Introduction
Patients in the orthognathic surgery process frequently turn to social media to obtain information and share their experiences. Accordingly, the present study analyzed YouTube comments to evaluate patient satisfaction, emotional tendencies, and social interactions, aiming to draw implications for improving communication and psychological support strategies.
Methods
A total of 250,582 comments from 8044 YouTube videos on orthognathic surgery were collected using Mozdeh (University of Wolverhampton, Wolverhampton, UK) and R software (version 4.2.2, Posit, PBC, Boston, Mass). After preprocessing, 111,953 comments remained. Sentiment analysis (AFFIN/NRC lexicons), structural topic modeling, and manual thematic analysis (10% sample) were performed. Gender-based discourse differences were examined using Chi-square tests with Benjamini-Hochberg correction.
Results
Ten main themes were identified through structural topic modeling and thematic analyses. Positive emotions were mostly expressed, though fear and negativity were also observed. Esthetic and emotional expressions were used more often by women, whereas cost-related terms were used more often by men. Thematic qualitative analyses revealed that YouTube videos and comments were perceived by patients as a source of psychological support, whereas certain misleading content was also found to have the potential to misguide patients during the treatment process.
Conclusions
This study has transcended the limited scope of traditional surveys, which are confined to specific time periods and small sample sizes, by making it possible to access patient experiences shared on the YouTube platform without temporal or spatial constraints. The analyses revealed prevalent misconceptions on social media and assessed patients’ expectations, concerns, and psychological states. The findings serve as a practical guide for healthcare professionals in developing patient-centered communication and care strategies.
Highlights
-
•
Patient experiences related to the surgical process were accessed without time or location limits.
-
•
Patients’ expectations and concerns during the orthognathic surgery process were identified.
-
•
Perioperative and postoperative challenges and healthcare service disruptions were identified.
-
•
The presence of misleading and inaccurate information on social media was demonstrated.
-
•
Findings supporting patient-centered, empathetic communication and expectation management were presented.
Although dental malocclusions can be corrected through orthodontic treatment, craniofacial deformities at the skeletal level require orthognathic surgery. In recent years, orthognathic surgery has become an increasingly preferred treatment approach that supports the maintenance of balanced anatomic and functional relationships in patients with severe skeletal discrepancies, while also providing an ideal and esthetically pleasing facial appearance. The need for orthognathic surgery arises from a variety of concerns, including respiratory, speech, and mastication disorders, temporomandibular joint (TMJ) dysfunctions, and, most importantly, the psychosocial issues caused by dentofacial disharmony. , Although orthognathic surgical procedures have become routine practices for surgeons, patients often perceive the process as a complex, lengthy, and uncertain sequence of events. Therefore, providing psychological support and equipping patients with adequate information are critically important during the decision-making phase. , Studies have reported that patients seeking orthognathic surgery often exhibit depressive symptoms in the preoperative period, along with low self-confidence and low self-esteem, poor facial image perception, and elevated levels of anxiety. These patients frequently benefit from orthognathic surgery in terms of improved social functioning, enhanced perception of facial attractiveness, positive life changes, reduced anxiety, and increased self-confidence and self-esteem. However, patients may also encounter different challenges during the postoperative adaptation period compared with the preoperative phase. Potential complications, such as facial swelling and altered dietary habits, may lead to psychological consequences, including anxiety or depression. Therefore, improving quality of life and achieving satisfactory postoperative outcomes often requires a prolonged recovery process. , Adult patients often consult internet-based sources during the decision-making process for orthognathic surgery. Although some patients obtain information from those who have previously undergone the procedure, others receive detailed explanations directly from a maxillofacial surgeon. In this context, the internet serves as a digital resource for patients seeking health-related information. These resources may include online services, such as search engines or social media platforms. Social media is regarded as a tool that reflects public opinions and attitudes by providing a forum in which individuals express their views and engage in discussions on various topics. This presents an opportunity for academicians and researchers to gain deeper insights into public perspectives. In this context, infodemiology has emerged as a scientific discipline that investigates the distribution and determinants of information in digital environments, particularly on the internet or within a population, with the ultimate goal of informing public awareness related to public health. ,
Among social media platforms, YouTube is an online video-sharing platform and the second most visited website after Google. As the internet has increasingly become a primary source of health-related information, YouTube has gained significant potential in shaping health-related discourse and currently hosts over 2 billion monthly users. , Big data analyses have transformed evolving healthcare approaches through comprehensive evaluations and have created potential for delivering improved clinical solutions. Analyzing user comments on this widely used platform is therefore of critical importance. Although surveys and other qualitative research methods generally indicate high levels of satisfaction among patients who have undergone orthognathic surgery, , studies have also found that preoperative explanations and information regarding potential side effects are often insufficient. , Among the advantages of social media are ease of use, accessibility, speed, and the practical ability to access information. This ease of use enables patients to express minor dissatisfactions or complaints that they may be reluctant to articulate, have not had the opportunity to communicate to their physicians, or have been unable to convey fully through traditional surveys via comments on these platforms. The analysis of YouTube comments offers a significant opportunity to understand societal perceptions on a global scale. Although there are studies examining the content and reliability of YouTube videos related to orthognathic surgery, these studies have generally been limited by small sample sizes and have not provided a comprehensive analysis of user comments. ,
This study aimed to explore the broad spectrum of comments posted under orthognathic surgery videos on YouTube by examining emotional trends over time, identifying relevant themes, and uncovering gender-based differences. The objective was to analyze these comments to reach a genuine understanding of patient satisfaction, to gain deeper insight into the dynamics of interactions among patients, and to shed light on potential areas for improvement in orthognathic surgical care.
