USE OF HEALTHCARE RESOURCES IN PATIENTS WITH LOW BACK PAIN AND COMORBID DEPRESSION OR ANXIETY

Abstract Background Context Psychological comorbidities are important prognostic factors for low back pain (LBP).  To develop improved treatment paradigms, it is first necessary to characterize and determine current patterns of treatment in this population. Purpose Identify how comorbid depression or anxiety in patients with LBP is related to use of healthcare resources. Study Design/Setting Retrospective cohort study using electronic health records from outpatient offices at a large multisite academic medical center. Patient Sample Data from 513,088 unique patients seen between 1/2010-7/2020 (58.0% female, 52.6±19.5 years) with a diagnosis of LBP, indicated by predetermined ICD-9 and ICD-10 codes. Outcome Measures Average self-reported pain scores, absolute differences and unadjusted risk ratios to compare opioid use, emergency department visits, hospitalizations, advanced imaging orders, spinal injections, and back surgeries between cohorts. Methods Clinical characteristics and data regarding use of healthcare resources were extracted from the electronic health record. Clinical features and patterns in healthcare utilization were determined for patients with depression or anxiety compared to those without. Results Depression or anxiety was coded for 21.4% of patients at first LBP visit. Those with depression or anxiety were more likely to be on opioids (unadjusted RR: 1.22, CI: [1.22,1.23]), go to the emergency department (1.31 [1.30-1.33]), be hospitalized (1.15 [1.13, 1.17]), receive advanced imaging (1.09 [1.08, 1.11]), receive an epidural steroid injection (1.16 [1.15, 1.18]), and less likely to have back surgery (0.74 [0.72, 0.77]). Differences in pain scores for those with depression/anxiety compared to those without were not clinically significant. Conclusions Depression/anxiety is associated with increased use of healthcare resources, and is not associated with clinically meaningful elevated pain scores. Limitations come from use of an aggregate data set and reliance on administrative coding.