Artificial intelligence (AI) is a rapidly evolving technology that has seen increased use within numerous fields, including the health care sector. With the transition from paper-based to electronic health records (EHR) throughout Canada, as well as an increased collection of health data, there is an increased need for analyzing and interpreting data that can be used to make health care systems more useful and efficient.
As a breakthrough technology, AI will not only revolutionize they way health care organizations improve health care delivery, but it will also change the roles of Health Information Management (HIM) professionals. Therefore, it is important for HIM professionals to have a good understanding of AI, and to attain knowledge about how AI can be utilized as a tool for managing health care data.
Types of Artificial Intelligence #
This PPB discusses 3 types of AI that are relevant for HIM professionals:
- Machine Learning (ML)
- Neural Networks
- Natural Language Processing (NLP)
Machine Learning #
ML is a field of AI that primarily focuses on developing mathematical algorithms that enable AI systems to make logical predictions based on the relevant training data it was provided with. Essentially, ML systems make predictions by recognizing pattens and relationships between the data it was trained with, and the data it was provided with.
Neural Networks #
Neural networks are a subsection of ML systems that seek to imitate the behaviours of biological brains to make human-like intelligent decisions, by utilizing interconnected “nodes” with multiple layers akin to biological neurons in brain. Neural networks are often trained using multiple datasets to enable them to recognize, learn, and interpret abstract patterns for making logical decisions from the data it was provided with.
Natural Language Processing #
NLP is a field of AI that focuses on processing human-generated natural language for computer interpretation, and for extracting and transcribing information into natural language for human interpretation. Essentially, NLP gives AI systems the ability to understand, process, and generate human language in both written and spoken form
Key Takeaways #
This PPB discusses potential applications of AI in healthcare, including diagnostics, data analytics, and computer-assisted coding. One of the major challenges of AI implementation, especially in healthcare, concerns the design and testing of AI systems.
Because AI Systems depend on existing datasets for training, the accuracy of the AI systems’ prediction largely depends on the data it was trained with. This may contribute to bias, as the training datasets may not be representative of the whole population.
Therefore, as HIM professionals, we must ensure that datasets being used for training AI are relevant, diverse, and comprehensive to ensure minimal bias in the system.
Further Reading: Artificial Intelligence in Health Care (CHIMA, 2020)