Medical error is the third leading cause of death in the US. In fact, preventable medical errors have been reported to be a lethal problem since two decades ago [1][2]. The latest studies have estimated the number of preventable deaths at over 22,000 a year in the United States [3]. A considerable portion of these preventable errors are caused by unintended deviation from the clinical best practice algorithms during a fast-paced and dynamically changing clinical scenario.
Our solution to this is computerized clinical guidance systems, which is a form of clinical decision support systems that provide dynamic guidance and assistance to healthcare practitioners during their workflow in real-time at the bedside. We use computational pathophysiology to transform the informal description of medical knowledge to precise software that tracks and records complex disease dynamics. Like GPS navigation transforming the navigation history, we shall demonstrate that AI-enhanced and computational pathophysiology model-based medical best practice guidance systems will transform clinical practice.
We collaborate with exceptional expert physicians and researchers from various hospitals and medical research centers to ensure the medical correctness and safety of our computational pathophysiology models and conduct clinical evaluations to verify the effectiveness of our guidance systems.
Poor quality cardiopulmonary resuscitation (CPR) is directly linked to detriments in patient morbidity and mortality. Despite advances in resuscitation continuous quality improvement (CQI) programs, in-hospital pediatric arrest survival rates remain low at a median of 36%. Our objective for this research is to develop a Pediatric Advanced Cardiovascular Life Support Guidance System that will uniformly improve the quality of resuscitation events in real-time by functioning as a dynamic cognitive aid, decision support tool and agent promoting adherence to resuscitation algorithms and CRM. We hypothesize the utilization of this system during cardiopulmonary arrests will improve both adherence to guidelines as well as CRM exhibited by team performance. The knowledge gained from this research will further inform the development of dynamic cognitive aids and decision support technology to improve the quality of resuscitation events and team performance in real-time.
Sepsis is life-threatening organ dysfunction caused by a dysregulated host response to infection. Sepsis can rapidly lead to tissue damage, organ failure, and death. According to CDC statistics, every year more than1.7 million adults in the U.S. develop sepsis, and nearly 270,000 die from sepsis. Identification of sepsis in its early stages upon presentation to the emergency department is vital in preventing significant organ injury, prolonged hospitalization, and potentially death. We propose a computerized pediatric sepsis best practice guidance system for early detection, diagnosis, and treatment of pediatric sepsis.
The surge of severely ill COVID-19 patients greatly exceeds the available medical staff who have been trained to care for them. To address the lack of training in COVID-19 treatment and minimize the risk of preventable medical error under this stressful pandemic, we are funded by C3.ai to create AI-assisted COVID-19 ARDS (Acute Respiratory Distress Syndrome, i.e. lung failure) guidance system. This system models ARDS treatment guidelines published by authoritative medical associations using computational pathophysiology and incorporates patient status prediction models powered by c3.ai suite.