Cancer
Conditions
Keywords
Communication, Social media, Cancer misinformation, Misinformation
Brief summary
The main purpose of this trial was to evaluate the impact of user interface designs on intervening with cancer misinformation on social media. The investigators conducted a randomized clinical trial with a sample of 294 US adults (ages 18-65) who previously had a cancer diagnosis or were a cancer caregiver. This registration record only covers the 5 timepoint trial under the same IRB number.
Detailed description
This 5 timepoint trial aimed to determine whether user interface designs encouraged intervening and reduced sharing of cancer misinformation on social media. The investigators enrolled 294 adults, ages 18-65, who had received a cancer diagnosis or were a cancer caregiver. Participants completed 5 online surveys, space approximately 2 business days apart. In the first survey, participant were randomized to 1 of 3 trial arms, described below. During each online survey, participants interacted with a simulated social media platform, branded as Invibe. Participants had passive exposure to stimuli cancer misinformation posts and cancer support posts in social media feeds. Participants also had forced exposure to stimuli posts in the online survey after interacting with Invibe. The investigators recorded social media behavior on Invibe and assessed self-reported measures in the online survey. This study was preregister on AsPredicted.com on March 24, 2025, prior to data collection and data analysis (https://aspredicted.org/te3su2.pdf).
Interventions
BEHAVIORALSocial cue prompt and policy
Cancer misinformation posts are shown with a prompt above the post caption that read, "This post has been flagged as false and potentially harmful by \[47 or 48\] people on Invibe. If you want to report, click the flag below." The social media launch screen had an overlaid policy for removing flagged posts: "We are committed to working with you and others on Invibe to reduce false and potentially harmful information. When a post is flagged by 50 people, we remove it until we verify the information. Please help us make the Invibe experience a good one for all."
BEHAVIORALStandard prompts
Cancer misinformation posts are shown with a prompt above the post caption that read, "This post has been flagged as false and potentially harmful. If you want to report, click the flag below."
Sponsors
UNC Lineberger Comprehensive Cancer Center
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
OTHER
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to 65 Years
Healthy volunteers
Yes
Inclusion criteria
* Are ages 18-65 * Received a cancer diagnosis or have been a cancer caregiver * Used social media * Were willing to complete all five study timepoints * Resided in the United States
Exclusion criteria
\- Were unable to scroll to see posts on Invibe in the screener survey
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Flagging | Up to 2 weeks | Measured with intervening behavior, defined as the flag icon selected. Response options were no (coded as 0) or yes (coded as 1). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Willingness to flag | Up to 2 weeks | Measured with 1 self-reported survey item: "How willing are you to flag this post?" Responses ranged from "not at all" (coded as 1) to "extremely" (5). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Willingness to like | Up to 2 weeks | Measured with 1 self-reported survey item: "How willing are you to like (heart) this post?" Responses ranged from "not at all" (coded as 1) to "extremely" (5). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Willingness to comment | Up to 2 weeks | Measured with 1 self-reported survey item: "How willing are you to comment on this post?" Responses ranged from "not at all" (coded as 1) to "extremely" (5). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Willingness to share | Up to 2 weeks | Measured with 1 self-reported survey item: "How likely are you to share or tell someone about this post?" Responses ranged from "not at all" (coded as 1) to "extremely" (5). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Comment valence | Up to 2 weeks | Measured with 1 self-reported survey item: "What type of comment would you write on this post?" Response options were "comment to agree" (share), "comment to disagree" (intervene), "other," or "I would not comment." Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Sharing recipient | Up to 2 weeks | Measured with 1 self-report survey item: "Who would you share about this post with?" Response options were "someone with cancer," "someone who is a cancer caregiver," "health-care provider," "other," or "no one." Assessed 6 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at first timepoint and last timepoint approximately 2 weeks apart. |
| Liking | Up to 2 weeks | Measured with intervening behavior, defined as the heart icon selected. Response options were no (coded as 0) or yes (coded as 1). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Past week intervening behavior | Up to 2 weeks | Measured with 1 self-reported survey item: "In the last 7 days, did you take action against false and potentially harmful cancer posts from social media? Select all that apply." Response options included muted accounts, flagged posts, disliked posts, commented to correct posts, chatted with others to disagree with posts information, encouraged someone to flag or report a posts, other or did not do anything. Assessed 1 time. Assessed at timepoint 4 approximately a week after the day of enrollment |
| Past week intervening platform | Up to 2 weeks | Measured with 1 self-reported survey item per reported interpersonal intervening behavior: "Who did you \[intervening behavior\]? Select all that apply." Response options were "someone with cancer," "someone who is a cancer caregiver," "health-care provider," "other," or "no one." Assessed up to 2 times. Assessed at timepoint 4 approximately a week after the day of enrollment. |
| Past week intervening topic | Up to 2 weeks | Measured with 1 self-reported survey item per reported intervening behavior: "What was the \[source for intervening\] about? Select all that apply." Response options were "to adopt a new mindset," "diet and nutrition advice," "physical activity advice," "advice to stop clinical care," "to seek treatment outside the US," or "other." Assessed up to 6 times. Assessed at timepoint 4 approximately a week after the day of enrollment. |
| Disliking | Up to 2 weeks | Measured with intervening behavior, defined as the thumbs down icon selected. Response options were no (coded as 0) or yes (coded as 1). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
| Sharing channel | Up to 2 weeks | Measured with 1 self-report survey item: "How would you share about this post?" Response options were "private message on social media," "text message," "in person," "repost on your social media," "other," and "none of these ways." Assessed 6 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at timepoint 1 and timepoint 5 approximately 2 weeks apart. |
| Willingness to dislike | Up to 2 weeks | Measured with 1 self-reported survey item: "How willing are you to dislike (thumbs down) this post?" Responses ranged from "not at all" (coded as 1) to "extremely" (5). Assessed 15 times. Assessed for 2 cancer misinformation posts and 1 cancer support post at each timepoint approximately 2 business days apart, starting on the day of enrollment. |
Countries
United States
Contacts
PRINCIPAL_INVESTIGATORAllison J Lazard, PhD
UNC Lineberger Comprehensive Cancer Center
Outcome results
None listed