Clinical risk assessment of chronic illnesses is a challenging and complex task which requires the utilisation of standardised clinical practice guidelines and documentation procedures in order to ensure consistent and efficient patient care. Conventional cardiovascular decision support systems have significant limitations, which include the inflexibility to deal with complex clinical processes, hard-wired rigid architectures based on branching logic and the inability to deal with legacy patient data without significant software engineering work. In light of these challenges, we are proposing a novel ontology and machine learning-driven hybrid clinical decision support framework for cardiovascular preventative care. An ontology-inspired approach provides a foundation for information collection, knowledge acquisition and decision support capabilities and aims to develop context sensitive decision support solutions based on ontology engineering principles. The proposed framework incorporates an ontology-driven clinical risk assessment and recommendation system (ODCRARS) and a Machine Learning Driven Prognostic System (MLDPS), integrated as a complete system to provide a cardiovascular preventative care solution. The proposed clinical decision support framework has been developed under the close supervision of clinical domain experts from both UK and US hospitals and is capable of handling multiple cardiovascular diseases.
Inhaltsverzeichnis (Table of Contents)
- INTRODUCTION
- Organisation of Thesis
- Motivation and aims
- Original Contributions
- Publications
- LITERATURE REVIEW
- Clinical Decision Support Systems
- Ontology Driven Clinical Decision Support Frameworks
- Clinical Decision Support Systems in Cardiovascular Care
- Cardiovascular Risk Estimation Systems for Disease Prevention
- Machine Learning Driven Cardiovascular Decision Support Systems
- Role of Feature Selection in Clinical Decision Support Systems
- Conclusion and Discussion
- A Novel Ontology and Machine Learning Driven Hybrid Clinical Decision Support Framework for Cardiovascular Preventative Care
- Proposed Framework
- ODCRARS for Cardiovascular Preventative Care
- Ontology driven intelligent context aware information collection component
- Patient Medical Records
- Ontology Driven Decision Support
- Machine Learning Driven Prognostic Modelling for Cardiovascular Preventative Care
- Machine Learning Driven Prognostic Model
- Data Acquisition
- Data Pre-Processing
- Feature Selection
- Prognostic Model Development
- Prognostic Model Validation and Evaluation
- Online Clinical Prognostic Model
- Conclusion and Discussion
- Ontology Driven Clinical Risk Assessment and Recommendation System (ODCRARS) for Cardiovascular Preventative Care
- Implementation of the Ontology Driven Clinical Risk Assessment and Recommendation System (ODCRARS)
- Ontology driven intelligent context aware information collection: Design and Implementation
- Ontology Driven Intelligent Context Aware Ontology Model
- Adaptive Clinical Questionnaire: Design and Implementation
- Proposed Novel Decision Tree based Approach
- Dynamic Adaptation
- Patient Medical Records
- Patient Semantic Profile : Design and Implementation
- Ontology Development
- Ontology Driven Clinical Decision Support: Design and Implementation
- Recommendation Ontology
- Clinical Rules Engine: Design and Implementation
- Clinical Rules Data - Patient Fact Representation
- Jess: Java based Rules Engine
- Partitioning the Rules
- Cardiovascular Risk Assessment
- System Implementation: Integration of ODCRARS and MLDPS
- Patient Module
- Doctor’s Module
- Integration of the ODCRARS with the machine learning driven cardiac chest pain and heart disease prognostic models
- Conclusion and Discussion
- Machine Learning Driven Prognostic System (MLDPS) for Cardiovascular Preventative Care
- Case Study 1: Rapid Access Chest Pain Clinic
- Background
- Aims
- RACPC Clinical Dataset 1
- Data Acquisition
- Data Preparation
- Missing Data Handling
- Feature Selection
- Prognostic Model Development: Experimental Setups and Results
- Final Diagnosis
- Evaluation of RACPC Results
- Results of Comparative Machine Learning Classification
- Analysis of Variance (ANOVA) Test for Performance Evaluation
- RACPC Clinical Dataset 2: Demonstrating Effects of missing Data on Verification Results
- Background
- Pre-processing of Missing Data using Probability Estimation
- Expectation Maximisation (EM) Approach
- Experiments
- Classification for the Incomplete Clinical Data
- Filling the Incomplete Data
- RACPC Clinical Case Study: RACPC Clinical Dataset 3
- Study Group 1: Clinical Risk Factors
- Evaluation
- Performance evaluation of experimental setups
- Study Group 2: Test Results
- Evaluation
- Performance evaluation of experimental setups
- Implementation of online Clinical Prognostic Models
