A Retrospective Analysis of Dermatologic Manifestations of Diabetes on TikTok

Abstract

This study evaluated 166 TikTok videos about three skin conditions associated with diabetes. It found that the quality of health information in these videos significantly differed based on the five creator types making the videos (person with diabetes, caregiver/parent, noncaregiver, health care professional, or health coach). There were also significant differences in the understandability and actionability of the health information provided. There is a need within the diabetes online community for higher-quality information from reliable sources regarding dermatological conditions associated with diabetes.

According to the Centers for Disease Control and Prevention, 38 million people in the United States had diabetes in 2023 (1), and the number of adults diagnosed with diabetes has more than doubled in the past 20 years (1). Diabetes can cause complications, including neuropathy, retinopathy, and nephropathy. Additionally, dermatological manifestations may arise in people with diabetes, including necrobiosis lipoidica (NL), granuloma annulare (GA), and neuropathic (diabetic) ulcers (NUs).

NL is a rare, granulomatous condition that is highly associated with diabetes and diabetes risk factors, including obesity, hypertension, and hyperlipidemia (2,3). The prevalence of NL in people with diabetes ranges from 0.3 to 2.8%; the reported prevalence of diabetes among people with NL is much higher, at ∼50% (4–6). Research on NL has been limited, but published cross-sectional studies have summarized the treatment options, treatment efficacy, comorbidities, and clinical features of the condition. This chronic disorder typically presents as red papules that can develop into discolored plaques or patches ranging from red to yellow-brown. Lesions are typically present in multiples and can be painful or itchy, ultimately leaving scarring, sclerosis, and continued inflammation, increasing the risk of future infection (2,3). Treatment options have demonstrated varied efficacy and include, but are not limited to, phototherapy, topical steroids, and calcineurin inhibitors (3,6).

GA is a poorly understood inflammatory disease characterized pathologically by ring-like, erythematous plaques (7,8). This condition is also very rare, with annualized prevalence and incidence rates of 0.04 and 0.06%, respectively, and most commonly occurs in middle-aged women (9). GA is strongly associated with preexisting diabetes, hypertension, and hyperlipidemia, but other risk factors include hypothyroidism and rheumatoid arthritis, with results suggesting that GA may be mediated by CD4+ T-cells (7). Research indicates that the main treatments for GA include topical corticosteroids, phototherapy, intralesional corticosteroids, oral tetracyclines, and hydroxychloroquine (8,9). Treatment regimens have been noted to be patient and subtype dependent, with some individuals having only localized lesions and others having much larger and more severe spreading (8).

NUs, also known as diabetic ulcers, are highly morbid lesions that extend into the dermis of the skin; they are a common complication of diabetic neuropathy, a classic finding in people with diabetes (10,11). An estimated 19–34% of people with diabetes will develop an NU, and of those, 20% will require amputation (11). In cases of lower-extremity amputation, >80% of those with diabetes had a preceding foot ulcer that either became infected, which happens in 50–60% of diabetic ulcers, or caused extreme deformity of the foot (10,12). Research indicates that factors such as chronic or end-stage kidney disease, peripheral artery disease, older age, history of amputation, and history of cardiovascular disease all increase the risk of mortality in relation to NUs (10–12). Diabetic neuropathy is highly associated with the formation of NUs, and treatments for their underlying pathology lack efficacy. Current modalities focus on glycemic control, lifestyle modification, and adequate podiatric care to help prevent ulcers from developing (10). However, once an NU is formed and fails to heal, treatment options include surgical debridement or amputation (11,12). Diagnoses of NUs have been associated consistently with feelings of anxiety, fear, and sadness, but few studies have assessed psychosocial quality of life in relation to this condition (11,12).

People with diabetes often seek information from the diabetes online community (DOC), part of a phenomenon labeled “peer-to-peer health care” that has been identified as a key coping strategy for diabetes (13). The DOC began with online blog sites and now can be found on all major social media platforms. Its members seek or offer health care advice and seek social support.

TikTok is a social media platform that has been gaining popularity. It features video posts that can be used in various ways, from physicians providing education to the public to individuals sharing their own experiences of having diabetes. The ever-increasing prevalence of diabetes, in combination with the ∼53% of the world’s population who use social media, translates to an abundance of videos posted online about diabetes (1,14).

This study aimed to identify and analyze TikTok video content about NL, GA, and NUs and specifically to assess the quality of the health information on dermatological manifestations of diabetes and the credibility, reliability, understandability, and actionability of the available information on these conditions.

