Automatic Extraction and Collection of Data from Electronic Medical Records for theCancer Registry using Natural Language Processing and Ensembles of Machine LearningAlgorithms

Cancer is the leading cause of death in many countries around the world. International and national efforts to prevent and control this disease are being facilitated by cancer registries which collect data and provide vital statistics for decision makers concerning the incidence, survival and prevalence of cancers within the population. The collection of data in cancer registries is mostly conducted manually by trained registrars through, for instance, reviewing patient medical records and hospital discharge lists. This manual process is becoming unsustainable, slow and prone to error especially with the increasing trend of new cancer cases every year. Artificial Intelligence (AI) technologies based on Natural Language Processing (NLP) and Machine Learning (ML) have shown great advances in recent years in the area of automated classification and information extraction from text. In the medical field, these technologies have demonstrated the ability to accelerate data collection, improve data quality, and enhance evidence-based decision making at the point of care. Such technologies can be implemented and evaluated in comparison to domain-experts' decisions and considerable evidence from literature indicate the potential of these technologies to match human-level performance with great consistency. The general aim of this study is to implement and evaluate advanced AI algorithms to automate data collection and information consolidation for the national cancer registry in Oman.