Article type
Abstract
Purpose: The Human Behaviour-Change Project (HBCP) is building an Artificial Intelligence system to scan the literature and extract key information and use this to build, update and interrogate a model of human behaviour change to determine: ‘What works, compared to what, how well, for whom, in what settings, for what behaviours, how long and why?’
Background: Behaviour change is essential to improve population health, disease self-management and clinical practice. Manual systems for evidence review and synthesis cannot keep up with the growth in the evidence base nor account of all the relevant features of interventions.
Method: We are: 1) developing an ontology of behaviour change interventions, populations, context, mechanisms of action and behaviours; 2) annotating published literature using the ontology to develop and train an automated system to extract key information from research reports; 3) developing and evaluating Machine Learning and automated Reasoning Systems to synthesise and interpret the evidence and make predictions; and, 4) developing and evaluating an online user interface to interrogate the knowledge base contained within the system.
Conclusions: The three main outputs will be: an ontology of behaviour-change interventions; an AI system capable of extracting and interpreting evidence from published literature and making predictions; and interfaces allowing users (human and machine) to access the knowledge base and answer specific questions about behaviour-change interventions.
Background: Behaviour change is essential to improve population health, disease self-management and clinical practice. Manual systems for evidence review and synthesis cannot keep up with the growth in the evidence base nor account of all the relevant features of interventions.
Method: We are: 1) developing an ontology of behaviour change interventions, populations, context, mechanisms of action and behaviours; 2) annotating published literature using the ontology to develop and train an automated system to extract key information from research reports; 3) developing and evaluating Machine Learning and automated Reasoning Systems to synthesise and interpret the evidence and make predictions; and, 4) developing and evaluating an online user interface to interrogate the knowledge base contained within the system.
Conclusions: The three main outputs will be: an ontology of behaviour-change interventions; an AI system capable of extracting and interpreting evidence from published literature and making predictions; and interfaces allowing users (human and machine) to access the knowledge base and answer specific questions about behaviour-change interventions.