Dr. David Morton, Chair of the Industrial Engineering and Management Services Department at Northwestern University, presented on November 16, 2017 on the topic of Pandemic Influenza Preparedness. Professor Morton delved into incidents of influenza pandemics in the last 100 years and the different ways these pandemics are addressed through both vaccines and mechanical ventilators–critical life saving devices for patients in respiratory failure. He spoke about his research on a project, funded by the CDC, involving the Texas Department of State Health Services (DSHS) to prepare the organization for future influenza pandemics. The goals of the project included optimizing on the central and regional stockpiles of ventilators, keeping in mind peak-week demand, spatial allocation of antivirals (which took into consideration both under-insured populations and hard to reach locations), and spatial allocation of multiple vaccines with differing suitability for each targeted population.
The focus of this project stemmed from resources through the Strategic National Stockpile (SNS), the nation’s largest supplier of life-saving pharmaceuticals and medical supplies. Factors taken into consideration included predictors of demand such as humidity, school closures, and influenza-like illness reporting systems. Also, for the stockpiling component of mechanical ventilators, factors such as the number of hospitalizations, number of people in the ICU, and how many mechanical ventilators were needed were researched. When looking at the 2009 H1N1 virus pandemic, Dr. Morton found that despite the fact that it usually makes sense to have a central stockpile, there was little benefit for DSHS to have a large central stockpile of mechanical ventilators. This is due to high spatial correlation (i.e., the demand for ventilators was high in Houston resulting in the demand being high in West Texas) and the wastage parameter (fourth-fifths of ventilators went to good use).
Dr. Morton then delved into the component focused on vaccines targeted for the 2009 H1NI virus with DSHS in collaboration with researcher Lauren Meyers. The goal was to find tools for training, planning, and real-time deployment using the UT Austin server. Dr. Morton discussed the focus on rural county areas and priority targeted groups such as pregnant women and healthcare personnel. The project looked at system performance from multiple perspectives and attempted to find a system that was proportionally fair and policies that would be relatively simple.
Finally, Dr. Morton delved into the separate issue of cell phone viruses which are capable of being spread through internet, bluetooth, and MMS. He discussed the history of these viruses tracing back from the first virus, Cabir. He discussed the challenges to virus detection and how the monitoring of cellphones is bandwidth limited. The issue being addressed was finding a way to maximize the probability of virus detection by a given time and minimize the expected detection time through the installation of honeypots, monitoring handsets which detect viruses. The two models compared were the Monotone Submodular Functions model, which resulted in an inaccurate evaluation, and the Greedy Heuristic model, to speed up computations. Overall the Greedy approximation was best able to achieve the project’s goals despite the fact that it wasn’t clear why this model worked better and there is still further research needed to look at virus behavior models and time dynamics of the network.
