Evaluating YouTube Videos on Facelift Surgery for Facial Rejuvenation as a Resource for Patients

Introduction

The Internet and social media platforms including YouTube, Facebook, and Instagram have become potential resources for patients searching for healthcare information. ¹ According to the Pew Research Center, more than 9 in 10 Americans use the internet as of 2019. ² The 2020 Health Information National Trends Survey (HINTS) discovered that approximately 72.1% of Americans use the internet to search for healthcare-related information. ³ YouTube is the second most popular website on the internet overall, and the most popular video-sharing platform on the internet, with an estimated 2 billion active monthly users watching over 1 billion hours of video content daily.^4,5 Many medical specialties have started to acknowledge and utilize YouTube as a repository of informational videos for patients, and YouTube as a resource for patients and clinicians has become of particular interest to otolaryngology and facial plastic surgery.^{2,6 -9}

As people age, the soft tissues of the face descend with gravity, the muscles atrophy, and the skin becomes more wrinkled and less elastic. ⁶ Rhytidectomy, also known colloquially as facelift surgery, is the workhorse of facial rejuvenation and addresses the lower one-third of the face by fixating and raising soft tissues of the neck and jawline. ⁷ The popularity of face-lifts is steadily increasing. ⁸ The American Society for Aesthetic Plastic Surgery statistics estimate that approximately 100 000 face-lifts were performed in 1999 in the United States, however in 2008, that number had increased by 30%, to approximately 130 000 face-lifts. ⁹

Facial plastic surgery is a highly esthetics-driven specialty, and many clinicians have begun utilizing social media to advertise procedure results, informally regarded as “before-and-after” pictures.^10,11 However, there is an alarmingly small amount of credible information posted to the top social media websites. In 2018, out of the top plastic surgery-related posts on Instagram, only 17.8% were made by American Society of Plastic Surgeons eligible board-certified plastic surgeons with no data currently available on the percentage provided by American Academy of Facial and Reconstructive Surgery members. ¹² Some accounts also labeled themselves as “cosmetic surgeons,” despite not being board-certified plastic surgeons. ¹³

Despite the growing popularity of rhytidectomy as a solution for facial rejuvenation and the increasing popularity of social media as a source of healthcare information for patients, there are no studies investigating the quality or credibility of information in YouTube videos on rhytidectomies. Facial plastic surgeons must understand the quality of online content related to this procedure. Furthermore, patient education and setting realistic goals is essential, especially in the field of facial plastic surgery where unrealistic expectations are a negative predictor of patient satisfaction. ¹⁴ This study aims to quantitatively assess if YouTube videos serve as a quality, comprehensive source of information for rhytidectomies for facial rejuvenation for patients.

Materials and Methods

Results

The first 50 video results were collected for each of the 4 search terms (n = 200). Subsequently, 100 (50%) duplicates were excluded. Of the remaining 100 videos, 57 met inclusion criteria and 43 were excluded. Of the 43 excluded, 3 were lecture format videos intended for medical professionals, 12 were operating room (OR) technique specific videos, 1 was non-English, 11 were irrelevant, 3 were videos without both audio and subtitles, and 13 were testimonials (Figure 1). Thirty-one unique uploaders were identified. Ten of these sources had more than one video uploaded and included for analysis (mean 1.81, standard deviation (SD) 1.84). Interrater reliability grading was “good,” with ICCs > 0.75 across all 3 grading systems (Table 1). ³

OPEN IN VIEWER

Table 1.

Intraclass Correlation Coefficients for All 3 Grading Systems According to Single and Average Measure. ²²

Grading system	Single measure	Average measure
GQS	0.868 ± 0.106	0.929 ± 0.068
Modified DISCERN	0.942 ± 0.036	0.970 ± 0.020
JAMA Benchmark	0.824 ± 0.159	0.904 ± 0.121

OPEN IN VIEWER

Figure 1.

Video selection process. Fifty-seven videos met inclusion criteria and subsequently underwent quality assessment.

Video Quality Scores Outcomes

For all videos the mean AVQ score was 2.94 (SD 0.22), the mean GQS 2.47 (SD 0.84), the mean modified DISCERN score was 11.50 (SD 2.46), and the mean JAMA Benchmark score was 0.85 (SD 0.36) (Table 2). There was a significant positive correlation between Total DISCERN score compared to GQS (P = .000, CC = .958) and DISCERN Bias score (P = .000, CC = .948). However, AVQ and JAMA grading systems were not significantly correlated with any other primary outcome.

OPEN IN VIEWER

Table 2.

Mean Scores, Score Ranges, and Percent of Highest Scores Achieved Across All Videos for Each Assessment.

Grading system	Mean score	Standard deviation	Possible score range	Percent of highest possible score
AVQ	2.94	0.22	0-3	98.00
GQS	2.47	0.84	1-5	49.40
Modified DISCERN	11.50	2.46	5-25	46.00
JAMA Benchmark	0.85	0.36	0-4	21.30

Video Quality Scores Versus Video Performance Outcomes

There were no significant differences in score across all grading systems when comparing videos above and below the median value for view (Table 3). There were significant correlations between Discern Bias score compared to video duration (P = .014, CC = 0.569); Discern Total score compared to video duration (P = .021, CC = 0.539); and JAMA score compared to likes (P = .045, CC = 0.492) and VPI (P = .046, CC = 0.489). There were no significant correlations between primary assessment outcomes and performance metrics.

OPEN IN VIEWER

Table 3.

