Does a Diagnostic Classification Algorithm Help to Predict the Course of Low Back Pain? A Study of Danish Chiropractic Patients With 1-Year Follow-up

BACKGROUND: A diagnostic classification algorithm, "the Petersen classification,' consisting of 12 categories based on a standardized examination protocol, was developed for the primary purpose of identifying clinically homogeneous subgroups of individuals with low back pain (LBP). OBJECTIVES: To investigate whether a diagnostic classification algorithm is associated with activity limitation and LBP intensity at follow-up assessments of 2 weeks, 3 months, and 1 year, and whether the algorithm improves outcome prediction when added to a set of known predictors. METHODS: This was a prospective observational study of 934 consecutive adult patients with new episodes of LBP who were visiting chiropractic practices in primary care and categorized according to the Petersen classification. Outcomes were disability and pain intensity measured with questionnaires at 2 weeks and 3 months, and 1-year trajectories of LBP based on weekly responses to text messages. Associations were analyzed with linear and logistic regression models. In a subgroup of patients, the numbers of visits to primary and secondary care were described. RESULTS: The Petersen classification was statistically significantly associated with all outcomes (P<.001) but explained very little of the variance (R ² = 0.00-0.05). Patients in the nerve root involvement category had the most pain and activity limitation and the most visits to primary and secondary care. Patients in the myofascial pain category were the least affected. CONCLUSION: The Petersen classification was not helpful in determining individual prognosis in patients with LBP receiving usual care in chiropractic practice. However, patients should be examined for potential nerve root involvement to improve prediction of likely outcomes.