Results
Eighty one of the 150 videos were excluded from the study due to exclusion criteria, and the remaining 69 videos were analyzed.
The descriptive statistics of the YouTube videos are presented in Table 1. The mean number of views for the videos on vertigo was 747,452.4 (min – max: 1,944 – 7,959,884, median: 166,229). The mean length of the videos was 357.2 seconds (min – max: 79 – 1,764 seconds, median: 266 seconds). The average number of days since the videos were uploaded was 1,447.1 days (min – max: 220 – 4,218 days, median: 1,288 days). The mean viewing rate of the videos was 44,928.2 (min – max: 153.7 – 348,506.3, median: 16,737.8). While the overall mean number of likes of the videos was 4,608.3 (min – max: 10 – 35,060, median: 1,698.5), the average number of dislikes was 222.7 (min – max: 1 – 1,772, median: 62). The mean number of comments made to the videos was 359.6 (min – max: 0 – 2,656, median: 137). The mean interaction index was 0.88 (min – max: 0.05 – 2.79, median: 0.73), while the mean video power index was 94.6 (min – max: 81 – 99.7, median: 95.6). The average number of views per day of the videos was 449.3 (min – max: 1.5 – 3,485.1, median: 167.4).
The mean total content score, GQS and modified DISCERN scores of the videos were 3.42 (min – max: 0 – 8), 2.48 (min – max: 1 – 5) and 2.09 (min – max: 0 – 5), respectively (Table 1).
It was seen that most of the videos were uploaded by healthcare professionals (n = 25, 36.2%), followed by other (n = 19, 27.5%), hospital / university (n = 18, 26.1%), commercial (n = 4, 5.8%) and layperson (n = 3, 4.3%). It was found that most of them were educational videos (n = 63, 91.3%), meaning videos that give information about vertigo, while a small part was testimonial videos (n = 6, 8.7%), meaning videos where people share their personal experiences (Table 2).
66.7% (n = 46) of the videos were included in the low content group, while 33.3% (n = 23) were included in the high content group (Table 3). The differences in sources of upload and video type between the high and low content videos were not statistically significant (p= 0.122, p= 0.168) (Table 3). Among the items used in content scoring, the most mentioned items in the videos were maneuvers (n = 47, 68.1%), treatment (n = 37, 53.6%), and symptoms (n = 38, 55.1%), respectively, while the least mentioned items were alarm symptoms (n = 6, 8.7%), prognosis (n = 8, 11.6%), and types of vertigo (n = 17, 24.6%), respectively.
In Table 4, high and low content videos are compared according to video characteristics and no significant difference was found between the two groups in terms of video characteristics (p> 0.05). GQS was found significantly higher in high content videos than low content videos (p< 0.001). Modified DISCERN was found significantly higher in high content videos than low content videos (p< 0.001) (Table 4).
In Table 5, educational and testimonial videos are compared according to video characteristics and no significant difference was found between the two groups (p> 0.05).
Spearman’s correlation analysis showed that significant positive correlations were found between GQS and total content score (r= 0.873, p< 0.001), between modified DISCERN and total content score (r= 0.883, p< 0.001) and between modified DISCERN and GQS (r= 0.900, p< 0.001) (Table 6). There was no statistically significant relationship between total content score and video demographics (p> 0.05). There was no statistically significant relationship between GQS and video demographics (p> 0.05). There was no statistically significant relationship between modified DISCERN and video demographics (p> 0.05) (Table 6).