EHMRG Based on Machine Learning
In 2012, Douglas Lee et al. developed a risk model called EHMRG that predicts the 7-day mortality of both heart failure patients who were discharged and those who were hospitalized. The model was validated on a data set of approximately 12,500 patients from 86 hospitals in Ontario, Canada, who visited an emergency department for heart failure.
OcularAI, a Toronto startup with specialty in building custom-made Machine Learning and Data Science solutions, is partnering with Dr. Lee to not only bring the 7-day EHMRG model to market, but also to help build a 30-day and 365-day version of the model equipped with Machine Learning capabilities. The enhancements will be built with the objective to scale and secure wide-spread adoption. Specifically, we will:
- Adjust the model to allow better flexibility and incremental learning, taking advantage of the existing in-house data that healthcare institutions might have available. This allows for the incorporation of additional variables that might be used in heart failure management strategies.
- Enable the model to incorporate heart failure management strategies in other regions and countries.