© The authors, 2024

Introduction

Global social media users are expected to rise from 2.86 billion in 2017 to 4.41 billion in 2025 [1]. However, these platforms also pose risks, such as the spread of misinformation, which is incorrect information shared unintentionally [2]. Frequently, this information is not verified, forcing individuals to independently distinguish between truth and falsehood [3].

Despite concerns like confidentiality breaches, legal issues, and ethical dilemmas in healthcare, the benefits of social networks—such as enhanced patient education, professional networking, increased visibility, real-time communication, and community building—often outweigh these risks [4]. Facebook, for instance, serves as a supplementary educational tool, fostering professional development through peer review and collaborative learning [5]. Additionally, social media promotes research by increasing the visibility and downloads of shared dermatological articles [6]. Learning through social media is also seen as less burdensome compared to other platforms like wikis [5].

In dermatology, various social media platforms have been studied. For example, YouTube has been analyzed for its viewership, comments, subscriptions, and shares to better understand interactions with dermatological content. It offers a diverse range of dermatological content, from rare diseases to educational videos, advocacy, entertainment, and preventative medicine. However, verifying evidence-based data and source credentials on Youtube remains a challenge [7]. Research on Instagram reveals the absence of top dermatological journals and professional organizations [8]. In contrast, Facebook is increasingly used by academic journals, associations, and patients for sharing diagnoses or joining support groups [9]. Thus, Facebook benefits both healthcare professionals and patients [10], with studies highlighting its value as a platform for professionals to communicate, share knowledge, and engage in dermatology discussions [11]. French dentists report a training gap in oral dermatology and are encouraged to use illustrated case studies to enhance their practice [12]. While photographs are important in dermatology [13], limited data exist on sharing oral dermatology case studies.

While research has explored the benefits of social media for general dermatology, little attention has been paid to its use in oral dermatology. Our study seeks to address this gap. Specifically, our objective was to evaluate the quality of communication in these cases, including their relevancy and the presence of misinformation. We conducted a cross-sectional study by selecting Facebook groups where health professionals share oral dermatology cases and analyzed these cases over six months. Each case was rated using a custom-made scoring grid to assess its clarity and educational value. Our secondary objective was to compare the quality between different groups. Given the limited data on sharing oral dermatology case studies via social media, our research offers critical insights into the quality and educational value of these cases.

Materials and methods

This cross-sectional study targets French-speaking dentists and dental students active in Facebook groups focused on oral dermatology. The design, conduct, and reporting adhere to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [14] (see Supplementary Materials). The protocol has been registered on OSF Registries: 10.17605/OSF.IO/J3CVR.

We selected Facebook groups to gather oral dermatology cases, adapting the selection process from D'Souza et al. [15]. We searched for visible groups (accessible via Facebook search) using the keywords ‘dentiste' (dentist) and ‘dermatologie buccale' (oral dermatology), and excluded non-French-speaking groups, those unrelated to dentistry or oral dermatology, and groups with fewer than 5000 members, as smaller groups are less likely to be discovered by users and thus deemed less relevant for our analysis. After joining, we screened for eligibility by excluding inactive groups, non-French-speaking groups, those closed to new members, and those with few clinical cases, specifically groups that posted less than one case per month in the last six months. This process resulted in the selection of all active oral dermatology groups for case inclusion.

We could not estimate the number of dermatology cases to be included or the impact of calendar parameters like holidays and seasons, so we established a 6-month collection period (from November 13, 2023, to May 13, 2024) to maximize data and minimize bias. We searched the ‘Media' section of Facebook groups for dermatology case photographs. Each post was recorded and rated using a custom communication criteria scoring grid. Only original posts were included, despite Facebook allowing edits after publication. The scoring grid was designed for this study, based on existing work on clinical and dermatology case presentations. It included criteria such as (a) clinical case presentation (reason for visiting, medical questionnaire, history, extraoral and intraoral examination, photograph, diagnosis) [16], and (b) dermatology case presentation (size/number of the lesions, outline, surface, base, color, consistency, lesion history) [17–20]. Table I details the scoring criteria and methods (see Supplementary Materials for more details). Scoring accounted for missing points, with each criterion rated as missing (1 point), present (0 points), or non-applicable (NA), aiming for the lowest communication score. If multiple lesions were present, each was evaluated separately. The main outcome was the global score; secondary outcomes included individual criterion scores, number of images, text length per publication, and whether the patient had been seen by the posting dentist. These additional data aimed to limit bias.

