Daniel Corkill, Joseph Gormley and Thomas Zisk
Healthcare is experiencing a crisis. Institutions are suffering multiple business and service level failures, including suboptimal outcomes, overburdened healthcare workers, institutional cost overruns, and reduced reimbursement levels. The healthcare system can no longer survive using static patient models, conflicting therapeutic guidelines, and ineffective care-coordination policies. New methods and tools are needed to effectively identify and quantify rising health risks and recommend effective interventions both during and between clinical encounters. This challenge applies to both U.S. healthcare organizations and global healthcare initiatives where the need for earlier interventions within low-resourced regions is particularly acute.
IOMICS developed PrecisionCARETM, an analytics system for personalized medicine, to be different along three important dimensions. First, PrecisionCare automatically mines a population dataset 24/7 to identify clinical and molecular features most relevant to risk profiling and predicting intervention impact. Using advanced machine learning technologies, each patient’s individual medical history, genetic profile, environment, and lifestyle are used to identify clinical strategies demonstrated to reduce comorbidities and minimize adverse events in similar patients. Second, PrecisionCARE incorporates real-time data quality management and automated model testing strategies to overcome the most debilitating issue in healthcare analytics today: effective patient modeling in the presence of missing, erroneous, or inconsistent clinical data. Third, PrecisionCARE automatically characterizes uncertainty within each intervention model’s response by assessing all available evidence, ensuring the most effective recommendation possible during each clinical encounter. Combined, these unique capabilities provide a new level of model transparency and advance the ability to quantify intervention confidence within a personalized medicine context.
PrecisionCARE has benefited client organizations in three significant ways. First, it creates and uses precision patient models in real time – an industry first. Second, it decreases the need for a large data science department, reducing staffing cost by approximately 60 percent while simultaneously eliminating process errors. Third, and most importantly, PrecisionCARE’s interventions have consistently demonstrated a higher level of clinical relevancy over existing static models and metrics for both high-risk individuals and complex patient types across tested disease models.