Mathematical Models to Predict Survivability from Cardiac Arrest
Statistical modeling has proven to be a powerful tool in biomedicine, transforming diagnostics and prognostics in many fields. However, to date it has not been applied sufficiently in the important area of critical care and resuscitation, despite its potential to have immediate impact on life-saving treatments.
Through a research collaboration across 20 hospitals, we are able to study how we can apply mathematical modeling to predict survivability from cardiac arrest. This is an ongoing collaboration and has allowed us to collect data from more than 500 patients in the U.S. and Europe. We also collected data on patients’ medical histories as well as the events that triggered cardiac arrest to occur. In all, we obtained more than 120 data points per patient. With this information, we aim to reach a comprehensive understanding of the current standard of care for cardiac arrest.
Using sophisticated mathematical modeling tools, we plan to use this data to develop a profile of characteristics for cardiac arrest patients in two categories: those who survived without severe brain damage and those who did not survive or who survived but with severe brain damage. By comparing certain characteristics of patients who do survive without severe brain damage and the treatments they received, we aim to identify subgroups of patients who are most responsive to cardiopulmonary resuscitation (CPR), as well as those who would benefit from augmented resuscitation techniques.
These powerful data will give physicians the unprecedented ability to make real-time decisions based on predictive science. In turn, the results of new resuscitation standards allow us to develop targeted methods to augment and improve resuscitation beyond standard CPR and therefore improve both the percentage of those who survive and their chances of avoiding neurological deficits.