Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.797862
Title: Intervention Evolution Engine (IEE) : a machine learning-driven, intelligent decision support platform for the advancement of eHealth service delivery
Author: Ahmed, Bakhtiyar
ISNI:       0000 0004 8505 6109
Awarding Body: Kingston University
Current Institution: Kingston University
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Advancement and progressive integration of ICT into health systems and services, known as eHealth, has opened the doors to endless opportunities to improve the health and wellbeing of individuals. However, there still seems to be a considerable divide in combining eHealth with the Healthcare Service Delivery process, making it difficult for both patients and healthcare professionals to attain the maximum benefits from routine clinical practice. Increasing choice and complexity of modern treatments make it difficult for clinicians to know and offer the best alternatives to patients as well as monitor outcomes to provide optimal care. Patients, in turn, struggle to adhere to complex treatments providing better results. Nonetheless, eHealth can overcome such a gap by improving the clinical decision-making process by automating the routine aspects of healthcare service delivery in addition to facilitating clinical optimisation by collecting and analysing outcome data. This research proposes an intelligent eHealth-based platform called Intervention Evolution Engine (IEE) that utilises the potential of Machine Learning to automate several routine clinical practices. These innovative practices will see the use of several predictive analytics to automate the NHS's Stepped Care model, alongside a recommender system that will facilitate the automation of randomised A/B Testing into the intervention recommendation process. Additionally, based on patient insight, using a continuous feedback mechanism, the IEE will also continuously optimise discreet interventions by replacing the least effective interventions with new alternatives/variants. In doing so, the intention is to discover not only the most effective treatments but also the optimal administration of these treatments. A further contribution of this research is the development of the Lean Design Thinking Methodology (LDTM), a novel System Development Methodology created to guide and improve the outcomes of Machine Learning projects. This study documents the Systems Development Life Cycle of the IEE, starting with a literature review and benchmark analysis, which identifies the relevant opportunistic areas of eHealth research and development unfulfilled by prior studies. This is followed by a requirements specification that documents the unique requirements of the IEE as identified by the stakeholders of this research. These requirements are then structured in the design description, which acts as a detailed blueprint for the subsequent implementation activities. Using the LDTM, the development of the IEE is carried out over several iterations to arrive at the final solution centred on the results of the IEE's underlying predictive analytics and recommender system, which have been improved incrementally throughout the development process. The research concludes with a verification and validation study conducted to establish confidence over the IEE's fitness for purpose.
Supervisor: Philip, Nada ; Dannhauser, Thomas Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.797862  DOI: Not available
Keywords: Intelligent Decision Support System ; eHealth ; information communication technology ; machine learning ; healthcare service delivery
Share: