In recent times there has been an increased availability of digital medical data. Much of the valuable data is stored as electronic health records (EHR), which contain a vast amount of information on patients’ medical history, data from treatments, diagnoses, prescriptions and test results. In this presentation, we discuss how the availability of this EHR data presents huge opportunities in areas such as development of data-driven precision medicines, more efficient healthcare, patients having increased control on their health outcome etc. We then talk through what the EHR data means for health insurance and how actuaries can use EHR data to produce more accurate predictive underwriting models. They can also better predict the expected cost of claims with the additional insights from EHR data and can integrate personalised digital health advice for customers. However, while EHR data open up a number of opportunities, they also present a number of challenges for analytics. Some of these challenges include the unstructured nature of data, high dimensionality of data, heterogeneity of data, quality of data etc. We discuss how different AI/ML techniques such as NLP can help address these challenges and the predictive analytics that follows. We also touch upon other broader challenges such as interoperability issues and how these can be addressed e.g. through the implementation of FHIR.