A critical appraisal of YouTube as a source of information for PET/CT

Objectives PET/CT is a rapidly growing nuclear imaging modality. Social media has become a widely-accessed source of healthcare information with potential to influence patients’ perspectives. However, there has been little critical appraisal of social media content imaging modalities. The purpose of this study was to assess the quality of information pertaining to PET/CT on YouTube, an archetypical and popular social media website.

Methods For this IRB-exempted study, a new YouTube account was created to avoid influence from prior internet search activity. From this account, YouTube was searched using the term “PET/CT.” First 200 videos (10 pages of search results) ranked by relevance (default option on YouTube, which uses a complex algorithm based on view count, rating, etc.) were screened for information on PET/CT. 166 (82.5%) videos had to be excluded: 73 (36.5%) in a non-English language, 59 (29.5%) advertorial, 30 (15.0%) for information not pertaining to PET/CT or pertaining to a highly specific clinical indication (e.g. PET/CT in lung cancer), and 2 (1.0%) duplicates and, 1 (0.6%) video uploaded by our institution. Video length, uploading source, upload date, likes, dislikes, and total views were recorded for the 34 included videos. Two board-certified Nuclear Medicine Technologists assessed each video for production quality, content accuracy, video position towards PET/CT, intended audience, main content theme, and global quality score on a 5-point scale. A validation set of 4 videos (excluded from final study sample) was used to train reviewers with these criteria. Discrepancies were resolved through adjudication by a board-certified Nuclear Medicine Radiologist.

Results Majority of the included 34 videos were short clips less than 4 minutes (65%), greater than 2-years old (58%), and uploaded by medical websites/vendors (42%). The study sample of 31 videos generated 328,414 total views (average: 365 views/month) and 539 net likes (total likes minus total dislikes). Most videos were rated as professional in production quality (84%) and as being intended for lay public (81%). All the included 34 videos were accurate or mostly accurate (100%) in terms of content, and exhibited either a neutral or favorable position towards PET/CT (100%). The most recurrent main content themes were introductory information and common uses of PET/CT and/or patient preparation and experience (98%). Overall, 58% of videos were deemed to be of good-to-excellent global quality. Professional quality and longer length videos were more often rated as good-to-excellent in global quality (p=0.0002 and 0.0149 respectively).

Conclusions An overwhelming majority of videos that appear in a YouTube search result using ‘PET/CT’ are extraneous due to either being advertorial or being in a non-English language. Although the remaining small number of videos that specifically focused on PET/CT had accurate content, they tended to be short-snippets, older than 2-years, predominantly uploaded by health websites or medical vendors and of variable overall utility. Therefore, there is need for active engagement by radiologists and professional organizations on social media to circulate updated information about advanced imaging modalities such as PET/CT. This may empower patients to be informed participants in decisions regarding medical imaging.