Leena Sharma, MD, ’89 GME, the Chang-Lee Professor of Preventive Rheumatology, was the first author of the study published in Annals of the Rheumatic Diseases.

Northwestern Medicine scientists have developed and validated a tool to predict which patients with pre-osteoarthritis are at high risk for developing disability in the future. Such information could prompt patients not yet afflicted with knee osteoarthritis to take steps to prevent disability.

“Prevention at this early stage could have dramatic effects on the course of an individual life and on the overall burden of knee osteoarthritis disability to society,” said Leena Sharma, MD, ’89 GME, the Chang-Lee Professor of Preventive Rheumatology and lead author of the study published in Annals of the Rheumatic Diseases.

Osteoarthritis is the most common form of human arthritis, and knee osteoarthritis in particular is a major cause of disability in older adults.

Strategies that can help prevent disability — such as increasing physical activity — become more difficult once patients have already developed knee osteoarthritis, due to pain and joint damage associated with the disease.

As such, focusing prevention strategies on people with pre-osteoarthritis — who are at high risk for knee osteoarthritis, but do not yet have established disease — may be a more effective approach. However, the population of adults with pre-osteoarthritis is large, and there is currently no reliable method to predict which patients are most at risk for developing disability in the future.

In the current study, investigators created and validated risk stratification trees, where patients are subdivided into branches based on their attributes.

To do so, the investigators began by considering 40 easily measurable patient variables — such as age, sex, health insurance status, knee symptoms and family history of knee replacement — from a prospective, longitudinal cohort study called the Osteoarthritis Initiative (OAI).

“We then used machine learning analytic techniques and the OAI data to develop the best trees, i.e., the combination of these variables that best predicted the outcome,” explained Sharma, a professor of Medicine in the Division of Rheumatology and of Preventive Medicine.

The investigators found two types of risk trees worked best. In the simpler version, baseline age and knee function were used to divide patients in one of three groups — two at various levels of high risk for impaired function and one at low risk.

In an alternative version, six variables — baseline age, knee function, education, strenuous activity, obesity and high depressive symptoms — were used to divide the population into seven different levels of risk.

The team then validated the risk trees using another large dataset, the Multicenter Osteoarthritis Study.

“Our findings suggest that in the pre-osteoarthritis population, risk of functional decline can be estimated using easily acquired data,” Sharma said.

Moving forward, the investigators will develop simple questionnaires based on the trees, which can help patients and providers predict future risk for disability.

“We will also study how best to disseminate and implement the questionnaires in clinic and community settings, as a tool to enhance awareness of risk of impaired function at an early stage — which may motivate steps to prevent decline,” Sharma said.

Other Northwestern Medicine co-authors of the study included Joan Chmiel, PhD, professor of Preventive Medicine in the Division of Biostatistics; Jungwha (Julia) Lee, PhD, associate professor of Preventive Medicine in the Division of Biostatistics; Alison Chang, PT, DPT, MS, associate professor of Physical Therapy and Human Movement Sciences; Jing Song; and Orit Almagor.

The study was supported by NIH/NIAMS grants R01 AR065473, R01AR066601 and P30AR072579.