- Machine Learning Driven Cardiac chest pain prognostic model’s integration with the recommendation system
- Case Study 2: Heart Disease
- Background
- Aims
- Data Preparation
- Feature Selection
- Prognostic Model Development
- Prognostic Model Validation and Evaluation
- Performance evaluation of experimental setups
- Implementation of online Clinical Prognostic Models
- Case Study 3: Breast Cancer Prognostic Modelling
- Background
- Aims
- Candidate Clinical Variable Selection
- Prognostic Model Development
- Prognostic Model Validation and Evaluation
- Performance Evaluation of Experimental Setups
- Online Clinical Prognostic Model
- Verification and Validation of the Clinical Prototypes
- Validation of the Machine Learning Driven System (MLDPS) and Ontology Driven Clinical Risk Assessment and Recommendation System (ODCRARS)
- Summary and Conclusion
- CONCLUSIONS AND FUTURE WORK
- Conclusions
- Discussion and Summary of Contributions
- Future Work
- Utilisation of Fuzzy Cognitive Maps for Collaborative Care
- Active Manifold Learning Strategy in Machine Learning Driven Prognsotic Modelling based on Big Data
- Limitations
Zielsetzung und Themenschwerpunkte (Objectives and Key Themes)
The primary aim of this research project is to develop a comprehensive clinical decision support framework for cardiovascular preventative care by combining evidence extrapolated from legacy patient data and clinical experts knowledge encoded in the form of clinical rules. The objectives of this research are:
- To design and develop a novel ontology and machine learning driven hybrid clinical decision support framework for cardiovascular preventative care.
- To develop an ontology driven clinical risk assessment and recommendation system (ODCRARS) that can provide a holistic cardiovascular decision support mechanism for clinicians and patients.
- To develop a machine learning driven prognostic system (MLDPS) that can predict the risk of cardiovascular events based on patient data.
- To integrate the ODCRARS and MLDPS to provide a comprehensive cardiovascular preventative care solution.
- To validate the proposed framework in other application areas, such as breast cancer.
Key themes of this work include:
- The use of ontologies to represent clinical knowledge and facilitate clinical decision making.
- The application of machine learning techniques to develop predictive models for cardiovascular risk assessment.
- The development of a hybrid clinical decision support framework that combines knowledge-based and data-driven approaches.
- The importance of learning from legacy patient data to improve clinical decision support systems.
- The need to address missing data and data sparsity issues in clinical datasets.
Zusammenfassung der Kapitel (Chapter Summaries)
Chapter 2 provides a comprehensive literature review of clinical decision support systems, focusing on the use of ontologies, machine learning, and hybrid approaches in cardiovascular care. It also discusses the role of feature selection in clinical decision support systems.
Chapter 3 introduces a novel ontology and machine learning driven hybrid clinical decision support framework for cardiovascular preventative care. The framework consists of two key components: an ontology driven clinical risk assessment and recommendation system (ODCRARS) and a machine learning driven prognostic system (MLDPS). The chapter outlines the design and development of the framework.
Chapter 4 focuses on the design, development, and validation of the ODCRARS. It covers the ontology driven intelligent context aware information collection component, patient semantic profile, and the NICE/Expert driven clinical rules engine.
Chapter 5 discusses the design, development, and validation of the machine learning driven prognostic system (MLDPS) through clinical case studies in the RACPC, heart disease, and breast cancer domains. The chapter explores the use of various machine learning and feature selection techniques, including missing data handling methods.
Chapter 6 summarizes the findings of the thesis and discusses future directions for research. It includes potential applications of fuzzy cognitive maps and active manifold learning for collaborative care and big data analysis, respectively.
Schlüsselwörter (Keywords)
The main keywords of this thesis are: Clinical Decision Support Systems, Cardiovascular Preventative Care, Ontology, Machine Learning, Hybrid Clinical Decision Support Framework, Prognostic Modelling, Feature Selection, Missing Data, Big Data, Fuzzy Cognitive Maps, Active Manifold Learning, RACPC (Rapid Access Chest Pain Clinic), Heart Disease, Breast Cancer.
- Quote paper
- Kamran Farooq (Author), 2015, A Novel Ontology and Machine Learning Driven Hybrid Clinical Decision Support Framework for Cardiovascular Preventative Care, Munich, GRIN Verlag, https://www.grin.com/document/334172
