AI-derived information on indeterminate lung nodule: How do patients want information presented to engage in Shared Decision-making?

By: Brian Christopher M. Nocon, Tracey Carr, Navdeep Hansra, Paul Babyn, Candace Skrapek, & Kevin Belitski

Lung nodules, while commonly benign, are often indeterminate on imaging. Because lung nodules are associated with lung cancer, these require further investigations leading to increased cost and anxiety. An emerging solution to assist radiologist in diagnosing these indeterminate nodules is Artificial Intelligence. As health-care shifts towards shared decision making (SDM), new technologies should be implemented with a focus on preserving the SDM model.

This project aims to determine:

• Comfort of patients with use of Artificial Intelligence in determining likelihood of malignancy on their indeterminate lung nodule.
• How best to convey AI lung nodule diagnostic data to patients, with the overall goal of increasing the ability for a patient to engage within SDM and be empowered in their decision-making process.
• How do patient’s beliefs, values, and preferences influence their decisions in managing their indeterminate lung nodule.
• At what threshold are patients willing to undergo further investigation, more so invasive procedures to determine significance of their indeterminate lung nodule.

Methods: This is a patient-oriented, mixed qualitative and quantitative interview-based study. Sampling size will be based on convenience sampling. Patients who are residents of Saskatchewan with newly diagnosed indeterminate lung nodule between the months of May 2020 to December 2020 will be recruited for the study. One-on-one semi-structured interviews will be conducted using an interview guide co-developed together with Patient/family advisors. Interviews will be recorded and transcribed.

Results (Work in progress): Line-by-line analysis & coding of interview transcripts will be performed to identify patterns and commonalities using thematic analysis. Analyses will be conducted to explore potential differences (or lack thereof) in responses based on demographic variables including age, sex, worldview/cultural lens, and language. Interviews will be conducted until data saturation is achieved.

We believe that results of this study will help inform future implementation of AI in diagnosing indeterminate lung nodules.

Share this story