The early detection and diagnosis of frailty is important in improving the quality of elderly life. Physicians must take many factors into account to achieve an objective and standardised method of frailty assessment. In this work, we present a novel mobile service infrastructure to improve the accuracy of frailty diagnosis in a population of elders, focusing on the main frailty risk factors and applying similarity algorithms based on Gower's coefficient. In addition, we have developed a mobile application to obtain the assessment results from the implemented services. The results are displayed on the mobile device using hierarchical visualisation structures known as treemaps. The system has been evaluated as a frailty diagnosis tool in a group of elderly patients.