Short answer

Design AI healthcare solutions that clearly delineate the AI's role as a supportive tool, ensuring human oversight and providing accessible explanations of its functionality and limitations to build trust and acceptance.

Field: User-Centred Design
Source: The Lancet Digital Health (2021)
Method: Mixed Methods Systematic Review
Evidence: Moderate effect

Despite general positivity towards Artificial Intelligence (AI) in healthcare, patients and the public express significant reservations, prioritizing human supervision and clear understanding of AI's role. This user-centred design research insight is drawn from a 2021 study published in The Lancet Digital Health. Using Mixed methods systematic review, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Design AI healthcare solutions that clearly delineate the AI's role as a supportive tool, ensuring human oversight and providing accessible explanations of its functionality and limitations to build trust and acceptance.

Study

User-Centred DesignHigh ImpactModerate effect

AI in Healthcare: Public Preference for Human Oversight and Clear Communication

The Lancet Digital Health · 2021

Key Findings

01Patients and the general public generally hold positive attitudes towards AI in healthcare.
02Significant reservations exist regarding AI, with a strong preference for human supervision.
03Concerns revolve around AI's conceptual understanding, acceptability, relationship with humans, development, implementation, strengths, benefits, weaknesses, and risks.

Application

Design takeaway

How to apply

When designing any AI-powered medical device or software, conduct user research specifically focused on understanding patient and clinician perceptions of AI's role, potential risks, and desired levels of human involvement.

Project actions

01When designing a product that uses AI, think about how you will explain its function to the user.
02Consider how your design will allow for human oversight or intervention.

Method & Evidence

AimWhat are the attitudes of patients and the general public towards clinical Artificial Intelligence, and what are their key concerns and preferences regarding its development and implementation?

MethodMixed Methods Systematic Review

ProcedureA systematic review of biomedical and computational databases was conducted to identify original research articles (quantitative, qualitative, and mixed methods) published between January 1, 2000, and September 28, 2020, focusing on patient and general public attitudes towards clinical AI. 2590 articles were screened, and 23 met the inclusion criteria.

ContextHealthcare, Clinical Artificial Intelligence

Variables

IV["Type of AI (hypothetical vs. realized)","AI application area"]

DV["Patient/public attitudes towards AI","Preferences for human supervision","Concerns and reservations"]

CV["Study design","Methodology of included studies"]

Strengths & Limitations

Strengths

+Comprehensive systematic review covering a decade of research.
+Mixed-methods approach captures diverse user perspectives.

Limitations

The studies reviewed often focused on hypothetical AI, so real-world user reactions might differ. The quality of the original studies varied.

Reliability & validity

The reliability of the review is enhanced by its systematic methodology and broad search strategy. Validity is supported by the inclusion of mixed methods, but is somewhat limited by the heterogeneity of the included studies and potential biases.

Think critically

How might the findings change if the AI was directly interacting with patients rather than solely assisting clinicians?

Design Principles

"Human-AI collaboration should be designed with transparency, explainability, and user control as core tenets."

For designers and engineers developing AI-driven healthcare solutions, understanding these user sentiments is crucial for successful adoption. Ignoring these concerns can lead to resistance, distrust, and ultimately, the failure of potentially beneficial technologies.

What This Means for Your Design

People generally like the idea of AI helping doctors, but they want to make sure a human doctor is still in charge and that they understand how the AI is helping.

How to use in your project

1.Use this research to justify the importance of user research and user-centred design principles in your design project, particularly when dealing with complex or sensitive technologies like AI.

Add to My Project

Quick Cite

Paragraph starter

This systematic review indicates that while there is a general acceptance of AI in healthcare, user-centred design must prioritize human oversight and transparent communication. Findings suggest that designs should clearly articulate the AI's supportive role and provide accessible explanations to address user reservations, ensuring a collaborative and trustworthy human-AI interaction.

Source

The Lancet Digital Health

Patient and general public attitudes towards clinical artificial intelligence: a mixed methods systematic review

journal · 2021

View source

Questions About This Research

What does the research say about ai in healthcare: public preference for human oversight and clear communication?: Design AI healthcare solutions that clearly delineate the AI's role as a supportive tool, ensuring human oversight and providing accessible explanations of its functionality and limitations to build trust and acceptance. Evidence: The Lancet Digital Health (2021).
Why does "AI in Healthcare: Public Preference for Human Oversight and Clear Communication" matter for design?: For designers and engineers developing AI-driven healthcare solutions, understanding these user sentiments is crucial for successful adoption. Ignoring these concerns can lead to resistance, distrust, and ultimately, the failure of potentially beneficial technologies.
How can designers apply this research?: Design AI healthcare solutions that clearly delineate the AI's role as a supportive tool, ensuring human oversight and providing accessible explanations of its functionality and limitations to build trust and acceptance.
What were the main findings?: Patients and the general public generally hold positive attitudes towards AI in healthcare.. Significant reservations exist regarding AI, with a strong preference for human supervision.. Concerns revolve around AI's conceptual understanding, acceptability, relationship with humans, development, implementation, strengths, benefits, weaknesses, and risks.
What research method was used?: Mixed Methods Systematic Review.
How strong is the evidence?: Evidence strength is rated Moderate effect, based on a 2021 journal from The Lancet Digital Health.
What should I do differently in my next project?: When designing any AI-powered medical device or software, conduct user research specifically focused on understanding patient and clinician perceptions of AI's role, potential risks, and desired levels of human involvement.
What are the limitations?: Heterogeneity in study populations, AI types, and study designs; potential selection bias in included studies; a majority of studies assessed hypothetical AI rather than realized tools.