Critical care units look after the most complex and unwell patients in hospitals, and acquire high volumes of physiological and laboratory measurements on their patients. These structured data need to be combined with information that is less well structured, such as diagnoses, examination findings, echocardiogram and imaging reports, in order to yield ‘big data’ insights that can improve clinical care. This PhD will address these challenges. The candidate will be able to use data from the NIHR Health Informatics Collaboration Critical Care theme, which contains data from over 20,000 patients from the past 10 years. However, it currently contains only structured data. Key information relevant to cardiovascular conditions such as examination findings and investigation reports are found only in free text clinical records.
The student will develop and implement methods to incorporate free text information from clinical notes, imaging reports and echocardiogram reports into the Critical Care HIC research platform and use the resulting resource to answer cardiovascular research questions such as:
- Is new left or right ventricular dysfunction in patients with sepsis associated with adverse prognosis?
- How are cardiac blood tests such as troponin values associated with outcomes in critical care patients with different underlying conditions (e.g. sepsis or heart failure)
- How is new onset atrial fibrillation, and its course and treatment, related to outcomes?