Factors associated with physiotherapists' preference for MRI in primary care patients with low back and leg pain

Background: Criticisms about overuse of MRI in low back pain are well documented. Yet, with the exception of suspicion of serious pathology, little is known about factors that influence clinicians' preference for magnetic resonance imaging (MRI) at first consultation. Objective: To explore factors associated with physiotherapists' preference for MRI for patients consulting with benign low back and leg pain (LBLP) including sciatica. Design: Cross-sectional cohort study. Methods: Data were collected from 607 primary care LBLP patients participating in the ATLAS cohort study. Following clinical assessment, physiotherapists documented whether he/she wanted the patient to have an MRI. Factors potentially associated with physiotherapists' preference for imaging were selected a priori from patient characteristics and clinical assessment findings. A mixed-effects logistic regression model examined the associations between these factors and physiotherapists' preference for MRI. Results: Physiotherapists expressed a preference for MRI in 32% (196/607) of patients, of whom 22 did not have a clinical diagnosis of sciatica (radiculopathy). Factors associated with preference for MRI included; clinical diagnosis of sciatica (OR 4.23: 95% CI 2.29, 7.81), greater than 3 months pain duration (2.61: 1.58, 4.30), high pain intensity (1.24: 1.11, 1.37), patient's low expectation of improvement (2.40: 1.50, 3.83), physiotherapist's confidence in their diagnosis (1.19: 1.07, 1.33), with greater confidence associated with higher probability for MRI preference. Conclusion: A clinical diagnosis of sciatica and longer symptom duration were most strongly associated with physiotherapists' preference for MRI. Given current best practice guidelines, these appear to be justifiable reasons for MRI preference at first consultation.