Low back pain and disability trajectories in primary care: a growth mixture modeling analysis

Background

People with low back pain are a heterogeneous group with diverse symptoms and recovery patterns, highlighting the need for a better understanding of trajectories toward either rapid resolution or chronicity with persistent disease burden.

Objectives

The aim of this study is to identify clustering of pain and disability trajectories and predictors thereof in adults seeking physical therapy care for low back pain.

Methods

People (n=347) were followed for 12 months in a prospective cohort. Growth classes were identified using growth mixture modeling (GMM) separately for pain and disability. Pseudo-class method was used to calculate the odd ratios between the assigned growth class for pain and disability.

Results

The best-fitting GMM identified two growth classes in both the pain and disability courses. One trajectory with people with moderate pain/disability at first visit that recovered and a second trajectory with people with moderate pain/disability at first visit that did not recover. People in the worst class for disability had higher odds of being the worst class for pain as well (OR 6.8, 95% CI 1.69-9.28). Predictors of class membership for the worst classes in pain and disability were longer duration of complaints (OR 1.24, 95% CI 1.02-1.46) for pain and higher baseline NPRS score (OR 1.31, 95% CI 1.09-1.52) and higher baseline STarT Back Screening Tool (SBST) score (OR 5.49, 95% CI 1.69-9.28).

Conclusion

The odds are high for people in the non-recovery class for disability to also be in the worst recovery class for pain. Longer duration of complaints, higher pain scores, and presence of more psychosocial factors were identified as predictors of slower recovery trajectories. People might benefit from tailored treatment based on these prognostic factors.

Registration

Clinicaltrials.gov 109643; Ethics committee RadboudUMC 2020-6295