The salient point of this research was to investigate the key factors of assistive devices for people with severe physical disabilities, i.e. spinal cord injury, when selecting and using an input device. The area of study was also concerned with validating a new computer device to enable those individuals with upper-limb impairments to engage the benefits of computer technology, via both user-issue and scientific-based evaluations. A specific methodology, concermng both user-Issue and scientific-evidence, was proposed for the studies related to assistive technology outcome measures. In order to validate the proposed methodology, the research work began with an in-depth survey (Study A), to give an insight into the present selection and utilisation of input devices among those computer users with spinal cord injuries and identify their specific needs when using a computer. Following the findings of this contextual survey, a SCI users' needs hierarchy was proposed for input device selection and use. Specific touchscreen devices, which matched the criteria in the hierarchy, were suggested as a possible solution for users with severe upper-limb disorders. Then, a series of user-centred validation studies, involving a pilot simulation study associated with a dimensional issue of an input device (Study B), followed by usability evaluations at the introductory phase (Study C), after short-term use and training (Study D) and after longer-term use and outcome comparisons (Study E), were carried out. The user perspectives and scientific data obtained from the usability assessments form the SCI subjects were used not only to demonstrate the effectiveness and efficiency of the assistive device, but also to fill the gap between the merely psychological/psychosocial-based measures and the merely scientific-focus evaluation. By merging a specific research technique and a systematic measuring procedure, a conceptual model for evaluating assistive technology outcome measures has been provided for this field of study. Moreover, this research has shown that the integration of user-issue and scientific-evidence can increase the reliability and validity of this type of device outcome measures and, therefore, attain a good match between users and the technology employed.