The ability to walk over various surfaces such as cobblestones, slopes or stairs is a very patient centric and clinically meaningful mobility outcome. Current wearable sensors only measure step counts or walking speed regardless of such context relevant for assessing gait function. This study aims to improve deep learning (DL) models to classify surfaces of walking by altering and comparing model features and sensor configurations. Using a public dataset, signals from 6 IMUs (Movella DOT) worn on various body locations (trunk, wrist, right/left thigh, right/left shank) of 30 subjects walking on 9 surfaces were analyzed (flat ground, ramps (up/down), stairs (up/down), cobblestones (irregular), grass (soft), banked (left/right)). Two variations of a CNN Bi-directional LSTM model, with different Batch Normalization layer placement (beginning vs end) as well as data reduction to individual sensors (versus combined) were explored and model performance compared in-between and with previous models using F1 scores.Introduction
Method
Management of bone and joint infection can be technically complex and often requires a prolonged course of antibiotics. Traditionally, bone and joint infection management utilises nurse-led outpatient parenteral antibiotic therapy (OPAT) where adherence is unlikely to be an issue. However, with increasing evidence in favour of oral therapy, the question of adherence merits further consideration. We describe the adherence of both oral (PO) and self-administered intravenous (IV) antibiotics in the treatment of bone and joint infection using paper questionnaires (8-item Modified Morisky Adherence Score (MMAS)) and, in a subset of participants, electronic pill containers (Medication Event Monitoring Systems*). All eligible participants enrolled in the OVIVA trial (2010–2015) were randomised to six weeks of either PO or IV antibiotic treatment arms. Self-administering patients were followed up with questionnaires at day 14 and 42. A subset of PO participants was also given the medication event monitoring system* in order to validate the adherence questionnaires. The results were correlated with treatment failures at one-year follow-up.Aim
Method