Qualitative variables were described in terms of frequencies and percentages. Quantitative variables were described in terms of mean and standard deviation for normally distributed data and median and quartiles for non-normally distributed data. The normality of distributions was assessed visually. Comparisons between two elements were conducted using the Student's t test for normally distributed data and the Mann–Whitney test for nonnormally distributed data. Correlations were assessed using the Pearson correlation coefficient and its associated test. Statistical analyses were conducted using GraphPad Prism v7.0 software (GraphPad Software, San Diego, CA, USA). The significance level was set at 5%.

Results

On November 10, 2023, we identified 133 closed and 76 public groups with the ‘dentiste' search, and 2 closed and 56 open groups with the ‘dermatologie buccale' search. After screening titles and content, we included 2 groups (Fig. 1). The first, ‘Dentistes de France' (28,000 members), is highly active (1000 posts per month) with diverse content, including a small proportion of oral dermatology cases; we refer to it as the All-Dentistry group. The second, ‘Pathologie et dermatologie des muqueuses buccales' (7000 members), is less active (50 posts per month) but focuses entirely on oral dermatology; we refer to it as the Oral Dermatology group. Both are closed groups requiring approval to join. The All-Dentistry group requires a request with questions about the user's dental background, though these questions were not always mandatory, allowing some members to join without verification. Both groups have moderators (15 and 3, respectively) but no content rules (Tab. II).

We gathered data published between November 13, 2023, and May 13, 2024, including 84 cases from the All-Dentistry group and 158 cases from the Oral Dermatology group, totaling 242 cases.

The mean score for missing points across both the All-Dentistry and Oral Dermatology groups was 6.21 ± 2.01, with scores ranging from 1 to 12. On average, 83% of patients were seen or likely seen by a dentist, 8% were not (photo sent by the patient, family member, or doctor), and 9% were undetermined.

All-Dentistry group scored more missing points (7.00 ± 2.09) than the Oral Dermatology group (5.80 ± 1.84, p < 0.001). Specifically, there was a significant difference between the groups on the Surface (mean = 0.39 ± 0.49 vs. mean = 0.22 ± 0.42, p = 0.008), Outline (mean = 0.23 ± 0.42 vs. mean = 0.10 ± 0.30, p = 0.009), History (mean = 0.69 ± 0.47 vs. mean = 0.37 ± 0.48, p < 0.0001), Lesion History (mean = 0.60 ± 0.49 vs. mean = 0.37 ± 0.49, p = 0.001), and Size/Number (mean = 0.17 ± 0.38 vs. mean = 0.06 ± 0.23, p = 0.005) criteria. The All-Dentistry group had a higher mean score on all criteria than the Oral Dermatology group, and displayed more heterogeneity, with the lowest and highest scores (Tab. III, Fig. 2).

Fig. 1 Flowchart of the Facebook groups inclusion.

Fig. 2 Scoring grid results for both groups. (a) Radar graph of all communication scores for each criterion. 0 means the presence of the point and 1 the absence. (b) Box-plots of average total score, comparing the All-Dentistry group and the Oral Dermatology group. *p < 0.05; **p < 0.01; ***p < 0.001.

Discussion

Facebook groups analyzed

Out of 287 groups evaluated, only two met the inclusion criteria: the All-Dentistry group and the Oral Dermatology group. It is unlikely that significant groups were missed or that those declining access would have qualified. This underscores the rarity of shared oral dermatology cases and the importance of detailed analysis of these two groups, showing that such cases are published in very few groups.

