Identifying spatial and temporal pain profiles to identify disease type and progression from the Manchester Digital Pain Manikin

点击翻译按钮获取中文摘要

Chronic pain drives disability in people with musculoskeletal and other chronic conditions and affects approximately one in five people worldwide. Chronic pain leads to deterioration people's physical and mental health, which in turn causes disability that results in lower productivity, increased work absenteeism and impaired social functioning. Precise figures on pain prevalence are still largely unknown and further knowledge gaps exist with respect to what causes pain and how best to manage it. To address this, researchers need validated methods to measure pain in large, representative populations.Pain manikins, also known as pain maps or pain diagrams, e human body-shaped figures that -compared to text-based questionnaires-enable intuitive self-reporting of pain location by shading or selecting affected body areas [1]. We have developed the Manchester Digital Pain Manikin which enables people to quickly and intuitively self-report pain location and location-specific pain intensity on their smartphone [2]. Pain manikins are currently used to accurately calculate a patient's pain extent, i.e. what percentage of their body is affected by pain. However, the detailed data provided through the pain manikin means that it may be possible to extract additional information about a patients's condition and prognosis by analysing the spatial patterns (i.e. where the pain is located) and temporal patterns (i.e. how does pain change over time).These types of complex patterns may be identified using machine learning methods. Such methods have previously been shown to be effective in medical imaging applications [3]. In the first instance, you will have access to a data set collected as part of the Manchester Digital Pain Manikin feasibility study. For this, 108 people with a clinician diagnosis of rheumatoid arthritis, osteoarthritis or fibromyalgia will submit daily manikin reports for 30 days, alongside a single item asking them about their overall pain intensity for that day. They will also complete a more extensive pain questionnaire at baseline and again at the last day of follow-up.In this PhD project, you will:1. Gain insight in the current state of play of (machine learning) methods for analysing digital manikin data2. Develop skills to apply and create machine learning techniques for analysis of digital pain manikin data to identify disease types and trajectories3. Learn how to incorporate manikin-based analytics into a clinical decision support tool for diagnosis and/or monitoring.