Data collection and dataset management form the first component of a new concept in healthcare: Digital Twin. Digital Twin means a multi-modal modelling approach to data management, analysis, prediction, and interpretation at the patient level.
But data collection is a problem within healthcare, and it’s a messy one. Healthcare data is fragmented and ever-changing. And -for a good reason- researchers deal with patient privacy and security. Safe, orderly and robust data collection is a necessity for the future of healthcare. To realize this first component of Digital Twin, we’ll perform a case study with the following research question:
How to identify and integrate various data sources for patients with cardiovascular risk factors during their entire life course?
Holistic Patient files
This work package will result in Holistic Patient files, covering a given patient in great detail. These files contain patient-specific data, including demographic data, weight, height, vital signs, medical history, physical examination, medication and allergies, laboratory test results, imaging, and more, as far as available in the Hospital Information System (HIX). We will use the Observational Health and Science and Informatics (OHDSI) standard data model to integrate the data sources and transform the identified Erasmus MC sources.
Hypertension is the proof of principle for this study. A common condition and associated with the increased risk of other cardiovascular diseases. It is also the leading risk factor responsible for all-cause morbidity and mortality worldwide. The focus starts on women with hypertensive complications around pregnancy and adult and elderly patients with complex, treatment-resistant hypertension.
Model for managing patient data
We will use our multi-model dataset and advanced techniques, including artificial intelligence (AI) algorithms, to develop a model to manage hypertension. Novel analysis techniques will enable us to process a lot of patient information and model the complex relationships between them.
In turn, this will provide us with more accurate predictions on hypertension outcomes. We will then utilize these predictions in the decision-making process of hypertension management strategies. Digital Twin can improve healthcare and patients’ outcomes by offering more personalized healthcare in this context.
Personalized management of hypertension
Behavioural risk factors of patients with hypertension such as unhealthy diets, physical inactivity, consumption of tobacco and alcohol, and obesity are modifiable. Therefore, clinical guidelines include both pharmacotherapy and lifestyle advice. However, these recommendations are often general, while optimal management of hypertension may differ based on patients’ characteristics such as age, sex, race and diet.
In this study, we aim to develop a model using AI algorithms that can assist in defining the optimal personalized management strategy, including both pharmacotherapy and lifestyle modifications for patients with hypertension.
Acute chest pain is one of the symptoms of Acute Coronary Syndrome, which is a common condition leading to medical contact. Doctors use an electrocardiogram as the first test. However, this test is not always precise and can lead to under-or over-diagnosis.
The main aim of this project is to use AI algorithms to identify Acute Coronary Syndrome more precisely from the electrocardiogram. Precision in this process can improve the prognosis of patients who should receive treatment. And also the other way around: it could prevent complications in overdiagnosed patients and avoid the extra cost imposed on the health system.