Jul 31 2018
CloudMedx, a healthcare artificial intelligence (AI) company based in Palo Alto, Calif., announced today a research collaboration with the UCSF Department of Orthopaedic Surgery to study how patient-generated health care data collected from consumer-grade wearable sensors may predict clinical outcomes following hip and knee replacement surgery.
By looking at structured and unstructured data from patient medical records, as well as from wearable devices, the UCSF research team, advised by the National Science Foundation’s Center for Disruptive Musculoskeletal Innovations, aims to create a new class of algorithms that can predict a patient’s individual outcome and recovery following surgery.
“We want to combine patient-reported outcomes, data from electronic medical records and sensor data to predict how patients will recover following joint replacement surgery,” said project leader Stefano Bini, MD, UCSF Health orthopedic surgeon and professor of orthopedic surgery at UCSF. “To give us a perspective on how patients are doing with predictive analytics, we partnered with CloudMedx to handle the large data sets that will be needed.”
According to Dr Bini, the current gold standard for patient evaluations are validated patient reported outcome surveys, which are obtained prior to surgery and at one year following surgery, when the maximum improvement is deemed to have been reached. However, data points gathered in the interim have not been validated and are generally discouraged.
“The hope is that we can overcome these shortcomings by gathering data from sensors that can objectively measure in real time steps, cadence, heart rate and other variables with a patient’s clinical record to accurately predict their outcomes,” Dr. Bini said. “There currently is a huge cache of unstructured information in the medical records in the form of physician notes, nurse progress notes, discharge summaries, radiology notes and patient-reported outcomes that is being overlooked due to lack of resources. By using CloudMedx’s robust AI to read clinical notes using machine-assisted natural language processing, we aim to surface insights in real time to improve patient outcomes.”
According to Dr Bini, “Engaging Cloudmedx as our analytics platform allowed us to derive relationships and incredible insights using state of the art ML algorithms. Rather than the standard and rather useless calculation of relative Hazard and Risk Ratios of one outcome versus another when looking at one variable and adjusting for all others (“useless” because the resulting data is incredibly hard to apply in clinical practice where patients present with multiple variables many of which like gender and age are not at all variable), the algorithm was able to clearly identify cohorts (clusters) of people whose variables (features) were more likely to be associated with a given outcome (PRO) and take any specific candidate and place them in a risk cluster. This is phenomenal and way more practical.”
Very preliminary data was presented this spring at the Orthopedic Research Society. Early signals have demonstrated what kind of data is predictive of what kind of outcome and that it may predict 6-week PROs with a high degree of accuracy in a given individual as early as 2 weeks following surgery.