Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis
Background: Ischemic heart disease (IHD) caused by the narrowing of coronary arteries is a major cause of morbidity and mortality worldwide. Clinical diagnosis involves complex, costly, and potentially invasive procedures. Objective: To address this problem, we introduce a novel clinical knowledge-e...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-12-01
|
| Series: | Cardiovascular Digital Health Journal |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666693622001700 |
| _version_ | 1797980572249751552 |
|---|---|
| author | James Meng, MA, MB, BChir Ruiming Xing, MSc |
| author_facet | James Meng, MA, MB, BChir Ruiming Xing, MSc |
| author_sort | James Meng, MA, MB, BChir |
| collection | DOAJ |
| description | Background: Ischemic heart disease (IHD) caused by the narrowing of coronary arteries is a major cause of morbidity and mortality worldwide. Clinical diagnosis involves complex, costly, and potentially invasive procedures. Objective: To address this problem, we introduce a novel clinical knowledge-enhanced machine learning (ML) pipeline to assist in timely and cost-effective IHD prediction. Methods: Unlike conventional data-driven “black box” ML approaches, we propose an effective mechanism to engage clinical expertise and gain insight into the “black box” at each stage of model development, including data analysis, preprocessing, selecting the most clinically discriminative features, and model evaluation. One-hot feature encoding is introduced to expose hidden bias and highlight the important elements and features. Results: Experimental results on the benchmark Cleveland IHD dataset showed that the proposed clinical knowledge–enhanced ML pipeline overperformed state-of-the-art data-driven ML models, using even fewer features. Our model based on one-hot feature encoding and support vector machine achieved the best accuracy of 94.4% and sensitivity 95% by using only 7 discriminative attributes. Conclusion: We share insights and discuss the effectiveness of incorporating clinical input in machine learning to improve model performance, as well as addressing some practical issues such as data bias and interpretability. We hope this preliminary study on engaging clinical expertise to explore the “black box” would improve the trustworthiness of AI and its potential wider uptake in the medical field. |
| first_indexed | 2024-04-11T05:56:09Z |
| format | Article |
| id | doaj.art-36088ddcfdf544518878c1b5cf9f9f83 |
| institution | Directory Open Access Journal |
| issn | 2666-6936 |
| language | English |
| last_indexed | 2024-04-11T05:56:09Z |
| publishDate | 2022-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Cardiovascular Digital Health Journal |
| spelling | doaj.art-36088ddcfdf544518878c1b5cf9f9f832022-12-22T04:41:52ZengElsevierCardiovascular Digital Health Journal2666-69362022-12-0136276288Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosisJames Meng, MA, MB, BChir0Ruiming Xing, MSc1Lancashire Teaching Hospitals NHS Foundation Trust, Preston, United Kingdom; Address reprint requests and correspondence: Dr James Meng, Lancashire Teaching Hospitals NHS Foundation Trust, Sharoe Green Lane, Fulwood, Preston, UK PR2 9HT.Department of Computer Science, Loughborough University, Loughborough, United KingdomBackground: Ischemic heart disease (IHD) caused by the narrowing of coronary arteries is a major cause of morbidity and mortality worldwide. Clinical diagnosis involves complex, costly, and potentially invasive procedures. Objective: To address this problem, we introduce a novel clinical knowledge-enhanced machine learning (ML) pipeline to assist in timely and cost-effective IHD prediction. Methods: Unlike conventional data-driven “black box” ML approaches, we propose an effective mechanism to engage clinical expertise and gain insight into the “black box” at each stage of model development, including data analysis, preprocessing, selecting the most clinically discriminative features, and model evaluation. One-hot feature encoding is introduced to expose hidden bias and highlight the important elements and features. Results: Experimental results on the benchmark Cleveland IHD dataset showed that the proposed clinical knowledge–enhanced ML pipeline overperformed state-of-the-art data-driven ML models, using even fewer features. Our model based on one-hot feature encoding and support vector machine achieved the best accuracy of 94.4% and sensitivity 95% by using only 7 discriminative attributes. Conclusion: We share insights and discuss the effectiveness of incorporating clinical input in machine learning to improve model performance, as well as addressing some practical issues such as data bias and interpretability. We hope this preliminary study on engaging clinical expertise to explore the “black box” would improve the trustworthiness of AI and its potential wider uptake in the medical field.http://www.sciencedirect.com/science/article/pii/S2666693622001700Heart disease diagnosisClinical knowledge–enhanced machine learningAI interpretability and trustworthinessData-driven predictive model |
| spellingShingle | James Meng, MA, MB, BChir Ruiming Xing, MSc Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis Cardiovascular Digital Health Journal Heart disease diagnosis Clinical knowledge–enhanced machine learning AI interpretability and trustworthiness Data-driven predictive model |
| title | Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis |
| title_full | Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis |
| title_fullStr | Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis |
| title_full_unstemmed | Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis |
| title_short | Inside the “black box”: Embedding clinical knowledge in data-driven machine learning for heart disease diagnosis |
| title_sort | inside the black box embedding clinical knowledge in data driven machine learning for heart disease diagnosis |
| topic | Heart disease diagnosis Clinical knowledge–enhanced machine learning AI interpretability and trustworthiness Data-driven predictive model |
| url | http://www.sciencedirect.com/science/article/pii/S2666693622001700 |
| work_keys_str_mv | AT jamesmengmambbchir insidetheblackboxembeddingclinicalknowledgeindatadrivenmachinelearningforheartdiseasediagnosis AT ruimingxingmsc insidetheblackboxembeddingclinicalknowledgeindatadrivenmachinelearningforheartdiseasediagnosis |