we are unlikely to be able to\ninfer mechanistic details from epidemic case data alone, underscoring the need\nfor broader data collection. Nonetheless, with sufficient data, conclusions\nfrom forecasting and some parameter estimates were robust to variations in the\nmodel structure, and comparative modeling can help determine how variations in\nmodel structure affect conclusions drawn from models and data.\n