Muhammad Mamdani, Professor - Status

Area of Research

Muhammad Mamdani’s current research uses machine learning and analytics to address health care problems, with a focus on improving patient outcomes and health care system efficiency.

Research Challenge

The health care system faces a number of challenges that limit its efficiency and affect patient outcomes. For example, emergency department overcrowding and long wait times are common in our health care system. In addition, patients in the hospital setting often deteriorate without advance warning. It can be extremely challenging for pharmacists to carefully assess each patient in the hospital for optimal management of medications, given how time consuming this can be.

Proposed Solution

Mamdani leads a team of data scientists who work closely with clinicians and hospital management to create advanced analytics and machine learning tools that address problems in health care.

His team works on developing artificial intelligence (AI) systems that, for example, analyze historical emergency department volumes, weather patterns, city events and other factors to accurately predict emergency department patient volumes to inform staffing decisions and reduce wait times. His team has also developed machine learning models to accurately predict patient deterioration (ICU transfer or death) 24 to 48 hours in advance so medical care can focus more efficiently on high risk patients. Further, Mamdani’s team has developed algorithms that automatically assess patients for their clinical stability and absorption status to generate daily lists of patients who may be eligible to convert from intravenous to oral antibiotics, assisting pharmacists in targeting patients for better medication management. 

Impact To Date

Mamdani’s technologies are improving the efficiency of health care services in the hospital setting. By predicting important health outcomes and the most effective interventions for individual patients, these technologies enable clinicians to make fast and informed decisions and will improve patient health outcomes and reduce health care costs. His work on predicting emergency department patient volumes affects staffing decisions in the emergency rooms. High-risk patients are more rapidly identified for intensive medical management to prevent death. ”Smart” algorithms rapidly identify patients on IV antibiotics who could be eligible for oral treatments, allowing pharmacists to save time and focus on patients who need their attention most.

In the past, Mamdani’s work has influenced drug policy changes through the Ontario Drug Policy Research Network (ODPRN), which bridges researchers with drug policy decision makers and enables rapid generation of data-driven insights. For example, a project focusing on data-driven policies for reimbursement of self-monitoring blood glucose test strips has saved the Ontario health care system at least $100 million over five years.