Association of Video Quality Scores by Views Above or Below the Median Value.

Grading system		Mean score	Standard deviation	P-value
AVQ	>8415 views	2.97	0.19	.34
	<8415 views	2.91	.24
GQS	>8415 views	2.33	0.84	.21
	<8415 views	2.61	.84
Modified DISCERN	>8415 views	11.38	2.37	.73
	<8415 views	11.61	2.55
JAMA Benchmark	>8415 views	0.84	0.38	.90
	<8415 views	0.86	.33

Video Views Versus Video Performance Outcomes

The median number of views for each video was 8415 (range 28-1 699 204) views. There were no significant differences between GSQ (P = .879), Total DISCERN (P = .903), JAMA (P = .472), and AVQ (P = .069) scores when comparing videos above (n = 28) and below (n = 27) the median number of views. Videos with views over the median were significantly associated with higher secondary outcome measures including dislikes (mean = 87, SD = 187, P = .016), duration (mean = 583.0, SD = 526.0, P = .001), comments (mean = 222.0, SD = 453.0, P = .012), views (mean = 197 069.0, SD = 403 727.0, P = .012), view ratio (mean = 260.0, SD = 659.0, P = .041), and VPI (mean = 251.0, SD = 642.0, P = .042). However, there was no significant difference between videos above and below the median view counts regarding likes (mean = 2887.0, SD = 8354.0, P = .072), like ratio (mean = 93.0, SD = 7.0, P = .476), and days since upload (mean = 1557.5, SD = 973.2, P = .413) (Table 4).

OPEN IN VIEWER

Table 4.

Association Between Video Views and Secondary Characteristics.

Characteristic	Above Median Mean (SD)	Below Median Mean (SD)	Significance
Likes	2887.0 (8354.0)	37.0 (47.0)	P = .072
Dislikes	87.0 (187.0)	1.0 (1.3)	P = .016
Like Ratio	93.0 (7.0)	90.0 (25.0)	P = .476
Duration (seconds)	583.0 (526.0)	222.0 (120.0)	P = .001
Comments	222.0 (453.0)	3.6 (8.5)	P = .012
Views	197 069.0 (403 727.0)	2452.0 (2204.0)	P = .012
View ratio	260.0 (659.0)	4.3 (7.8)	P = .041
VPI	251.0 (642.0)	4.1 (7.5)	P = .042
Days Since Upload	1557.5 (973.2)	1337.9 (1034.2)	P = .413

Discussion

This study investigated the current landscape of facelift education videos available to patients. Videos scored poorly across all quality measurements. For example, across all videos the mean modified DISCERN score was 11.50 (SD 2.46) which corresponds to video content not meeting any quality criteria (if all videos did not meet any of the quality content criteria, then each would receive a score of “no” which equals score of 1 × 5 = 5/25) to only partially meeting quality criteria (if all videos only partially met each quality content criteria then each video would score “partially” which equals score of 3 × 5 = 15/25). Previous studies demonstrate similar findings where preclinical videos were scored and performed poorly according to the same methods used in this investigation.^{3,15,23 -25} Videos with more views and more interaction (likes and dislikes) were more likely to have positive comments from viewers. Additionally, videos of longer duration correlated with more positive comments and were associated with more views. This might be secondary to longer videos taking the time to explain surgery specifics compared to shorter videos. This explanation can be corroborated by the positive association of video duration with Total DISCERN scores. Interestingly, higher video views were not associated with higher quality assessment scores. This finding might indicate that videos are of poor quality regardless of popularity as measured in views. However, videos with more views were significantly associated with interactions (dislikes, comments, view ration, and VPI). This might be secondary to the nature of the YouTube algorithm where videos are more likely to be recommended to an individual based on the watch times and interactions of similar viewers. ²⁶

The current landscape of facelift videos available on YouTube are highly biased to positive outcomes and do not adequately explain the risks, alternatives, or postoperative courses for the procedures offered. Videos that properly explain the indications, benefits, risks, and postoperative courses in a nonbiased manner have the opportunity to set realistic expectations and better guide patient decision-making. Additionally, videos that are patient centered and address common questions have the potential to decrease the time required counseling patients in the office by addressing concerns beforehand. ² However, no video can replace a face-to-face consultation with a surgeon, rather videos should complement the clinical visit by providing a starting point that both the patient and surgeon can reference upon their visit.

This study was limited by only reviewing the first 50 videos for each search term and limiting search terms to 4. Additionally, the words and syntax composing each term may have affected the video returned by each query. However, 50% (n = 100) of these videos were duplicates indicating considerable overlap with this sample size. These results may also have some degree of bias due to the subjective nature of assessing videos. However, the ICC for all assessment criteria was good, (>0.75) potentially reducing individual subjective bias. Additionally, while some individual videos may not have contained indications, alternatives, complications, etc., we recognize that uploaders may have chosen to provide singular videos dedicated to each of these topics which may have subsequently resulted in low quality content scores. Regardless, the authors believe this practice to be inferior to providing a single, comprehensive video as there is no guarantee viewers will watch or find the other provided videos. Future studies might investigate patient understanding and satisfaction when provided with educational videos immediately before their initial content.

Conclusion

Patients are increasingly using online sources to learn about medical conditions/procedures. Despite many videos being available, the current quality of videos related to facelift surgery is poor. Filling this gap with high quality videos has the potential to benefit both patients and physicians alike.

Supplemental Material

Supplemental Material

Supplemental Material

Supplemental Material