The cost to society of mental illness is substantial. A large scale international study has identified mental illnesses as the second leading cause of disability and premature mortality in the developed world [Murray, C.L., Lopez, A.D. (Eds.), 1996. The Global Burden of Disease: A comprehensive assessment of mortality and disability from disease, injuries, and risk factors in 1990 and projected to 2020. Harvard University. Cambridge. MA]. Unfortunately, research also suggests that the majority of people suffering from treatable mental health disorders do not have access to the required treatment. Furthermore, even when treatment is accessible many sufferers are unable to successfully engage with professional services [Surgeon General, 1999. Mental Health: A Report of the Surgeon General - Executive Summary. Department of Health and Human Services, Washington, DC, Retrieved August 2006, from http://www.surgeongeneral.gov/library/mentalhealth/home.html; WHO World Mental Health Survey Consortium, 2004. Prevalence, severity, and unmet need for treatment of mental disorders in the World Health Organization World Mental Health Surveys. Journal of the American Medical Association, 291(21)]. Computer assisted mental health interventions have the potential to help in addressing this imbalance. However, a review of literature shows that to date this potential has been largely unexplored. One of the primary reasons for this is that few researchers from a HCI or technical background have engaged in this area. The primary purpose of this paper is to provide a foundation and set an agenda for future research on the design of technology for talk-based mental health interventions. Theoretical approaches to the treatment of mental illness are reviewed, as is previous research on the use of technology in this area. Several significant factors effecting design and evaluation are identified and based on these factors a broad set of design guidelines are proposed to aid the development of new technologies. Of the issues identified, ethical requirements along with the sensitivity and stigma associated with mental illness pose particular challenges to HCI professionals. These factors place strict limitations on access to mental health care (MHC) settings by non-MHC professionals and create difficulties for the direct application of traditional HCl methods, such as participatory, user-centred and iterative design. To overcome these difficulties this paper proposes a model for collaborative design and evaluation, involving both HCI and MHC professionals. The development of adaptable technologies is an important element of the proposed approach. The final contribution of the paper is to suggest future research directions and identify ways in which HCI researchers can contribute to this work. (C) 2007 Elsevier B.V. All rights reserved.