How useful are YouTube videos on labial frenectomy? A quality review

Abstract

Background

The labial frenulum plays a crucial role in oral health, and its surgical removal, known as frenectomy, is often necessary for various dental and orthodontic conditions. This study evaluated the quality and educational value of YouTube videos on labial frenectomy, highlighting potential misinformation risks in online health resources.

Methods

A cross-sectional study was conducted via the search term “labial frenectomy” on YouTube, and 68 relevant videos were identified on the basis of predefined criteria. Content quality and credibility were evaluated via the total content score (TCS), video information and quality index (VIQI), modified DISCERN, global quality scale (GQS), Journal of the American Medical Association (JAMA), and Health on the Net Code (HONcode). The data were analyzed with the Kolmogorov‒Smirnov, Kruskal‒Wallis, and Bonferroni correction. Numerical variables were examined via Spearman’s rho and point‒biserial correlation, whereas categorical variables were assessed via chi-square and Fisher’s exact tests.

Results

TCS was positively correlated with GQS (r = 0.392; p < 0.001), VIQI (r = 0.379; p < 0.001), and modified DISCERN (r = 0.396; p < 0.001), indicating that higher-quality videos received better scores. The GQS was strongly correlated with the VIQI (r = 0.747), modified DISCERN (r = 0.711), and HONcode (r = 0.721; p < 0.001). A significant association was observed between TCS and presenter gender (p = 0.03), with higher-quality content more frequently found in videos featuring female presenters.

Conclusions

YouTube videos on labial frenectomy generally provide low- to moderate-quality content and often lack reliability and medical accuracy. While they offer basic educational value, they should not replace professional guidance. Healthcare professionals should direct patients toward expert-reviewed, evidence-based sources to ensure comprehensive and accurate medical information.

Background

The labial frenulum is a fold of mucosa that connects the inner lip to the gingiva [1]. When abnormally positioned near the gingival margin, it may impede oral hygiene, cause gingival tension, and contribute to periodontal issues [2]. Labial frenectomy is commonly indicated in such cases and may also be performed to address diastema, limited lip mobility, or prosthetic and orthodontic concerns [3].

Labial frenectomy can be performed using different techniques, including traditional scalpel methods and newer options, such as electrosurgery or lasers. Each method has its own advantages and limitations, including intraoperative bleeding, postoperative infection, scarring, and in some cases, recurrence of the frenulum [4, 5]. The selection of an appropriate technique is typically guided by multiple factors, such as the patient’s age, extent of the frenulum anomaly, and the surgeon’s level of expertise [6].

In the current digital landscape, patients have unprecedented access to medical information through social media, including content on labial frenectomy procedures. This occurs across various stages of the treatment journey, such as pre-treatment decision-making (e.g., assessing surgical necessity), intra-treatment expectations (e.g., anticipated pain or procedural visuals), and post-operative recovery (e.g., wound care and dietary guidance). The internet, particularly platforms like YouTube, has become a prominent source of health information due to its accessibility and visual appeal. However, the quality and reliability of such content vary significantly, potentially shaping patient perceptions in both beneficial and harmful ways [7–9].

Understanding the indications, procedures, and potential risks of labial frenectomy is crucial for both clinicians and patients. As online platforms increasingly serve as primary sources of medical information, evaluating the accuracy and educational quality of YouTube content on labial frenectomy is essential. This study aimed to assess the reliability, accuracy, and educational value of these videos. To achieve this, we applied multiple validated assessment tools, including the modified DISCERN instrument, JAMA criteria, Global Quality Score (GQS), Video Information and Quality Index (VIQI), Health On the Net Code (HONcode), and a custom Total Content Score (TCS) derived from the relevant literature. We hypothesized that a significant number of these videos would lack sufficient academic rigor and present incomplete or poor-quality educational content.

Methods

This retrospective observational study analyzed publicly available YouTube videos without involving human participants or identifiable data; hence, ethical approval was not required. The study adhered to the Declaration of Helsinki and institutional ethics guidelines. Two licensed dentists independently selected, screened, and evaluated the videos. The data were recorded in Microsoft Excel (Redmond, WA, USA) and remainded accessible via publicly available YouTube links.

This cross-sectional study systematically evaluated YouTube videos on labial frenectomy. Based on a Google Trends analysis, “labial frenectomy” was identified as the most searched term and used as the primary search query.

On November 1, 2024, a YouTube search was conducted using the keyword “labial frenectomy,” selected via Google Trends as the most relevant term. To reduce personalization bias, the search was performed in incognito mode after clearing the browser history, cache, and cookies, following previousq recommendations [10]. The “Sort by relevance” filter was applied to the search. The first 250 videos were recorded and archived for this study. The inclusion criteria were English-language videos lasting 3–20 min. Videos shorter than 3 min were excluded to ensure sufficient content for quality assessment, as supported by prior studies [11, 12]. Videos were excluded if they were duplicates (n = 10), not related to subject (n = 4), had no audio/visual (n = 9), were not in English (n = 8), contained unintelligible speech (n = 2), had no explanation (n = 24), were shorter than 3 min (n = 120), or longer than 20 min (n = 5) (Fig. 1). Following the application of these criteria, 68 videos exhibiting appropriate characteristics were selected for further evaluation.

