Early detection of Rheumatoid Arthritis (RA) and other neurological conditions is vital for effective treatment. Existing methods of detecting RA rely on observation, questionnaires, and physical measurement, each with their own weaknesses. Pharmaceutical medications and procedures aim to reduce the debilitating effect, preventing the progression of the illness and bringing the condition into remission. There is still a great deal of ambiguity around patient diagnosis, as the difficulty of measurement has reduced the importance that joint stiffness plays as an RA identifier. The research areas of medical rehabilitation and clinical assessment indicate high impact applications for wearable sensing devices. As a result, the overall aim of this research is to review current sensor technologies that could be used to measure an individual’s RA severity. Other research teams within RA have previously developed objective measuring devices to assess the physical symptoms of hand steadiness through to joint stiffness. Unfamiliar physical effects of these sensory devices restricted their introduction into clinical practice. This paper provides an updated review among the sensor and glove types proposed in the literature to assist with the diagnosis and rehabilitation activities of RA. Consequently, the main goal of this paper is to review contact systems and to outline their potentialities and limitations. Considerable attention has been paid to gloved based devices as they have been extensively researched for medical practice in recent years. Such technologies are reviewed to determine whether they are suitable measuring tools.

A system for automatically training and spotting signs from continuous sign language sentences is presented. We propose a novel multiple instance learning density matrix algorithm which automatically extracts isolated signs from full sentences using the weak and noisy supervision of text translations. The automatically extracted isolated samples are then utilized to train our spatiotemporal gesture and hand posture classifiers. The experiments were carried out to evaluate the performance of the automatic sign extraction, hand posture classification, and spatiotemporal gesture spotting systems. We then carry out a full evaluation of our overall sign spotting system which was automatically trained on 30 different signs.

Background: The increased use of wearable sensor technology has highlighted the potential for remote telehealth services such as rehabilitation. Telehealth services incorporating wearable sensors are most likely to appeal to the older adult population in remote and rural areas, who may struggle with long commutes to clinics. However, the usability of such systems often discourages patients from adopting these services. Objective: This study aimed to understand the usability factors that most influence whether an older adult will decide to continue using a wearable device. Methods: Older adults across 4 different regions (Northern Ireland, Ireland, Sweden, and Finland) wore an activity tracker for 7 days under a free-living environment protocol. In total, 4 surveys were administered, and biometrics were measured by the researchers before the trial began. At the end of the trial period, the researchers administered 2 further surveys to gain insights into the perceived usability of the wearable device. These were the standardized System Usability Scale (SUS) and a custom usability questionnaire designed by the research team. Statistical analyses were performed to identify the key factors that affect participants’ intention to continue using the wearable device in the future. Machine learning classifiers were used to provide an early prediction of the intention to continue using the wearable device. Results: The study was conducted with older adult volunteers (N=65; mean age 70.52, SD 5.65 years) wearing a Xiaomi Mi Band 3 activity tracker for 7 days in a free-living environment. The results from the SUS survey showed no notable difference in perceived system usability regardless of region, sex, or age, eliminating the notion that usability perception differs based on geographical location, sex, or deviation in participants’ age. There was also no statistically significant difference in SUS score between participants who had previously owned a wearable device and those who wore 1 or 2 devices during the trial. The bespoke usability questionnaire determined that the 2 most important factors that influenced an intention to continue device use in an older adult cohort were device comfort (τ=0.34) and whether the device was fit for purpose (τ=0.34). A computational model providing an early identifier of intention to continue device use was developed using these 2 features. Random forest classifiers were shown to provide the highest predictive performance (80% accuracy). After including the top 8 ranked questions from the bespoke questionnaire as features of our model, the accuracy increased to 88%. Conclusions: This study concludes that comfort and accuracy are the 2 main influencing factors in sustaining wearable device use. This study suggests that the reported factors influencing usability are transferable to other wearable sensor systems. Future work will aim to test this hypothesis using the same methodology on a cohort using other wearable technologies.

Objective: Current methods of assessing the affect patients' health has on their daily lives are extremely limited. The aim of this paper is to develop a sensor-based approach to health status measurement in order to objectively measure health status. Methods: Techniques to generate human behavior profiles, derived from the smartphone accelerometer and gyroscope sensors, are proposed. Experiments, using SVM regression models, are then conducted in order to evaluate the use of the proposed behavior profiles as a predictor of health status. Results: Experiments were conducted on data from 171 participants, with an average of 114 h of data per participant. Regression models were trained and tested on the 10 SF-36 self-ratings. Results showed that the eight individual SF-36 scales and two component scores could be predicted with an average correlation of 0.683 and 0.698, respectively. General health was predicted with an average correlation of 0.752. Conclusion: Research shows that the clinically important difference for SF-36 self-ratings are approximately 10 points. Health status prediction errors in this study were 11.7 points on average. While the problem has not been fully solved, this paper presents a hugely promising direction for health status prediction. Significance: Using the proposed techniques, health status could be measured using unobtrusive, inexpensive, and already available hardware. It could provide a means for clinicians to accurately and objectively assess the daily life benefits of treatments on an individual patient basis.