A prognostic model for the patient‐reported outcome of surgical treatment of carpal tunnel syndrome

Many prognostic factors have been studied in carpal tunnel decompression, but most studies consider only a subset of variables. Methods: Three thousand three hundred thirty‐two operations were used to develop prognostic models, and 885 operations were used for validation. Outcome recorded on a Likert scale was dichotomized into success or failure. Modeling was performed with both logistic regression and artificial neural networks using 87 candidate variables. Results: Both approaches produced predictive multivariate models for outcome with areas under a receiver operating characteristic curve of 0.7 in the validation data set. Patients with moderately severe nerve conduction abnormalities, night waking, a family history of carpal tunnel syndrome, a good response to corticosteroid injection, and women have better outcomes. Greater functional impairment, diabetes, hypertension, and surgery on the dominant hand are associated with poorer outcomes. Discussion: A multivariate model partially predicts the outcome of carpal tunnel surgery, aids decision making, and helps to manage patient expectations. Muscle Nerve 58 :784–789, 2018