| Criteria for judging whether large data sets are suitable for use in high-value clinical AI applications |
Need for understanding the use of algorithms in machine learning Limited access to informational resources for current students enrolled in clinical degrees Useful and relevant topic for research training and participation
|
| General ethical implications |
|
| Machine learning—neural networks and deep learning |
Specific concepts to include the applications, implications, consequences, and limitations of using machine learning Focus on future needs Targeted and focused development of a group of individuals rather than the whole workforce
|
| Specific patient engagement and adherence applications |
Specialized training needs will be required depending on the care pathway, specific apps, and progress of these technologies Having established governance structures around value-based health care to complement education Changes to cognitive load within overall clinical workflow Tailoring levels of educational uptake for different disciplines “Models of care using these tools need to be built and clinically governed”
|
| Changes to cognitive load within overall clinical workflow |
|
| Change management processes when AI is integrated within clinical workflows |
|
| Human-machine interaction in clinical settings |
|
| Specific diagnosis and treatment applications of AI |
Might not be relevant to certain workforce roles End users of specific diagnosis and treatment applications of AI might not require in-depth specialist training and education This area will need to evolve to meet future needs (ie, development of standards and clinical governance regarding skill competencies)
|
| Specific administrative applications |
Area of interest given that a workforce competent in specific administrative applications would bring about productivity and clinical quality benefits Digital and ICTb specialist workforce will require knowledge of specific administrative applications; the health care workforce could contribute by providing clinical input in this area AI and process automation in the area of change management Inclusion of risk management strategies in education and training “Knowing about the very many near misses is more important for the purposes of refining AIMLc than critical incidents alone”
|
| Rule-based expert systems |
Area of great potential and benefit (ie, reduction in cognitive load errors in emergency and intensive care settings) Further analysis required to understand the health workforce’s receptivity toward using rule-based expert systems and the implications for clinical practice in the next decade Specific health knowledge management applications
|
| Specific health knowledge management applications |
Limited access to resources (ie, databases) Participation in research projects was a way to promote learning
|
| Physical robots |
Highly relevant to medical professionals, nursing, aged care, and allied health A lack of clarity around the implications of using physical robots in clinical practice (ie, concrete examples would be required to understand how health workforce job roles might interact with physical robots) The need for training to be value-adding to ensure that physical robots improve and do not hinder health care workflow “Doctors probably learn more about robots from their kids’ toys than from their training.”
|
| NLPd
|
Lack of access to real-world health data to teach learners about using algorithms; limited number of education opportunities and digital health literacy resources to support learning The clinical workforce might only require a general understanding of how NLP tools work, its applications, limitations, and consequences of use in health care Expertise of NLP specialists could be leveraged
|
| RPAe
|
Important for the digital and ICT workforce to acquire skills in this area to support the health workforce in automating processes Health workforce could benefit from greater knowledge of ways to identify opportunities to apply RPA.
|
