Objective and Scope
Upon successful completion of the course, students will be knowledgeable about sources from which electronic health data can be extracted, the linkage and management of such data, as well as methods for constructing data sets ready for further statistical analysis. They will also be familiar with regulatory data governance frameworks.
Εκπαιδευτικοί στόχοι του μαθήματος
Μετά την επιτυχή ολοκλήρωση του μαθήματος οι φοιτητές θα είναι σε θέση να γνωρίζουν τις πηγές από τις οποίες μπορούν να εξαχθούν ηλεκτρονικά δεδομένα υγείας, τη διασύνδεση και διαχείριση τέτοιων δεδομένων καθώς και τις μεθόδους για τη δημιουργία συνόλων δεδομένων έτοιμων προς περαιτέρω στατιστική ανάλυση. Επίσης θα είναι εξοικειωμένοι με τα ρυθμιστικά πλαίσια διακυβέρνησης προσωπικών δεδομένων.
Course Content
- Introduction to the basic principles of precision medicine
- Sources and systems related to health data science in Greece and abroad (including Electronic Health Records)
- Application of core epidemiological study designs (prospective, clinical trials, and others) to electronic health data
- Basic concepts of controlled clinical terminologies and medical entities for recording diagnoses, medications, laboratory measurements, and procedures (e.g., ICD-10, SNOMED-CT, etc.)
- Ethical considerations and procedures, including data ownership, consent, algorithmic ethics, and relevant data governance mechanisms. Information governance, anonymization, privacy, and regulatory frameworks (e.g., GDPR)
- Quality issues in routinely collected data, such as coding/misclassification, probabilistic and deterministic data linkage, and missing values
- Construction of analytical data sets from Electronic Health Records using various study designs, creation and use of code lists, development and evaluation of phenotyping algorithms, measurement of exposures, outcomes, and covariates (including drug exposure data and ICD-10 classification)
- Data management (data wrangling and data cleaning)
- Major sources of bias in electronic health registries
- Examples of supervised and unsupervised machine learning and artificial intelligence for research using electronic health registries, including natural language processing and medical imaging
- Software engineering skills for accessing data, organizing analysis, and ensuring analyses are shareable and reproducible
- Mechanisms for ensuring secure access, including Trusted Research Environments
- Best practices for storing, managing, and analyzing large and complex data sets
Teaching Methods
The course content is delivered through 13 four-hour blended (theory–application) lectures, where theoretical concepts are taught alongside real-time practical computer applications. Students are assigned relevant coursework during the semester.
