An Artificial Intelligence-Powered Supportive Care Chatbot to Address the Supportive Care Needs of Young Adult Cancer Survivors

Primary outcomes

1. Acceptability of AI-powered supportive care chatbot

   Time frame: At end of intervention, assessed up to 12 weeks

   Acceptability will be supported if mean scores on the Acceptability E-Scale are ≥ 4 (on a 5-point scale). Will be described (i.e., means, medians, standard deviations, and ranges) at the post-intervention time point.

2. Demand of AI-powered supportive care chatbot

   Time frame: Up to 12 months

   Demand will be demonstrated by successful recruitment of the target sample (N=30) within 12 months.

3. Implementation of AI-powered supportive care chatbot

   Time frame: During intervention use, assessed up to 12 weeks

   Implementation will be assessed by engagement with the chatbot, defined as ≥ 70% of participants reporting at least one use per week during the initial 4-week period, rather than a fixed duration of use, given the self-directed nature of the intervention. Will be described (i.e., means, medians, standard deviations, and ranges) weekly. Given the pilot nature of the study, no hypothesis testing or formal comparisons will be conducted.

4. Retention

   Time frame: Up to 12 weeks

   Retention will be considered feasible if ≥ 80% of participants complete 4-week assessments, and ≥ 50% elect to continue to the optional extended use period.

5. Usability of AI-powered supportive care chatbot

   Time frame: At end of intervention, assessed up to 12 weeks

   Usability will be supported if mean System Usability Scale scores are ≥ 70, indicating acceptable usability. Will be described (i.e., means, medians, standard deviations, and ranges) at the post-intervention time point.

6. Patient Reported Outcomes Measurement Information System measure

   Time frame: At baseline, 4 weeks, and/or 12 weeks

   Will be summarized using descriptive statistics (e.g., means, medians, standard deviations, and ranges) at each time point. Changes over time (baseline, post-intervention, as applicable) will be examined descriptively.

7. Digital Health Literacy Scale

   Time frame: At baseline

   Will be summarized using descriptive statistics (e.g., means, medians, standard deviations, and ranges) at the baseline time point. The Digital Health Literacy Scale is a 0 to 12 point score (based on 3 items), with higher scores indicating greater digital health care literacy.

8. Interview themes and subthemes

   Time frame: At end of intervention, assessed up to 12 weeks

   The audio-recorded interviews will be transcribed verbatim by a professional transcription company and verified for accuracy by another study team member. The finalized transcripts will be imported into NVivo 12 (QSR International Pty Ltd). Inductive content analysis will be used to analyze the interview transcripts. Two study team members will review the transcripts and the interview guide to create an initial list of codes. Three transcripts will be independently coded using the initial codebook. After three interviews are coded, two study team members will meet to resolve any coding discrepancies and to revise the codebook further. The same process will be repeated after three more interviews are coded. After the codebook is finalized, one study team member will code the remaining interviews. Subsequently, the study team will meet as a group to review the transcripts in their entirety, making sense of the data and generating potential major themes and subthemes.