Fig. 1.

Flowchart of the videos included from YouTube

Two independent researchers, both licensed dentists, evaluated the videos using pre-defined scoring tools. Each video was assessed individually. Discrepancies were resolved through discussion and reference to the relevant literature, including clinical guidelines and peer-reviewed studies. Specifically, literature sources were used to clarify the relevance and importance of certain content topics, particularly in scoring items in the custom-developed TCS. When disagreements occurred regarding the presence or adequacy of a topic (e.g., indications, surgical steps, or postoperative care), relevant guidelines, peer-reviewed articles, or textbook references were consulted to guide the scoring decisions and ensure consistency.

All videos were evaluated under standardized conditions to minimize bias and maintain consistency. Both researchers used the same 15.6-inch high-resolution laptop (1920 × 1080 pixels) in a uniformly lit environment and wore high-quality headphones to ensure audio clarity. To avoid fatigue and ensure accuracy, evaluations were conducted in multiple sessions, each lasting no more than three hours.

For each video, detailed data were systematically recorded, including the number of views, likes, dislikes, comments, duration, upload date, country of origin, upload source, video category, and uploader’s gender. To ensure consistency and reduce bias, one researcher recorded the likes, dislikes, and comments during each viewing. Viewer engagement was quantified using two key metrics: the interaction index, calculated as [(likes–dislikes)/total views] × 100, and the viewing rate, determined as [(views/days since upload)] × 100.

The videos analyzed in this study were produced by a diverse range of content creators, including healthcare professionals, such as specialists and dentists, hospitals, universities, commercial entities, individuals, and other organizations, such as television networks and news agencies [13]. To categorize these content creators, the authors conducted a qualitative assessment based on several indicators. These included the information provided in the channel’s “About” section, any self-introduction or professional claims within the video content, and links to institutional websites or social media accounts. The uploader’s channel name was also considered when it provided a clear context regarding professional affiliation. This comprehensive evaluation enabled a reliable classification of the uploader. The intended audience varied, encompassing professionals, the general public, and a combination of both. The primary objectives of these videos include education, patient experience sharing, product promotion, and entertainment [14]. To further categorize content creators, they are classified based on their follower count into five distinct influencer tiers: nanoinfluencers (social media users with approximately 1–10,000 followers), microinfluencers (10,000–50,000 followers), midtier influencers (50,000–500,000 followers), macroinfluencers (500,000–1 million followers), and megainfluencers (1–5 million followers) [15]. To evaluate the effect of visual content complexity, the videos were grouped into three categories. Group 1 included videos without any educational visuals (e.g., no intraoral footage or diagrams), although a speaker may be visible. Group 2 included videos with a single type of educational visual (e.g., animation only), whereas Group 3 featured multiple types of visuals (e.g., both clinical footage and text overlays). Videos without any visual components (e.g., audio-only files or black screen recordings) were excluded from the study [15].

Before the full data collection, both reviewers independently assessed 20 pilot videos to calibrate the scoring process and categorization criteria. Discrepancies were discussed and resolved by consensus to ensure the consistency of the results.

Based on a literature review, the videos were evaluated for the presence of 13 key topics related to labial frenectomy: (1) definition of frenectomy; (2) frenectomy procedure; (3) indications; (4) contraindications; (5) advantages and disadvantages; (6) complications; (7) prognosis and survival; (8) alternative treatments; (9) cost; (10) postoperative pain; 11) swelling; 12) bruising; and 13) bleeding. Each topic was assigned 1 point, and the TCS was calculated from 13 points: a score of 0–6 was considered low content and a score of 7–13 was considered high content [16]. For each video, one point was awarded for each topic that was present and adequately covered, resulting in a maximum score of 13. No points were deducted for missing content; the scoring system was additive in nature. This approach allowed for the objective quantification of the breadth of educational content presented in each video.

The VIQI employs a 5-point Likert scale to assess the flow of information, factual accuracy, and audiovisual quality [17–20]. Each video was systematically analyzed based on critical attributes such as information flow, factual accuracy, and overall production quality. In addition to these core criteria, an extra point was assigned for including specific elements that enhance viewer engagement and comprehension, including still images, animations, community interviews, video captions, and report summaries. Moreover, the accuracy of each video was rigorously assessed by examining the alignment between its title and actual content to ensure that the information presented was both relevant and consistent [21].

The Modified DISCERN tool was developed as a standardized instrument to assist both individuals seeking healthcare services and information providers in evaluating the credibility of treatment options for various health conditions using reliable written materials. Originally designed by Charnock et al. and Singh et al., this five-question survey serves as a widely recognized framework for assessing the reliability and quality of health-related information (https://www.discern.org.uk/discern_instrument.php). This tool is commonly used to evaluate the accuracy and trustworthiness of health-related websites, brochures, and other informational sources. In the survey, positive responses were given a score of 1, whereas negative responses were given a score of 0, with higher scores indicating greater reliability of the video content [22, 23].

The HONcode is an eight-item ethical framework used to evaluate the transparency and reliability of online health information. It includes the following principles: authoritativeness, complementarity, privacy, attribution, justifiability, transparency, financial disclosure, and advertising policies. Each principle is scored as 0 (absent) or 1 (present), with a total possible score of 8, indicating maximum ethical compliance and trustworthiness [24–26].

