JMIR Publications

ABSTRACT

Background:

Background:

Artificial intelligence (AI) has been applied and showed promising results in various fields of medicine. It has the potential to facilitate shared decision making (SDM), which is the process in which patients and their healthcare providers collaborate to make a screening or treatment decisions based on evidence and patient values and preferences. However, there is no comprehensive mapping on how AI may be used for SDM.

Objective:

Objective:

We intended to identify and evaluate published studies that have tested or implemented AI for facilitating SDM as a first step towards a comprehensive mapping of the literature.

Methods:

Methods:

We performed a scoping review informed by the methodological framework proposed by Levac et al modifications to the original Arksey and O'Malley framework of a scoping review and the Joanna Briggs Institute scoping review framework. We reported our results following to PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analysis-Scoping Reviews) reporting guideline. At the identification stage, an information specialist performed a comprehensive search of six electronic databases from their inception to May 2021. Inclusion criteria were as following. Population: All populations who provided care or received care. Intervention: All AI interventions used to facilitate SDM. If the AI intervention was not used for the decision making point in SDM, it was excluded. Comparators: No restrictions. Outcome: Any outcome related to patients, healthcare providers and/or healthcare systems. Setting and study design: Studies in any health care setting. Only studies published in English were included. All studies using qualitative, quantitative or mixed method studies were included. Two reviewers independently performed the study selection process and extracted data on study characteristics, AI intervention characteristics, aspects of the AI intervention, how the AI interventions supported the decision-making step of SDM, and any reported outcomes related to patients, healthcare providers or healthcare systems. Disagreements were resolved by a third reviewer. A descriptive analysis was performed.

Results:

Results:

After the removal of duplicates, 894 documents were screened, and six peer-reviewed publications met our inclusion criteria. Two were conducted in North America, two in Europe, one in Australia and one in Asia. The majority was published after 2017. Three papers focused on primary care and three on secondary care. All papers used machine learning methods. Three included healthcare providers in the validation stage of the AI intervention, and one included both healthcare providers and patients in clinical validation, but none of the papers included healthcare providers or patients in design and development of AI system. All papers utilized AI to support SDM by providing clinical recommendations or predictions.

Conclusions:

Conclusion: The evidence on the use of AI in SDM is in its infancy. We found AI was supporting SDM in similar ways across the included papers. We observed poor reporting of the AI interventions, resulting in a lack of clarity about different aspects. Moreover, little effort was done to address the topics of explainability of AI interventions, as well as to include end users in the design and development of the interventions. Further efforts are required to strengthen and standardize use of AI in different steps of SDM as well as evaluate its impact within various decisions, populations, and settings.