Material and methods
This study was designed as a cultural-infodemiology based analysis aiming to reveal large-scale patient perceptions and emotional responses related to orthognathic surgery, and to contextualize these findings in relation to clinical practice. This study did not involve any human or animal participants. Only publicly available YouTube comments were analyzed, and no personally identifiable information was collected or reported. Therefore, ethical approval was not deemed necessary.
Within the scope of this study, English-language user comments under YouTube videos related to orthognathic surgery were analyzed. For video selection, Google Trends ( https://trends.google.com ) (Google LLC, Mountain View, Calif) was used, and based on the analysis, 3 search terms were identified: orthognathic surgery, orthognathic jaw surgery, and jaw surgery. These terms were then used on April 15, 2025, to retrieve data via the YouTube Data API (version 3; Google LLC, Mountain View, Calif). A total of 24,061 videos were identified from searches conducted with these 3 keywords for each year since the establishment of YouTube. During the initial screening, duplicate videos were removed, resulting in 11,801 videos. Subsequently, videos deemed irrelevant in terms of content were excluded, and a final set of 8044 videos was deemed eligible for analysis. All subsequent analyses were conducted based on this video set. A total of 250,582 user comments associated with these videos were collected using the Mozdeh software (University of Wolverhampton, Wolverhampton, UK) via the YouTube Data API (version 3). To prevent popular videos with a high number of comments from disproportionately influencing the overall analysis, a maximum of 700 comments per video was included in the study, resulting in a final dataset of 164,446 comments. Data analysis was conducted using a combination of Mozdeh and R Studio (version 4.2.2, Posit, PBC, Boston, Mass).
A multistage preprocessing procedure was applied to the collected raw data. In this process, only comments written in English were included, and duplicate comments were removed with the help of the Mozdeh software. Before analysis, a comprehensive data cleaning and structuring process was carried out, which included steps such as cleaning, transforming, reformatting, and merging the data. Subsequently, data cleaning was performed using the stop_words function from the tidytext package in the R programming language (The R Foundation for Statistical Computing, Vienna, Austria). During this step, standard stop-word dictionaries, such as SMART, Snowball, and Onix, were used. Pronouns (eg, I and me), verb forms (eg, am and are), contractions (eg, I’m and you’re), auxiliary verbs with negations (eg, won’t and couldn’t), adverbs and conjunctions (eg, and, but, and than), and prepositions (eg, in, under, and for) were systematically removed from the text. After stop-word removal, lemmatization was performed using the textstem package, which reduced inflected forms of words to their base or root form. This step aimed to reduce lexical diversity and enhance semantic coherence in the context of natural language processing. Finally, before conducting topic modeling and sentiment analysis, all numbers, punctuation marks, and special characters were removed, and all text was converted to lowercase for standardization. As a result of these preprocessing steps, a final dataset consisting of 111,953 comments was obtained, and all subsequent analyses were performed on this refined dataset. The methodological workflow of the preprocessing procedure is illustrated in Figure 1 .
Flowchart depicting the data collection, wrangling, and analysis process.
To visualize the prominent words within the comments, word clouds were generated using the wordcloud2 package in the R programming language. The wordcloud2 package provides an HTML5-based interface for data visualization. This word cloud served as an effective visual analysis tool to highlight the overall themes present in user comments. Slang or inappropriate words were excluded from the word clouds.