The GQS is a five-point ordinal scale used to assess the overall quality, flow, and usefulness of online health content, particularly for patient education. Scores ranged from 1 (very poor) to 5 (excellent) [27].

The JAMA scoring system was used to evaluate the credibility of health information or sources based on four key criteria: authorship, attribution, disclosure, and currency. Each item is scored 0 or 1, with a maximum score of 4 indicating high transparency and credibility [28].

To assess inter-rater reliability, Cohen’s kappa coefficients were calculated for each scoring tool based on the independent evaluations of the two reviewers prior to reaching a consensus. The calculated kappa values indicated a high level of agreement between the reviewers for all scoring instruments. After this assessment, any discrepancies were resolved through structured discussion and literature-based consensus, and the final agreed-upon scores were used for the subsequent statistical analyses.

All analyses were performed using SPSS for Windows version 22. Normality was assessed using the Kolmogorov–Smirnov test, as well as skewness and kurtosis values. Depending on the data distribution, either parametric or non-parametric tests were applied. For group comparisons, the Mann–Whitney U test (two groups) and Kruskal–Wallis H test (three or more groups) were used, with Bonferroni correction for multiple comparisons. Relationships between numerical variables were analyzed using Spearman’s rho and point-biserial correlation, and categorical variables were assessed using the chi-square or Fisher’s exact test. Statistical significance was set at p < 0.05.

Results

The first 250 videos retrieved from YouTube via the search term “labial frenectomy” underwent an initial screening process. On the basis of the exclusion criteria established in our study, 182 videos were deemed ineligible and subsequently excluded. As a result, all analyses and evaluations were conducted on the remaining 68 videos.

Cohen’s kappa coefficients were calculated to assess inter-rater agreement before the consensus process. The results indicated substantial to almost perfect agreement across all tools: VIQI (κ = 0.824), Modified DISCERN (κ = 0.893), GQS (κ = 0.836), JAMA (κ = 0.883), TCS (κ = 0.872), and HONcode (κ = 0.848). These findings confirm the reliability of the initial assessments and reinforce the robustness of the study’s methodological approach.

Among the videos included in the analysis, those from the United States accounted for the greatest number of uploads (55%, n = 37), followed by those from India (23%, n = 19). Two videos were uploaded by users from the Philippines and Iran, (each 3%, n = 2). The remaining videos were uploaded by users from Poland, Brazil, Australia, Egypt, Ireland, Iraq, and Kuwait (each 2%, n = 1). Descriptive statistics regarding the characteristics of the YouTube videos are presented in Table 1.

Table 1.

The statistical distribution of YouTube videos based on their characteristics

	Minimum	Maximum	Mean	Standard Deviation	Median
Number of views	19.00	2515624.00	63394.47	305886.21	4596.00
Likes	0.00	9600.00	302.60	1170.69	33.50
Dislikes	0.00	1500.00	30.32	181.82	0.00
Days since upload	241.00	5506.00	2047.88	1242.56	1874.50
Duration in minutes	3.00	19.35	6.80	3.49	6.06
Number of comments	0.00	893.00	30.85	112.52	2.00
Viewing rate	3.19	86181.02	2484.89	10502.69	255.16
Interaction index	0.00	17.79	1.44	2.58	0.68
Total Content score	1.00	8.00	3.15	2.28	2.00
VIQI	4.00	17.00	10.09	3.92	10.00
Modified DISCERN	0.00	3.00	1.46	0.97	1.00
JAMA	0.00	3.00	1.49	0.76	1.00
HONcode	1.00	6.00	2.57	1.50	2.00
	Group			n	%
Source of upload	Healthcare professionals (specialist. dentist)			42	61.76
	Hospital/University			6	8.82
	Commercial			12	17.65
	Layperson			7	10.29
	Other (tv channels. new agencies)			1	1.47
Target audience	Professional			13	19.12
	Layperson			31	45.59
	Both			24	35.29
Influencer group	Nano-influencers			41	60.29
	Micro-influencers			16	23.53
	Mid-tier influencers			10	14.71
	Macro-influencers			1	1.47
Video type	Educational			45	66.18
	Patient testimonial			10	14.71
	Product advertisement			7	10.29
	Amusement purposes			6	8.82
Quality of communication	Group 1 – no visual			8	11.76
	Group 2 – single visual			9	13.24
	Group 3 – multiple visuals			51	75.00
Gender	Male			52	76.47
	Female			16	23.53
Total Content Score (groups)	Low Content Video			58	85.29
	High Content Video			10	14.71
GQS	Score 1 – very poor quality			19	27.94
	Score 2 – low quality			25	36.76
	Score 3 – moderate quality			16	23.53
	Score 4 – high quality			8	11.76

VIQI Video Information and Quality Index, JAMA Journal of the American Medical Association, HONcodeHealth On the Net Code, GQSGlobal Quality Scale

A comparison of videos based on TCS groups revealed that videos classified within the high-content group presented significantly higher VIQI scores than those in the low-content group did (z = −3.10; p < 0.01). Regarding the modified DISCERN scores, the high-content video group achieved a mean score of 2.30 ± 0.48, with a median of 2.00, whereas the low-content video group had a considerably lower mean score of 1.31 ± 0.96, with a median of 1.00. This difference was statistically significant, indicating that high-content videos were generally more reliable and informative than their low-content counterparts were (z = −3.24; p < 0.01). Similarly, an analysis of HONcode scores across TCS groups demonstrated that videos in the high-content category received significantly higher scores than those in the low-content category did (z = −2.30; p < 0.02). A detailed summary of these statistical comparisons is presented in Table 2.

Table 2.

Comparison of YouTube video characteristics based on total content score groups

	Total Content Score Groups										z/p
	Low Content Video					High Content Video
	Min	Max.	Median	Mean	SD	Min.	Max.	Median	Mean	SD
Number of views	54.00	2515624.00	6412.00	72112.07	330676.89	19.00	83698.00	1957.00	12832.40	26211.21	z= −1.25/p = 0.21
Likes	0.00	9600.00	63.50	324.17	1261.27	0.00	911.00	18.00	177.50	327.51	z= =−0.66/p = 0.51
Dislikes	0.00	1500.00	0.00	34.79	196.74	0.00	26.00	0.00	4.40	8.36	z= −0.49/p = 0.62
Days since upload	256.00	5506.00	1919.50	2132.84	1285.17	241.00	2627.00	1548.00	1555.10	846.77	z= −1.24/p = 0.22
Duration in minutes	3.00	19.35	6.06	6.43	2.85	3.23	17.85	7.16	8.99	5.75	z= −0.67/p = 0.50
Number of comments	0.00	893.00	3.00	34.16	121.32	0.00	74.00	1.00	11.70	23.60	z= −0.48/p = 0.63
Viewing rate	3.19	86181.02	384.92	2604.45	11285.34	5.58	11093.77	90.69	1791.44	3733.40	z= −1.06/p = 0.29
Interaction index	0.00	17.79	0.61	1.40	2.70	0.00	5.88	1.14	1.69	1.80	z= −1.27/p = 0.20
VIQI	4.00	17.00	9.00	9.47	3.77	8.00	17.00	14.00	13.70	2.71	z= −3.10/p < 0.01*
Modified DISCERN	0.00	3.00	1.00	1.31	0.96	2.00	3.00	2.00	2.30	0.48	z= −3.24/p < 0.01*
JAMA	0.00	3.00	1.00	1.47	0.80	1.00	2.00	2.00	1.60	0.52	z= −0.68/p = 0.50
HONcode	1.00	6.00	2.00	2.45	1.55	2.00	5.00	3.00	3.30	0.95	z= −2.30/p = 0.02*

Min Minimum, Max Maximum, SD Standart deviation, VIQI Video Information and Quality Index, JAMA Journal of the American Medical Association, HONcode Health On the Net Code, GQS Global Quality Scale

Mann Whitney U Test; ^* p < 0.05

Statistical analysis revealed no significant differences between the low and high-content groups in terms of the following parameters: source of upload (χ² = 2.89; p = 0.58), target audience (χ² = 1.31; p = 0.52), influencer group (χ² = 6.39; p = 0.09), video type (χ² = 3.96; p = 0.27), and quality of communication (χ² = 0.48; p = 0.79). A significant association was found between the TCS groups and the gender parameter (χ² = 4.57; p = 0.03; Cramer’s V = 0.259). Among videos featuring males, 90.38% had low content quality, whereas 9.62% had high content quality. In contrast, 68.75% of the videos featuring females had low content quality, and 31.25% had high content quality (Table 3).

Table 3.

The distribution of demographic data for YouTube videos across total content score groups

Parameter	Group	Total content score groups				X2	p; Cramer’s V
		Low Content Video		High Content Video
		n	%	n	%
Source of upload	Healthcare professionals	34	80.95	8	19.05	2.89	0.58
	Hospital/university	5	83.33	1	16.67
	Commercial	12	100.00	0	0.00
	Layperson	6	85.71	1	14.29
	Other	1	100.00	0	0.00
Target audience	Professional	12	92.31	1	7.69	1.31	0.52
	Layperson	27	87.10	4	12.90
	Both	19	79.17	5	20.83
Influencer group	Nano-influencers	36	87.80	5	12.20	6.39	0.09
	Micro-influencers	13	81.25	3	18.75
	Mid-tier influencers	9	90.00	1	10.00
	Macro-influencers	0	0.00	1	100.00
Video type	Educational	36	80.00	9	20.00	3.96	0.27
	Patient testimonial	10	100.00	0	0.00
	Product advertisement	7	100.00	0	0.00
	Amusement purposes	5	83.33	1	16.67
Quality of communication	Group 1 – no visual	7	87.50	1	12.50	0.48	0.79
	Group 2 – single visual	7	77.78	2	22.22
	Group 3 – multiple visulas	44	86.27	7	13.73
Gender	Male	47	90.38	5	9.62	4.57	p = 0.03 Cramer’sV:0.259*
	Female	11	68.75	5	31.25

^*p < 0.05; X², chi-square test; %: Row percentage

When the interaction index, TCS, VIQI, modified DISCERN, JAMA, and HoNcode parameters were compared on the basis of GQS scores, statistically significant differences were found between the GQS score groups. With respect to the interaction index scores, the Bonferroni post Hoc correction revealed that the interaction index value for Score 4 was significantly greater than that for Score 1 (p < 0.02). For the TCS (KW = 37.01; p < 0.01), VIQI (KW = 37.86; p < 0.01), modified DISCERN (KW = 34.95; p < 0.01), and HONcode (KW = 36.30; p < 0.01), the Bonferroni post Hoc correction revealed that the GQS scores for Scores 3 and 4 were significantly higher than those for Scores 1 and 2 across all four parameters. With respect to the JAMA scores, the Bonferroni post Hoc correction revealed that the JAMA index values for Scores 3 and 4 were significantly greater than those for Score 1 (p < 0.01) (Table 4).

Table 4.

Comparison of video characteristics based on demographic and evaluation parameters across GQS score groups

Parameter	Group	GQS					KW	p	Comparison between groupsa
		Minimum	Maximum	Median	Mean	Standard Deviation
Number of views	Score 1	65.0	275824.0	5816.0	44579.21	78683.98	0.38	0.94
	Score 2	119.0	92467.0	3223.0	17857.80	28141.59
	Score 3	19.0	2515624.0	9984.0	183408.19	623225.17
	Score 4	51.0	27537.0	7622.0	10355.38	10896.38
Likes	Score 1	0.0	911.0	29.0	155.58	258.16	1.36	0.72
	Score 2	0.0	837.0	28.0	145.92	222.14
	Score 3	0.0	9600.0	52.0	793.06	2367.73
	Score 4	3.0	658.0	130.0	160.50	211.74
Dislikes	Score 1	0.0	70.0	0.0	10.05	21.75	0.92	0.82
	Score 2	0.0	85.0	0.0	7.36	17.51
	Score 3	0.0	1500.0	1.0	104.19	372.78
	Score 4	0.0	11.0	1.0	2.50	3.78
Days since upload	Score 1	640.0	5506.0	2510.0	2645.63	1513.64	5.41	0.14
	Score 2	256.0	4092.0	1904.0	1847.76	898.91
	Score 3	340.0	5251.0	1527.5	1990.38	1327.88
	Score 4	241.0	2627.0	1392.0	1368.63	830.56
Duration in minutes	Score 1	3.2	8.2	4.8	5.08	1.51	7.28	0.06
	Score 2	3.2	19.4	6.2	7.09	3.65
	Score 3	3.0	17.9	6.9	7.68	3.88
	Score 4	3.8	15.5	6.9	8.26	4.55
Number of comments	Score 1	0.0	227.0	2.0	28.16	57.24	0.61	0.89
	Score 2	0.0	65.0	1.0	10.04	15.96
	Score 3	0.0	893.0	6.0	67.50	220.73
	Score 4	0.0	145.0	3.0	29.00	53.23
Viewing rate	Score 1	3.2	6496.9	214.4	1462.36	2262.23	0.26	0.97
	Score 2	6.4	3782.9	242.5	785.33	1147.64
	Score 3	3.5	86181.0	376.0	6611.63	21294.69
	Score 4	5.6	11093.8	617.5	1971.01	3745.16
Interaction index	Score 1	0.0	17.8	0.3	1.32	4.01	9.91	0.02	1 < 4
	Score 2	0.0	9.4	0.7	1.41	2.07
	Score 3	0.0	4.6	0.9	1.35	1.44
	Score 4	0.5	5.9	1.5	2.04	1.68
Total Content Score	Score 1	1.0	7.0	1.0	1.53	1.39	37.01	0.00	1,2 < 3,4
	Score 2	1.0	7.0	2.0	2.36	1.50
	Score 3	1.0	8.0	4.0	4.63	1.96
	Score 4	5.0	8.0	7.0	6.50	1.31
VIQI	Score 1	4.0	16.0	6.0	6.68	2.77	37.86	0.00	1,2 < 3,4
	Score 2	5.0	14.0	10.0	9.20	2.74
	Score 3	9.0	17.0	13.0	12.94	2.79
	Score 4	14.0	17.0	15.5	15.25	1.16
Modified DISCERN	Score 1	0.0	2.0	1.0	0.68	0.67	34.95	0.00	1,2 < 3,4
	Score 2	0.0	3.0	1.0	1.20	0.65
	Score 3	1.0	3.0	2.0	2.13	0.81
	Score 4	2.0	3.0	3.0	2.75	0.46
JAMA	Score 1	0.0	2.0	1.0	0.95	0.52	18.54	0.00	1 < 3,4
	Score 2	0.0	3.0	1.0	1.48	0.77
	Score 3	1.0	3.0	2.0	1.88	0.72
	Score 4	1.0	3.0	2.0	2.00	0.53
HONcode	Score 1	1.0	3.0	1.0	1.42	0.61	36.30	0.00	1,2 < 3,4
	Score 2	1.0	5.0	2.0	2.12	1.05
	Score 3	2.0	6.0	4.0	3.94	1.48
	Score 4	3.0	6.0	4.0	4.00	1.07

GQS scores were interpreted as follows: Score 1 = very poor quality, Score 2 = low quality, Score 3 = moderate quality, Score 4 = high quality

