ABSTRACT Fear avoidance behaviour is associated with slow recovery from mild traumatic brain injury (mTBI). This study is a preliminary evaluation of graded exposure therapy (GET), which directly targets fear avoidance behaviour, for reducing post-concussion symptoms (PCS) and disability following mTBI. In a historical comparison design, we compared two groups from independent randomized trials. The GET + UC group (N = 34) received GET (delivered over 16 videoconference sessions) in addition to usual care (UC). The historical comparison group (N = 71) received UC only. PCS severity (Rivermead Post Concussion Symptoms Questionnaire; RPQ) and disability (World Health Organization Disability Assessment Schedule; WHODAS 2.0 12-item) were measured at clinic intake (M = 2.7, SD = 1.1 months after injury) and again at M = 4.9 (SD = 1.1) months after injury. Between-group differences were estimated using linear mixed effects regression, with a sensitivity analysis controlling for injury-to-assessment intervals. The estimated average change on the RPQ was −14.3 in the GET + UC group and −5.3 in the UC group. The estimated average change on the WHODAS was −5.3 in the GET + UC group and −3.2 in the UC group. Between-group differences post-treatment were −5.3 on the RPQ and −1.5 on the WHODAS. Treatment effects were larger in sensitivity analyses. Findings suggest that a randomized controlled trial is warranted.