Eight percent of adults suffer from chronic back pain, 80% will experience back pain at some point in their lives. Back pain is a leading global cause of disability and absenteeism from the workplace with broad health and economic consequences. There is no universally effective treatment for chronic back pain that helps the majority of patients.

Researchers Angshuman Roy, Mor Peleg, and Ori Davidov are interested in developing statistical methods that could help find treatments that can improve back pain along multivariate outcomes: level of pain, enjoyment of life, ability to conduct general activities.

Toward this end, they developed a general statistical tool to test for order among multiple multivariate multinomials and applied it to back pain data to show that patients who followed doctors’ recommendations more, performed better.

They developed a statistical tool to test for order among K independent multivariate multinomials. Additionally, they also consider: (i) testing for order under longitudinal sampling; and (ii) ordered analysis in the presence of covariates. The Proposed methodology is then applied to patient reported data for studying back pain.

As expected, they found that patients who followed treatments recommended by physiotherapists performed better than those who did not adhere to treatment. Most of the recommendations related to strengthening and stretching exercises, including Yoga, stretching, massage, and postural modification exercises, which are the main treatments that have been shown to be effective in clinical trials.