How effectively do experts predict elderly target-users of assistive devices? importance of expert knowledge in device development

Abstract

Many assistive devices have been developed to assist in the activities of daily life (ADLs) of frail elderly people. However, one of the significant barriers to device development is to conduct multiple user tests due to burdensome to the elderly. Therefore, the potential need for devices must be determined efficiently with some limited opportunities. In this study, we examined whether it is possible to identify target users who can adapt to, and continuously use the device, by utilizing the knowledge of experts who are familiar with the mental and physical conditions of the elderly. As a case study, we analyzed a two-month user test of 57 elderly people with the use of a four-wheel electrically assisted cycle. As a result, the accuracy rate of the expert’s prediction of continuous/discontinuous use over the whole study period was 66.7% before use, but it was improved to 87.7% when the prediction was made again after two rides by elderly people. We also attempted to model this identification rule by multiple regression analysis, and found that experts could predict whether users would use the device long-term by evaluating the following three factors: 1) subjects’ willingness to exercise, 2) their anxiety and dissatisfaction associated with the product, and 3) the ease with which they became fatigued. It is thus proposed that target users can be identified by having a small number of elderly people use the equipment several times, and obtaining expert predictions regarding whether users will continue to use or stop using the device.