For the combined set of both groups, the mean score for missing points was 6.21 ± 2.01 on a thirteen-criteria grid. Cases with 8 or more missing points are considered poor quality, but defining a quality threshold is difficult because some criteria were rated NA, lowering the final score. The number of missing points ranged from 1 to 12, indicating significant variability in the quality of online clinical cases. Consequently, while Facebook groups can help users learn from discussions about oral dermatology cases, they are not sufficient for mastering the specialty without prior foundational knowledge. Users may pick up both good practices and bad habits from these posts.

The Oral Dermatology group, being smaller and more focused, performed better in sharing cases, possibly due to increased public interest and training in this field. Neither group had rules for sharing cases, resulting in significant variability in case quality. Implementing additional rules could improve the quality of clinical cases. However, the entrance questions for the All-Dentistry group did not positively impact the quality of cases shared.

An interesting aspect is that the two groups are not strictly separate, with some members belonging to both. During the study, 242 cases were published across both groups, with seven appearing in both groups simultaneously or after being advised to post in the other group. These duplicate cases were evaluated in both groups, receiving the same grade overall. However, one of the cases included additional information when reposted in the Oral Dermatology group by a different member, resulting in a more favorable evaluation in that group. The high number of cases helps reduce bias and cherry-picking, despite some overlap. Notably, one publication was more detailed in the Oral Dermatology group than in the All-Dentistry group.

Scoring grid criteria

Some criteria were not scored, despite their clinical utility. Histological exams were excluded since most shared cases focused on initial measures rather than confirmed diagnoses, and radiography is often irrelevant for oral dermatology lesions [21,22]. We also did not score confirmed diagnoses, as those seeking opinions rarely share biopsy results. However, we required a diagnostic hypothesis, as it is central to the hypothetico-deductive method of diagnosis.

The best possible score is 0, while the worst is 13. However, using ‘NA' as 0 in the total score can make the worst score lower than 13. Some criteria were not evaluable in certain cases, potentially leading to misinterpretation of low scores as good. For instance, there were 141 non-evaluable criteria, often due to the ‘Base' criterion, which was only applicable if lesions were on the lips, tongue, cheeks, or soft palate. The reason for visit could be rated NA as well for patients who were not seen in consultation (for example, if they merely sent a photo to their dentist). Thus, it is more accurate to interpret the total score as the number of points missing rather than present.

The history criterion was challenging to evaluate since dentists rarely note what the patient lacks (e.g., no allergies). This made it difficult to distinguish between missing information about a condition and its absence. We expected dentists to specify if no other relevant elements existed beyond what was mentioned.

Scores

The results indicate three categories of criteria: frequently provided (color, photograph), rarely provided (reason for visit, extraoral exam, intraoral exam, base, consistency, diagnostic hypothesis), and better shared in the Oral Dermatology group (history, size, outline/number, surface, lesion history).

The History criterion was better in the Oral Dermatology group (p < 0.001), indicating that dentists provided more context to aid diagnosis. However, since dentists may deem it unnecessary to report a lack of peculiarities, this might suggest that patients in the All-Dentistry group had fewer comorbidities.

The radar graph (Fig. 2) shows that the All-Dentistry group consistently has higher averages than the Oral Dermatology group across all criteria. This trend indicates that publications in the All-Dentistry group have more missing points for each criterion.

While descriptions are often insufficient, photos help achieve better results. Sometimes posts lack descriptions, but clear photos allow validation of criteria like size/number, outline, surface, color, and photo quality. More photos are published on average in the specialized group compared to the general dentistry group (2.32 ± 1.52 vs. 1.90 ± 1.26), suggesting that specialized dentists recognize the importance of images. Some dentists may rely on clear photos rather than detailed descriptions as a more convenient communication method. As noted in Cura's study, medical photography is becoming common [23]. American dermatologists report a usage rate of photography of 99% [24]. However, students are taught not to rely solely on photos for diagnoses, as additional information like consistency is also relevant. Dentists might assume that clear photos make further textual descriptions unnecessary.

Some dentists merely forward photos sent to them without seeing the patient, which may account for the often brief descriptions. Sometimes dentists indicate they haven't seen the patient, or clues like the absence of gloves or casual settings suggest the lack of a consultation. Consequently, information not conveyed by the photo, such as consistency or patient history, cannot be included in the text.

