Investigate the effectiveness of Concussion Coach, an interactive smartphone application, as a treatment for residual neurobehavioral symptoms and distress in Veterans with a history of mild traumatic brain injury (mild TBI). Veterans with mild TBI were randomized to Concussion Coach (n = 238) or Treatment-as-Usual (TAU) (n = 241) in a 3-month randomized controlled trial. Primary outcome measures included postconcussive symptom (PCS) severity as measured by the Neurobehavioral Symptom Inventory (NSI), and psychological distress as measured by the Brief Symptom Inventory-18 (BSI-18). Measures of self-efficacy, social support, and comfort with technology were administered as potential moderators and mediators. An intention-to-treat (ITT) analysis was performed (N = 461: Concussion Coach = 231 and TAU = 230) using Bayesian Network (BN)modeling. The probability of decreased PCS severity was significantly greater for those assigned to Concussion Coach, .35 [.32,.37], than for TAU, .29 (.27, .32), with an odds ratio (OR) of 1.29. Also, Concussion Coach showed a significantly greater probability of increased self-efficacy (.36 [.32, .39]) than did TAU (.28 [.25, .30], OR = 1.42). In turn, self-efficacy (increased vs. decreased) showed a significantly greater probability of decreased PCS severity (.51 [.47, .54] vs. .27 [.24, .30], OR = 2.71) and decreased psychological distress (.53 [.49, .56] vs. .32 [.29, .35], OR = 2.35), suggesting that self-efficacy may have mediated Concussion Coach effects. Concussion Coach is effective at reducing PCS severity and psychological distress. Increased self-efficacy/perception of self-management of symptoms may be key to successful treatment of residual symptoms in those with history of concussion.