KW Kruskal Wallis value, GQS Global Quality Scale, VIQI Video Information and Quality Index, JAMA Journal of the American Medical Association, HONcode Health On the Net Code

^aScore 1 vs. Score 2 vs. Score 3 vs. Score 4 groups, post hoc: Bonforreni–corrected Mann–Whitney U tests were conducted

Spearman’s rho correlation analysis revealed positive, moderate, and statistically significant relationships between the TCS groups and the GQS (r = 0.392; p < 0.001), VIQI (r = 0.379; p < 0.001), and modified DISCERN scores (r = 0.396; p < 0.001). These results indicate that videos in the high-content video category received higher scores across these parameters, suggesting that videos with higher content quality are consistently rated more favorably in terms of the GQS, VIQI, and modified DISCERN. Additionally, a positive but weak, statistically significant relationship was found between HONcode scores and TCS groups (r = 0.281; p < 0.02) (Table 5).

Table 5.

The correlation matrix between the characteristics of the videos and the obtained scores

		Total content score groups	GQS	VIQI	JAMA	Modified DISCERN	HONcode
Total content score groups	r	1.00	0.392	0.379	0.08	0.396	0.281
	p		0.00	0.00	0.50	0.00	0.02
GQS	r	0.361	1.00	0.747	0.522	0.711	0.721
	p	0.00		0.00	0.00	0.00	0.00
VIQI	r	0.379	0.747	1.00	0.588	0.819	0.790
	p	0.00	0.00		0.00	0.00	0.00
JAMA	r	0.08	0.522	0.588	1.00	0.613	0.688
	p	0.50	0.00	0.00		0.00	0.00
Modified DISCERN	r	0.396	0.711	0.819	0.613	1.00	0.773
	p	0.00	0.00	0.00	0.00		0.00
HONcode	r	0.281	0.721	0.790	0.688	0.773	1.00
	p	0.02	0.00	0.00	0.00	0.00
Number of views	r	−0.15	−0.03	0.08	0.242	0.07	0.14
	p	0.21	0.78	0.49	0.05	0.57	0.26
Likes	r	−0.08	0.14	0.277	0.274	0.21	0.22
	p	0.51	0.25	0.02	0.02	0.09	0.07
Dislikes	r	−0.06	0.06	0.15	0.276	0.13	0.13
	p	0.63	0.61	0.23	0.02	0.30	0.28
Days since upload	r	−0.15	−0.250	−0.19	0.04	−0.11	−0.07
	p	0.22	0.04	0.11	0.73	0.35	0.57
Duration in minutes	r	0.08	0.312	0.494	0.353	0.504	0.437
	p	0.51	0.01	0.00	0.00	0.00	0.00
Number of comments	r	−0.06	−0.01	0.11	0.18	0.03	0.07
	p	0.64	0.94	0.38	0.15	0.81	0.57
Viewing rate	r	−0.13	0.03	0.16	0.253	0.12	0.17
	p	0.29	0.81	0.20	0.04	0.32	0.16
Interaction index	r	0.16	0.335	0.334	−0.01	0.240	0.15
	p	0.20	0.01	0.01	0.93	0.05	0.23

Bold indicates that the difference between groups is statistically significant. The relationship between numerical and nominal variables was assessed using the point-biserial correlation coefficient

VIQI Video Information and Quality Index, JAMA Journal of the American Medical Association, HONcode Health On the Net Code, GQSGlobal Quality Scale

r: Spearman’s rho correlation analysis

Further analysis of the GQS scores revealed strong positive and statistically significant correlations with the VIQI (r = 0.747; p < 0.001) and modified DISCERN scores (r = 0.711; p < 0.001). Moderate positive and statistically significant relationships were also observed with HONcode (r = 0.721; p < 0.001) and JAMA scores (r = 0.522; p < 0.001). Strong positive and statistically significant correlations were found between the VIQI scores and both the modified DISCERN (r = 0.819; p < 0.001) and HONcode scores (r = 0.790; p < 0.001). A moderate positive and statistically significant relationship was noted with the JAMA score (r = 0.588; p < 0.001) (Table 5).

Analysis of the JAMA scores revealed moderate positive and statistically significant relationships with both the modified DISCERN (r = 0.613; p < 0.001) and HONcode scores (r = 0.688; p < 0.001). A highly positive and statistically significant correlation was found between the modified DISCERN score and HONcode score (r = 0.773; p < 0.001). Furthermore, a positive but weak, statistically significant relationship was identified between the Likes variable and both the VIQI (r = 0.277; p = 0.02) and JAMA scores (r = 0.274; p = 0.02). A similar relationship was observed between the Dislikes score and the JAMA score (r = 0.276; p = 0.02) (Table 5).

A negative but weak statistically significant relationship was found between the ‘Days since upload’ variable and the GQS score (r = −0.250; p < 0.05). Moderate positive and statistically significant correlations were found between the ‘Duration in minutes’ variable and the VIQI score (r = 0.494; p < 0.001), modified DISCERN score (r = 0.504; p < 0.001), and HONcode score (r = 0.437; p < 0.001). A positive but weak statistically significant relationship was also observed between the viewing rate and JAMA score (r = 0.253; p < 0.05). Finally, positive and moderate statistically significant relationships were found between the interaction index and both the GQS (r = 0.335; p < 0.05) and the VIQI (r = 0.334; p < 0.05). Additionally, a positive but weak statistically significant relationship was noted between the interaction index and the modified DISCERN score (r = 0.240; p < 0.05) (Table 5).