One of the lowest-scoring criteria is the reason for consultation, with an average score of 0.84 ± 0.37. According to Hartman et al., mentioning the reason for encounter aids in diagnosis, making it a crucial part of a clinical case description [25]. While dentists might assume the visit is for a painful lesion, it could be a minor issue that the patient barely mentions. Conversely, a patient may be unaware of a painless lesion, visiting for another reason. Regardless, the reason should be included in the description as it is essential for medical records.

The extraoral exam criterion is the lowest-scoring, with an average of 0.94 ± 0.24, and is rarely mentioned. This omission may result from the exam being forgotten or from finding nothing relevant during the examination.

The size and number criterion is rarely omitted, with an average score of 0.10 ± 0.30. This may be because these details can be inferred from the photo, allowing readers to estimate them even if not explicitly stated by the dentist.

It is interesting that the community group and the specialized group showed a highly significant difference, likely due to their purposes: the specialized group consists of specialists or those seeking specific answers in oral dermatology, while the community group includes a broader range of dental professionals who may excel in other fields.

Additional remarks

We chose not to analyze comments for additional information due to time constraints and complexity. Determining an appropriate timeframe and navigating intricate responses would be challenging. While updates from dentists on biopsy results would be valuable, it is not common practice for them to share these results.

We did not consider whether the original posts were modified, as determining a timeframe for tracking modifications would be problematic.

Cases without at least one photo of the lesion were excluded, which may introduce bias. However, in dermatology, it is uncommon to share cases without photos, as describing a dermatological lesion with text alone is challenging [13].

Facebook is an aging social network compared to newer platforms like TikTok and Instagram [26]. This may bias the study toward older dentists whose academic training is more distant. The dentists whose posts we analyzed are active Facebook users who have time to publish. They might lack nearby colleagues proficient in dermatology or may not be skilled in it themselves, otherwise they wouldn't seek advice online. It's also possible they aren't dentists at all, given the easy admission process to these groups, which only requires stating one's profession without proof (Tab. II).

We must also consider the legal framework, as it is unlikely that patients have consented to having their photos published online. In France, and likely other French-speaking countries, posting patient photos without consent is illegal.

Additionally, some photos were sent directly by patients or acquaintances. A report indicated that 74.5% of respondents had received a photo from a patient, with only 47.4% sent through secure messaging systems [24]. This highlights the need for secure communication channels between health professionals and patients.

In some cases, anonymization was not respected by the original author, as patient names were visible on radiographs.

Conclusion

The study's findings from two representative Facebook groups reveal significant variability in the quality and completeness of clinical cases shared by French dentists, particularly in oral dermatology. While photographs play a central role in case analysis, the inconsistencies in case details suggest the need for clearer guidelines to ensure more comprehensive and reliable information sharing. The study also highlights the legal and ethical challenges of patient consent and data anonymization, which are often overlooked. To enhance the educational value of these groups, contributors should follow standardized criteria for case presentation, including detailed descriptions alongside photographs. To enhance the educational value of these groups, contributors should follow standardized criteria for case presentation, including detailed descriptions alongside photographs. Moreover, the groups should establish publication rules and strengthen moderation to ensure higher standards of content quality. Implementing secure, professional communication channels and strictly enforcing privacy regulations would further improve case quality and protect patient confidentiality.

Funding

No funding was received for this article.

Conflicts of interest

The authors have no conflicts of interest to declare.

Data availability statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Supplementary Material

Supplementary material 1: Criteria for a clinical case presentation.

Supplementary material 2: Complete description of the scoring grid.

References

All Tables

All Figures

	Fig. 1 Flowchart of the Facebook groups inclusion.
In the text

	Fig. 2 Scoring grid results for both groups. (a) Radar graph of all communication scores for each criterion. 0 means the presence of the point and 1 the absence. (b) Box-plots of average total score, comparing the All-Dentistry group and the Oral Dermatology group. *p < 0.05; **p < 0.01; ***p < 0.001.
In the text