Sentiment analysis is a method used to determine the emotional tone of digital texts and can be conducted in various ways using different lexicons. In the R AFFIN approach, sentiments are scored on a scale from–5 (most negative) to +5 (most positive), with a score of 0 indicating a neutral sentiment. In the R NRC approach, comments are categorized as positive or negative and further classified into specific emotional categories, such as trust, anticipation, joy, fear, anger, sadness, surprise, and disgust. In this study, both sentiment analysis approaches were applied in combination to gain a more comprehensive understanding of the emotional content within the YouTube comments.
Topic modeling is an unsupervised learning technique that aims to identify latent terms and topic groupings within a given text corpus. In this context, various text mining methods, such as latent semantic analysis, probabilistic latent semantic analysis, latent Dirichlet allocation, and structural topic models (STM), have been developed to autonomously uncover hidden themes in textual data. In this study, the STM model was selected for topic modeling. Similar to other topic modeling approaches, STM operates as a generative model based on word frequencies. However, unlike traditional models, STM offers a flexible framework that allows for the inclusion of document-level covariate information into the modeling process. By incorporating such covariates, STM not only improves inference but also enhances the interpretability of the resulting topics. These covariates can influence topic prevalence, topic content, or both, enabling the generation of richer and more nuanced insights.
In addition to the automated topic modeling conducted in R, a simple random sample of 11,200 comments, approximately 10% of the total 111,953 comments, was selected for manual thematic analysis to achieve methodological triangulation. Although previous studies have typically used smaller subsamples (approximately 5%) for qualitative data analysis, a more recent approach was adopted in this study by selecting a 10% sample. To ensure an unbiased selection process, the sample function in R was used to select the comments randomly. Thematic analysis was then applied to the selected comments.
To enhance methodological rigor and address potential concerns regarding the adequacy of the initial sample, an additional validation process was implemented. In line with methodological insights drawn from previous large-scale content analysis studies that employed similar sampling strategies, a secondary qualitative analysis was conducted on an independent 5% subsample that had not been included in the initially coded 10% dataset. During the secondary analysis, no additional or divergent themes emerged, indicating that the themes identified in the initial sample were sufficient to capture the content meaningfully. This consistency reinforced the reliability and validity of the original thematic framework. As a result, a total of 16,800 comments (15%) were qualitatively analyzed.
Thematic analysis is a systematic method that enables researchers to identify patterns of meaning within qualitative data and gain deeper insight into the phenomenon under investigation. The method involves a 6-phase process that facilitates theme extraction from large datasets. These phases include the following: (1) familiarization with the data, (2) identifying patterns through repeated reading, (3) coding (assigning labels to words, phrases, sentences, and paragraphs), (4) generating themes (grouping codes that share similar meanings), (5) reviewing themes (reorganizing, merging, or splitting themes based on coherence), (6) defining and naming themes, and (7) reporting the findings. Accordingly, 2 authors initially coded the data independently, each identifying their own themes. They then met to review the initial coding and resolve any discrepancies that emerged during the process through discussion. Based on this joint evaluation, necessary modifications were made to the themes, and the final themes were established. In addition, to evaluate intracoder reliability, a random 10% subsample was selected from the primary dataset. This subset was recoded by 1 of the researchers approximately 1 week after the initial coding process. The Cohen kappa coefficient between the 2 coding rounds was found to be 0.92, indicating an excellent level of intracoder reliability.
In this study, thematic analysis was used alongside sentiment analysis and topic modeling to enhance the validity of the analytical outcomes through methodological triangulation. To protect personal information and ensure clarity, user comments cited in the study were anonymized before presentation.
To identify gender-based discourse differences in user comments, gender-based analyses were conducted in this study. As YouTube does not provide gender information for users, gender was inferred based on username analysis. The usernames were segmented based on spacing and compound word structures; the resulting components were then matched against an expanded list of names derived from the U.S. Social Security Administration’s baby name dataset covering the years 1880 to 2024. To broaden the name pool and enhance the reliability of the analysis, several modifications were made to the Mozdeh software to enable the integration of the entire name dataset from this period. Names were classified as male or female if they had a minimum 90% gender association across the dataset. In addition, honorifics, such as Mr, Ms, and Miss, were used to support the classification process. To explore lexical variation between male and female commenters, an association mining-based analysis was conducted to identify words that showed statistically significant gender-based differences. A word was considered gendered if it appeared in at least 2 user profiles, occurred more frequently in 1 gender group, and the frequency difference was statistically significant. Based on this analysis, gender-associated lexical patterns were identified, and sentiment analysis was subsequently conducted separately for male and female users.
