New Software Measures Mortality Risk at Admission

There is a new software tool which has the ability to calculate a patient’s mortality risk at the moment of hospital admission.

This is a tool, say the developers, which can be used to guide the care of older patients who may have higher risks of complications.

Researchers from NYU Langone Health are tackling the scourge of geriatric trauma with their recent work, “How Does Frailty Factor Into Mortality Risk Assessment of a Middle-Aged and Geriatric Trauma Population?” The article appeared in the December 2017 edition of Geriatric Orthopaedic Surgery & Rehabilitation.

Sanjit Konda, M.D., with the Department of Orthopedic Surgery at NYU Langone Health, told OTW, “Despite a plethora of publications on hip fractures over the last 30 years—and lots of innovation—the hip fracture mortality rate in the elderly has not changed.”

“If we are successful in fixing a patient’s hip the mortality rate is approximately 25-30% at one year; without surgery that rate is around 60%. We wanted to develop a risk model of mortality in this population so that we could predict the risk of inpatient mortality (including 30 day and 1 year postop mortality risk) and thus help guide their care.”

“Our team created a risk-cost prediction model looking at value based treatment methods for these patients when they come into the hospital. We incorporated this into a software program—PersonaCARE—that uses predictive analysis and a deep learning algorithm to help identify those most at risk of mortality. The algorithm has previously been published as The Score for Trauma Triage in the Geriatric and Middle-Aged (STTGMA).”

“STTGMA evaluates four physiologic criteria: age, comorbidities, vital signs, and anatomic injuries. Our work, which included 1,486 patients over the age of 55, looked at whether adding specific frailty variables to STTGMA would improve risk stratification of a middle-aged and elderly trauma population. It turns out that using these additional frailty factors did not increase the model’s predictive ability. So essentially, our original algorithm already accounts for enough frailty variables to help guide patient care.”

“Right from the outset we use PersonaCARE to determine a patient’s risk of inpatient and long-term mortality, something that helps us manage patient expectations about mobility and the need for home health. If it appears that someone is not going to do well then we can ‘target’ them (and their families) for more support.”

“We are also using PersonaCARE to access palliative care. Many people think that you only get palliative care involved when someone is on the brink of death. The fact is, however, that 60% of patients who undergo a palliative care consult live for another five years.”

“Going forward I would like to add to PersonaCARE. We have modeled it for hip and femur fracture, both arthroplasty and fixation, and will expand it to all other orthopedic service lines. The software and underlying models are also sufficiently robust to transfer to other medical specialties.”