Background: Major Depressive Disorder (MDD) and Generalised Anxiety Disorder (GAD) have the highest comorbidity rates within the internalising disorders cluster, yet no internet-based Cognitive Behavioural Treatment (iCBT) exists for their combined treatment. Methods: We designed a 6-lesson therapist-assisted iCBT program for mixed anxiety and depression. Study 1 was a Randomised Controlled Trial (RCT) comparing the iCBT program (n=46) versus Wait-List Control (WLC, n=53) for patients diagnosed by structured clinical interview with MDD, GAD or co-morbid GAD/MDD. Primary outcome measures were the PHQ-9 (depression), GAD-7 (generalised anxiety), K-10 (distress) and WHODAS-II (disability). The iCBT group was followed-up at 3-months post-treatment. In Study 2, we investigated the adherence to, and efficacy of the same program in a primary care setting, where patients (n=136) completed the program under the supervision of primary care clinicians. Results: The RCT showed that the iCBT program was more effective than WLC, with large within- and between-groups effect sizes found (>.8). Adherence was also high (89%), and gains were maintained at 3 month-follow-up. In Study 2 in primary care, adherence to the iCBT program was low (41%), yet effect sizes were large (>.8). Thirty per cent of non-completers experienced benefit. Conclusions: Together the results show that iCBT is effective and adherence is high in research settings, but there is a problem of adherence when translated into the ‘real world.’ Future efforts need to be placed on developing improved adherence to iCBT in primary care settings.