Discussion

YouTube has emerged as a major platform for accessing health-related information; however, its unregulated nature raises concerns regarding the accuracy and reliability of its content. This is particularly important for procedures such as labial frenectomy, where misleading information may affect patient understanding [29, 30]. Although previous studies have assessed YouTube content on other dental procedures, this is the first to focus specifically on labial frenectomy [13, 21, 31–36]. This study evaluated the quality and educational value of related videos using multiple established tools.

Our findings demonstrated substantial variability in the reliability and educational quality of YouTube videos on labial frenectomy. Using established evaluation tools (VIQI, DISCERN, GQS, JAMA, TCS, and HONcode), we found that most videos were of low to moderate quality. This finding is consistent with earlier studies that assessed YouTube content on topics such as oral biopsies [35], dental implants [36–38], and maxillary expansion [21]. The similarity in the results suggests an ongoing problem with the accuracy of health-related content on video-sharing platforms.

Our study highlights the ongoing concerns regarding the reliability of online health content. As in previous research, only a small portion of the videos met high-quality standards, especially when assessed using the modified DISCERN and JAMA tools [11, 12]. This is important because many patients and students use YouTube as their main source of information. Similar findings were reported by Aurlene et al. and Ozisci et al., who highlighted the prevalence of misleading or incomplete health-related videos [39, 40].

Our analysis found no significant differences in video quality based on the upload source, target audience, video type, or communication style. These results align with previous studies [39, 40], except for the “upload source” parameter. This difference may be due to the small number of videos from TV stations, advertisements, or amateur creators in our sample, which could have affected the distribution. Another possible explanation is that not all healthcare professional-generated videos are intended to be comprehensive educational tools. Some may prioritize branding or procedural demonstration without addressing key aspects such as indications, risks, and postoperative care. Consequently, even professionally produced videos may receive lower quality scores. Ercan and Yavan [41] reported that videos uploaded by professionals had higher reliability and content scores than those uploaded by patients, despite the latter generating greater engagement. In contrast, our study found no such distinction, likely due to the small number of non-professional sources. These findings suggest that for niche topics such as labial frenectomy, content quality may depend more on informational depth than uploader identity. Future research should examine how platform algorithms affect the visibility of evidence-based content on social media.

A notable finding was that many videos uploaded by healthcare professionals and hospitals were classified as low-content according to the Total Content Score. Several factors may explain these findings. First, creators may assume a baseline level of medical knowledge in their audience, resulting in less detailed explanations. Second, institutional videos are often designed for promotional or informational purposes rather than for comprehensive patient education. Finally, production time constraints or a preference for brevity may limit content depth. Similar patterns have been observed in a previous study, in which professional sources did not always correlate with higher content quality scores [18]. These findings emphasize that credibility alone is insufficient; the completeness and educational value of the content must also be assessed.

Our analysis showed that videos with female presenters tended to receive higher content scores than those with male presenters. Although this association has not been widely studied, it may be influenced by differences in communication style, such as clearer speech, a more engaging tone, or a greater focus on educational detail. However, this result should be interpreted with caution, as these differences may stem from confounding factors such as audience expectations, cultural communication styles, or video format rather than the presenter’’s gender alone. Care must be taken to avoid overgeneralizing these results or reinforcing the unintended gender stereotypes. Nonetheless, presenter-related factors may influence the effectiveness of videos and deserve further investigation.

The TCS reflected How thoroughly each video addressed the 13 key labial frenectomy-related topics. Videos with higher TCS scores also tended to score better on the VIQI, modified DISCERN, and HONcode, indicating that detailed content enhances both quality and reliability. The GQS showed strong correlations with other tools, including VIQI, JAMA, DISCERN, TCS, and HONcode, further emphasizing the value of informative content. However, TCS and JAMA scores were not significantly correlated, likely because JAMA focuses on authorship, citations, and update dates, which are unrelated to audiovisual clarity or educational depth. These results support the use of multiple evaluation tools to comprehensively assess online health information. A combined approach that integrates ethical, structural, and audio-visual criteria offers a more complete picture. Future research should explore how algorithms and platform policies can promote the visibility of high-quality educational material through collaboration between health professionals and digital platforms.

A major strength of this study was the use of multiple scoring systems, which allowed for a more comprehensive evaluation. Most previous studies used only a single tool [11, 34, 42, 43], whereas our multi-tool approach captured different dimensions of the video quality. Each instrument has strengths and limitations: VIQI assesses audiovisual flow and structure but not medical accuracy [11]; DISCERN focuses on trust and reliability but overlooks engagement-related factors; GQS captures overall usefulness but can be subjective [44]; JAMA emphasizes authorship and referencing but not educational clarity [11]; and HONcode evaluates ethical standards but ignores visual communication [45]. By combining these tools, we mitigated the individual limitations and achieved a more balanced assessment. This integrated method is also supported by the recent literature [46], which advocates multi-criteria approaches for evaluating online health content.

