Technologies and Strategies for Continuous Learning through Electronic Health Records Data

Achieving a comprehensive view of a patient’s health using data from Electronic Health Record systems requires the use of advanced analytics. However, effectively managing and curating this data requires carefully designed workflows. While digitization and standardization enable continuous health monitoring, issues such as missing data values and technical glitches can jeopardize data consistency and timeliness. On the other hand, the Efficiency in processing the large volume of data from disparate sources generated by the healthcare industry is critical. In this chapter, we try to provide an overview of how distributed computing and Artificial Intelligence can be used in the context of smart healthcare and big data in practical use cases, enabling insights to improve patient care. In addition, we propose a workflow for developing prognostic models that uses the SMART BEAR infrastructure and leverages the capabilities of the Big Data Analytics engine to standardize and harmonize data. Our workflow improves data quality by evaluating different imputation algorithms and selecting the one that preserves the distribution and correlation of features similar to the original data. We applied this workflow to a subset of data in the SMART BEAR repository and evaluated its impact on predicting future health conditions, such as cardiovascular disease and mild depression. We also explored the potential for model validation by clinicians in the SMART BEAR project, the transfer of subsequent actions within the decision support system, and the estimation of the required number of data points.