1. Digital data processing and analytics to support decision making in cardiac care
- Author
-
Iftikhar, Aleeha, Bond, Raymond, McGilligan, Victoria, Peace, Aaron, and Leslie, Stephen
- Subjects
Digital forms ,Health care ,Usability evaluation ,PPCI ,Data analytics ,Digital data processing - Abstract
Cardiovascular disease (CVD) is the main cause of death globally. This PhD was carried out to improve clinical decision making in cardiac care by enhancing how data is collected and analysed. Firstly, the PhD includes the development and assessment of the usability of three different interactive digital health forms, namely, 1) a single page digital form, 2) a multi-page digital form, and 3) a conversational digital form (a chatbot). After comparing these three digital form designs, it was discovered that healthcare professionals perform better when using a single page digital form (p(HCI)/form design suggestions for gathering high quality data that can be used to facilitate reliable real world data analytics, which could, in turn, provide new useful and actionable insights to improve clinical decision making. Moreover, given that data science/data analytics is an emerging area to improve patient care, this PhD carried out series of analyses to elicit beneficial insights from analysing referral datasets and pathways. These analyses contain a series of analyses including time series analysis, supervised and unsupervised machine learning and process mining, all applied to real-world data (datasets of patients who were referred to the primary percutaneous coronary intervention (PPCI) service/Cath-Lab for cardiac reperfusion therapy). The primary findings include that time series analysis of all the patient's data exhibit various fluctuations over time. Furthermore, cluster analysis was used to discover patient archetypes as well as new taxonomy for naming archetypical patients. Also, using the PPCI patient referral datasets, 30 days and 1-year mortality was predicted using various ML algorithms. This PhD also illustrates the use of process mining methods for detecting patient pathway patterns in cardiac care. Secondary data also allowed the PhD to investigate how computers and humans make clinical decisions when interpreting electrocardiograms (ECGs).
- Published
- 2022