Social media platforms, particularly YouTube, have become major sources of health-related information, a trend accelerated by the COVID-19 pandemic. While this shift has improved access, it has also increased the spread of unverified or misleading content [11, 40]. YouTube’s algorithm prioritizes engagement metrics, such as likes, views, and watch time, over educational accuracy. Consequently, attention-grabbing videos often rank higher, regardless of their factual reliability [40]. Our findings support this concern, showing no significant pilot correlation between engagement indicators and video quality and quantity based on tools such as the GQS. This disconnect highlights the need for systems that promote evidence-based content on digital platforms such as TikTok.

Our findings showed that patient testimonial videos were common among low-content categories and scored lower on objective quality metrics. However, these videos often had high viewer engagement, suggesting their potential to influence audiences regardless of their educational quality. Patient-generated content, such as testimonial videos, substantially influences the public perception of surgical procedures. Given that clinical outcomes may vary based on technique, surgeon expertise, and individual factors, personal accounts may not always represent typical patient experiences. Overly positive or negative portrayals may shape expectations that misalign with clinical realities. This highlights the importance of supplementing such anecdotal content with accurate professional information to guide informed patient decision-making.

To combat the ongoing spread of misinformation on platforms such as YouTube, both institutional and systemic actions are needed. Healthcare organizations and academic societies should take the lead in producing accessible, evidence-based educational videos that meet high-quality standards. These institutions are uniquely positioned to ensure scientific accuracy, uphold ethical principles, and address patients’ information needs. Additionally, developing clear guidelines for digital patient education, such as using plain language, citing reliable sources, and ensuring audiovisual clarity, can raise the overall standard of online health content. Collaboration among content creators, digital platforms, and regulatory bodies may also support the development of verification systems and visual indicators that help users identify trustworthy information.

Our findings have several practical implications for both clinicians and patients. High-quality videos of labial frenectomy typically include clear procedural explanations, accurate visuals, references to clinical guidelines, and thorough coverage of pre- and postoperative care. These features may serve as markers of reliable content in the absence of a peer review. Clinicians should guide patients toward such videos and consider incorporating these elements when creating their educational materials. This approach can improve patient understanding, reduce misinformation, and support better clinical outcomes in the future.

This study had several limitations that should be acknowledged. First, we did not conduct a systematic analysis of the specific products or brands promoted in the videos. While some content clearly aimed to advertise services or devices, we did not explore the prevalence, strategies, or ethical implications of such commercialization. Since promotional messaging in health-related content can influence viewer trust, decision-making, and health literacy, future studies should examine this aspect in greater depth. Second, the dynamic nature of YouTube content poses reproducibility challenges, as videos may be deleted, altered, or newly uploaded over time. Third, the study only included English-language videos, which limits the generalizability of the findings to other linguistic and cultural settings. Fourth, although a 3-minute minimum duration threshold was established to ensure adequate educational content, this criterion was not based on a standardized reference. As a result, some brief but potentially informative videos may have been unintentionally excluded from the study. Additionally, while the TCS captured the presence of key topics, it did not evaluate the accuracy of the information presented in them. This limitation was minimized through the use of additional tools, such as the modified DISCERN, which evaluates reliability and scientific soundness. Finally, some evaluation tools used, such as the GQS, involve a degree of subjectivity that may affect scoring consistency. Future research should consider multilingual, longitudinal designs and potentially incorporate automated evaluation tools to enhance objectivity and scalability.

Conclusions

Our findings support the hypothesis that a significant proportion of YouTube videos related to the labial frenectomy procedure exhibit poor educational quality and provide incomplete or misleading information. While a handful of videos offered valuable educational content, many others fell short of meeting the essential standards for both reliability and comprehensiveness. This highlights an urgent need for stronger content regulation and the promotion of evidence-based, expert-reviewed educational materials across digital platforms. By incorporating structured evaluation frameworks, encouraging the submission of peer-reviewed content, and refining platform algorithms to prioritize quality over engagement, we can substantially increase the educational value of online health information. Furthermore, fostering greater collaboration between content creators, healthcare institutions, and digital platforms is critical in bridging the gap between content accessibility and accuracy. By adopting these strategies, policymakers and educators can better leverage the potential of social media while minimizing its risks, ultimately improving the quality and trustworthiness of online health education for global audiences.

Abbreviations

TCS

Total content score

VIQI

Video information and quality index

GQS

Global quality scale

JAMA

Journal of the American Medical Association

HONcode

Health on the Net Code

SPSS

Statistical package for the social sciences

Authors’ contributions

M.I. conceived and designed the study, conducted the literature review, collected the data, provided critical feedback, and wrote and edited the manuscript. M.E.A. performed the analysis, conducted the literature review, collected the data, and wrote and edited the manuscript. Both authors meet the authorship criteria and both reviewed and approved the final draft.

Funding

This study was conducted without any external funding or financial support from any funding agencies.

Data availability

The data used in this study were obtained from publicly available YouTube videos. The dataset, including video links and extracted information, can be shared upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable. This article does not contain any studies with human participants or animals performed by any of the authors. No ethical committee approval is required since this study is performed on the publicly available Internet data. For this type of study, formal consent is not required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.