Operative Treatment is Not Associated with More Relief of Depression Symptoms than Nonoperative Treatment in Patients with Common Hand Illness

Abstract

Depression symptoms are prevalent in the general population, and as many as one in eight patients seeing a hand surgeon may have undiagnosed major depression. It is not clear to what degree lower mood is the consequence or cause of greater symptoms and limitations. If depressive symptoms are a consequence of functional limitations, they might be expected to improve when pathophysiology and impairment are ameliorated. Because surgical treatment is often disease-modifying or salvage, surgery might have a greater impact than nonoperative treatment, which is more often palliative (symptom relieving) than disease-modifying. (1) For which hand or wrist conditions are depression symptoms lower after operative compared with nonoperative treatment? (2) Among the subset of patients with the highest depression scores, are depression symptoms lower after operative treatment compared with nonoperative treatment? (3) Among the subset of patients who had nonoperative treatment, are depression symptoms lower after a corticosteroid injection compared with no specific biomedical intervention? At an academic orthopaedic department, 4452 patients had a new office visit for carpal tunnel syndrome, benign neoplasm, primary hand osteoarthritis, de Quervain's tendinopathy, or trigger digit. We analyzed the 1652 patients (37%) who had a return visit at least 3 months later for the same diagnosis. Patients completed the Patient-reported Outcomes Measurement Information System (PROMIS) Depression computerized adaptive test at every office visit (higher scores indicate more depression symptoms) and PROMIS Pain Interference (higher scores indicates greater hindrance in daily life owing to pain). Patients with a return visit were more likely to have surgical treatment and had greater Pain Interference scores at the first visit. Thirteen percent of patients (221 of 1652) had incomplete or missing scores at the initial visit and 33% (550 of 1652) had incomplete or missing scores at the final return visit. We used multiple imputations to account for missing or incomplete data (imputations = 50). In a multivariable linear regression analysis, we compared the mean change in Depression scores between patients treated operatively and those treated nonoperatively, accounting for PROMIS Pain Interference scores at the first visit, age, gender diagnosis, provider, and treatment duration. A post-hoc power analysis demonstrated that the smallest patient cohort (benign lump, n = 176) provided 99% power (α = 0.05) with eight predictor variables to detect a change of 2 points in the PROMIS Depression score (minimally important difference = 3.5). After controlling for potentially confounding variables such as pain interference and age, only carpal tunnel release was associated with a slightly greater decrease in depression symptoms compared with nonoperative treatment (regression coefficient [RC] = -3 [95% confidence interval -6 to -1]; p = 0.006). In patients with the highest PROMIS Depression scores for each diagnosis, operative treatment was not associated with an improvement in depression symptoms (carpal tunnel release: RC = 5 [95% CI -7 to 16]; p = 0.44). Moreover, a corticosteroid injection was not associated with fewer depression symptoms than no biomedical treatment (carpal tunnel release: RC = -3 [95% CI -8 to 3]; p = 0.36). Given that operative treatment of hand pathology is not generally associated with a decrease in depression symptoms, our results support treating comorbid depression as a separate illness rather than as a secondary effect of pain or physical limitations. Level II, therapeutic study.