Analysis of the quality and credibility of health-related content on YouTube: cross-sectional study

Abstract

Introduction. Social media plays a massive role in informing people and promoting valuable, poor-quality content. The openness of social media to all users makes users vulnerable to viewing low-quality, misleading videos. In today's world, the number of users using information on social media is growing, so assessing the quality and credibility of the material provided is important. This study aims to thoroughly analyze the quality and credibility of YouTube health-related content (HRC) impacting public health. Methods. Data were collected through the YouTube API using key queries: 'vaccination,' "COVID-19," "folk remedies," "health," "mental health," and' nutrition.' One hundred seventy-seven videos with more than 10,000+ views and duration between 3 and 30 minutes were selected for analysis. Quality was assessed using the Global Quality Scale (GQS), and reliability was assessed using a modified DISCERN scale. All data were analyzed using SPSS version 26; the Kappa coefficient was used to assess consistency. Results. The final analysis included 162 videos, of which 74.1% were helpful, and 25.9% contained misleading information. Most of the videos (27.8%; n = 45) were classified as "Medical educational," followed by "Educational" (17.9%; n = 29) and "News" (15.4%; n = 25). The least represented categories were "Entertainment" (5.6%; n = 9) and "Other" (4.9%; n = 8). Helpful videos showed higher scores on the GQS and DISCERN scales (P < 0.001). The GQS scores of the valuable videos were significantly higher than the misleading videos, with a median (min-max) score of 4 (4–5) (P < 0.001). The DISCERN found the scores of the valuable videos to be significantly higher (P < 0.001). The median number of views for misleading videos was 591,060 (13,370–107,969,406), compared to 353,567 (range: 10,314–13,557,501) for valuable videos. Videos with mostly misleading content had many views, and the difference between the two comparable categories of groups was not statistically significant (P>0.05). Daily views and likes in valuable videos were lower than in misleading ones; the difference was significant (P < 0.05). The likes rate was lower in the valuable videos, but the difference was insignificant (P = 0.778). The average number of views of misleading videos is much higher, although the quality and credibility are lower. Conclusion. YouTube significantly impacts the dissemination of medical information, but the platform's accessibility also facilitates the dissemination of inaccurate content. The results of our study emphasized the need to create a system of monitoring and control over the quality of medical information in the YouTube social network, as well as the involvement of medical professionals in creating educational medical content. The development of information verification mechanisms will help to reduce the risk of users making wrong decisions based on unreliable video content.