Statistical analysis
Gender-based differences in the usage of each word were analyzed using a 2 × 2 Chi-square test. Because of the large number of statistical tests performed, the Benjamini-Hochberg procedure was applied to reduce the risk of false positive results. This procedure adjusts the Chi-square significance threshold to maintain the probability of at least 1 false positive finding below 5%.
Results
The STM analysis of YouTube comments related to orthognathic surgery identified associated keywords along with their frequency percentages, based on which relevant thematic categories were established ( Table I ).
Table I
Keyword distribution and STM in YouTube comments on orthognathic surgery
| Topic number | Topic contribution of tokens | Relevant terms | Main themes |
|---|---|---|---|
| 1 | 18 | look, hope, video, thank, beautiful, happy, recovery, amaze, soon, wish | Viewer responses to orthognathic surgery experiences |
| 2 | 12 | feel, bad, time, day, help, week, eat, swell, lot, post | Postoperative outcomes and recovery in orthognathic surgery |
| 3 | 10, 8 | love, girl, god, sorry, omg, person, painful, friend, poor, comment | Social perceptions of orthognathic surgery |
| 4 | 10, 6 | surgery, jaw, month, chin, low, hello, scare, ago, upper, double | Perceptions of fear, anxiety, and uncertainty regarding orthognathic surgery |
| 5 | 10, 5 | people, change, life, yes, mean, plastic, money, hate, Korean, understand | Esthetic concerns and economic accessibility in orthognathic surgery |
| 6 | 8, 8 | bite, fix, underbite, surgeon, nose, cause, lip, issue, orthodontist, sleep | Functional benefits of orthognathic surgery |
| 7 | 8, 7 | brace, tooth, pain, tell, please, hurt, dentist, cut, cover, wear | The role of orthodontic treatment in the orthognathic surgery process |
| 8 | 7, 6 | doctor, bone, kid, wrong, body, difference, maybe, mew, patient, age | Optimal timing of orthognathic surgery in relation to growth and development |
| 9 | 6, 6 | pretty, okay, try, actually, nice, job, imagine, health, reason, literally | Positive perceptions of orthognathic surgery |
| 10 | 6, 4 | mouth, happen, wait, smile, cost, break, wire, hospital, lose, shut | Postoperative hospital course and care after orthognathic surgery |
Most frequently used words in YouTube comments: The word frequency analysis of YouTube comments was conducted using the word cloud visualization method ( Fig 2 ). This analysis revealed that the most frequently occurring term was surgery (n = 25,511), followed by look (n = 23,211), jaw (n = 15,409), feel (n = 8437), brace (n = 7357), and hope (n = 6754). Emotionally expressive words, such as love (n = 5661), thank (n = 5641), beautiful (n = 5033), happy (n = 3920), and hope (n = 6754), appeared with high frequency, whereas negative experience related words, such as bad (n = 5405), pain (n = 4213), and swell (n = 2780), were also notably prevalent.
Word cloud of the most frequently used keywords in YouTube comments on orthognathic surgery.
To analyze the emotional content of YouTube comments, as part of the triangulation process, a random sample of 11,200 comments was manually reviewed. The qualitative findings were consistent with the sentiment outcomes obtained through data mining. In the sentiment-based analysis of YouTube comments related to orthognathic surgery, the most frequently observed emotion was positive (n = 119,302), followed by fear (n = 78,538) and negative (n = 72,857) emotions ( Fig 3 , A ). Notable differences were observed between male and female users. Among female commenters, the most common emotion was positive (n = 7262), followed by fear (n = 4705) and joy (n = 4332) ( Fig 3 , B ). Similarly, among male commenters, positive was the most frequently expressed sentiment (n = 7943), followed by fear (n = 5012) and trust (n = 4478) ( Fig 3 , C ). Figure 4 illustrates the temporal trend of the daily average sentiment scores of YouTube comments related to orthognathic surgery between 2007 and 2025. Between 2007 and 2019, daily average sentiment scores were distributed across a wide range, from–4 to +4, with a high frequency of both extremely positive and extremely negative outliers. After 2019, a narrowing of these outliers was observed, and from 2021 onward, the distribution formed a moderately positive plateau, with scores ranging between approximately–1.5 and +2.5. Figure 5 presents the annual distribution of sentiment categories in YouTube comments between 2007 and 2025. Overall, positive expressions were found to have the highest proportion across the years. Other positively associated emotions, such as joy and trust, remained at moderate levels, with trust showing a gradual increase over time. Negative and sadness sentiments consistently remained lower than positive emotions throughout most of the time period. Although both showed a noticeable spike in 2021, they subsequently declined to previous levels in the following years. The fear sentiment generally followed a similar trend to positive emotions but exhibited a short-term increase in 2021. In contrast, anger, disgust, and surprise appeared at relatively low levels throughout the analyzed period.
Comparative distribution of sentiment categories in YouTube comments related to orthognathic surgery: A, The overall sentiment distribution; B and C , The sentiment distributions among comments posted by female and male users, respectively.
Temporal distribution of daily average sentiment scores in YouTube comments on orthognathic surgery between 2007 and 2025.
Yearly distribution of emotion categories in YouTube comments on orthognathic surgery between 2007 and 2025.
Main topics discussed in YouTube comments: When the results of the STM topic modeling and manual thematic analysis were evaluated together, 10 main themes were identified. Each of these themes was further divided into a varying number of subthemes based on content density ( Table II ). Anonymized user comments (presented in italics) were used to support inferences related to the main themes.
Table II
Themes identified through qualitative analysis of YouTube comments related to orthognathic surgery, along with subthemes derived from patient remarks associated with these themes
| Themes | Subthemes derived from patient remarks |
|---|---|
| Shared experiences of the postoperative process | Swelling potentially persisting for up to several years, and the prolonged adaptation of facial muscles to the new structure |
| Difficulties caused by persistent severe pain despite the use of analgesic medications in the postoperative period | |
| Delayed achievement of the final facial appearance, potentially taking up to several years, and challenges in adapting to the altered facial appearance | |
| Regret expressed by older patients for not having undergone the surgery at a younger age | |
| Emotional challenges, such as temporary regret, low mood, and questioning the necessity of surgery immediately after the operation | |
| Concerns arising from the clinical appearance after the resolution of decompensation during presurgical orthodontic preparation | |
| Difficulties in adapting to eating and drinking in the early postoperative period | |
| Postoperative edema temporarily concealing the outcomes of the surgical intervention | |
| Questions and information seeking regarding the postoperative process and surgical intervention | Risk of permanent mandibular nerve injury |
| Concerns regarding the grafting process and donor site morbidity | |
| Apprehensions about sneezing and vomiting during the postoperative period | |
| Postoperative sleep positioning and associated challenges | |
| Potential positive impact of the procedure on TMJ function | |
| Uncertainty about the presence of a visible scar | |
| Concerns about changes in nasal morphology | |
| Worries about postoperative skin laxity or sagging | |
| Timeline for resuming intimate activities (eg, kissing) and normalization of menstrual function | |
| Concerns about weight loss and muscle mass reduction during recovery | |
| Fear of lower lip numbness impairing trumpet performance | |
| Apprehensions about returning to sports activities postoperatively | |
| Psychosocial and functional impacts of skeletal jaw anomalies | Low self-esteem because of persistent dissatisfaction with facial appearance |
| Negative social experiences, including bullying because of visible dentofacial anomalies | |
| Frustration and impatience among adolescent patients regarding the recommended wait for skeletal maturation before orthognathic surgery | |
| Daily functional limitations resulting from skeletal deformities | |
| Facial dissatisfaction driven by internalized beauty norms | |
| Appearance-based stigma, bullying, and the pursuit of social acceptance before and after orthognathic surgery | |
| Nonsurgical alternative approaches lacking scientific evidence | Mewing technique |
| Orthotropics approach | |
| Economic barriers and cost-related concerns in accessing treatment | |
| Positive emotional outcomes and satisfaction after orthognathic surgery | Increased self-confidence and body image satisfaction after orthognathic surgery |
| Perception of the surgery as a life-changing and worthwhile decision | |
| Adaptation to postoperative changes in facial appearance over time | |
| Improved functional and psychological well-being in the long term | |
| Positive reassessment of the surgical experience after full recovery | |
| Negative emotional outcomes and surgical complications after orthognathic surgery | Risk of postoperative relapse because of surgery performed before the completion of craniofacial growth |
| Development of soft tissue asymmetries (eg, lip imbalance) after orthognathic surgery | |
| Need for second and third clinical opinions to better evaluate potential surgical complications | |
| Patient feedback regarding orthognathic surgeons | Concerns about unethical clinical practices and consent violations |
| Patient concerns regarding financially motivated treatment recommendations by certain surgeons | |
| Trust and satisfaction resulting from attentive and competent surgical care | |
| User comments expressing motivation derived from surgical experience videos |
Stay updated, free dental videos. Join our Telegram channel
Leave a Reply
VIDEdental - Online dental courses