Research Design and Methods

The digital application Apify (apify.com), a TikTok content scraper, was used to gather all the videos within the time frame of December 2022 to December 2023 with the following hashtags: #necrobiosislipoidica, #necrobiosislipoidicadiabeticorum, #granulomaannulare, #neuropathiculcer, and #diabeticulcer. This app created a Microsoft Excel spreadsheet containing key details about the TikTok videos with any of these hashtags. Each of the authors of this article was assigned ∼45 TikTok videos to code using Qualtrics, a tool for creating online surveys.

A Qualtrics form was created in which to input data for each video. Three TikTok videos on the topic of acanthosis nigricans were used to practice answering the Qualtrics survey questions. The survey questions were based on DISCERN (15), a short instrument to assess the quality of health information, and the Patient Education Materials Assessment Tool for Audiovisual Materials (PEMAT-A/V) (16), an instrument to evaluate the understandability and actionability of the videos (Table 1).

Table 1.

Qualtrics Coding Items

Your initials Video ID TikTok User ID  Video Web URL How many likes does this video have? Does the content creator mention having an autoimmune condition? Does the content creator mention that this skin condition was caused by an autoimmune condition? Who is the video creator? Does the content creator live with diabetes? Please describe the video content. What type of diabetes is this content about? Which skin condition is this content about? DISCERN items PEMAT-A/V items

A total of 230 TikTok videos were compiled in the Apify-generated spreadsheet. Eight duplicate videos were removed. Coders noted any videos that were not in English, unavailable with the provided link, or irrelevant (i.e., videos that used one of the dermatological hashtags from the study but did not discuss the condition). Sixty-four videos were excluded, and 166 were included in the final data analysis (Figure 1).

Figure 1.

Flowchart of the methods in the study, including the total number of videos scraped by Apify with the study hashtags used, the number of videos excluded, and the number of videos included for analysis.

Statistical Analysis

The 166 included videos were categorized by creator type as general user (subcategorized as person with diabetes, caregiver/parent, noncaregiver, no relation to diabetes, or other), health care professional (HCP), or health coach. The proportion of users classified in each respective category was then compared using a three-sample proportions test. Likewise, the content within the videos was categorized as misleading/misinformation, useful tips/tricks, personal anecdotes, or comedy/humor. The videos were also compared based on the dermatological skin condition discussed in them. These categories were compared using a four-sample proportions test.

Video content was analyzed using three separate measures. The DISCERN survey assessed the quality of the health information. The PEMAT-A/V understandability and actionability surveys evaluated comprehension and application of the information presented. Three separate one-way ANOVA tests were conducted to examine the variability of the DISCERN and PEMAT-A/V tools across creator types. The subcategories “no relation to diabetes” and “other” were removed from the ANOVA tests because they only had a sample of one each. The significance level for all tests was set at α = 0.05. The data were analyzed using RStudio, v. 2023.03.1+446.

Results

Of the 166 included videos, 117 (70.5%) were created by general TikTok users (Table 2). This was greater than the other two major categories of HCP and health coach, which had 38 (22.2%) and 11 (6.4%) videos, respectively. The proportion of video creators that were general users, HCPs, and health coaches varied significantly (P <0.001).

Table 2.

Types of Users

TikTok User Type	n (%)
General user
Person with diabetes	23 (13.5)
Caregiver/parent	34 (19.9)
Noncaregiver (e.g., sibling of someone with diabetes)	58 (33.9)
No relation to diabetes	1 (0.6)
Other	1 (0.6)
Health care professional	38 (22.2)
Health coach	11 (6.4)

Most of the content within the videos was either useful tips/tricks or personal anecdotes (Table 3). Approximately 8.2% of all videos contained some misleading information or misinformation. The videos varied significantly by video content (P <0.001).

Table 3.

Classification of Video Content

Description of Video Content	n (%)	P
Misleading/misinformation	14 (8.2)	<0.001
Useful tips, tricks, or information	78 (45.6)
Personal anecdotes	67 (39.2)
Comedy/humor	7 (4.1)

The median number of hearts, or “likes,” the videos received was 18. The overall range of hearts was (0–6,685). The videos received a mean of 176.3 ± 715.5 hearts.

The majority of videos contained information on NUs (Table 4), which were covered in 97 of the 166 videos (56.7%). The second largest group was on GA; this topic was covered in 52 videos (30.4%). The proportion of videos varied significantly by dermatological condition (P <0.001).

Table 4.

Breakdown of Videos by Type of Dermatological Condition

Video by Skin Condition	n (%)	P
NL	12 (7.0)	<0.001
GA	52 (30.4)
NUs	97 (56.7)
Other	5 (2.9)
Total	166 (100.0)

The reliability and validity of the videos varied significantly by creator type (Table 5). The quality of health information, as determined by DISCERN, was significantly different across the five creator types (F = 30.1, P <0.001). There was also a significant difference in understandability of health information (F = 4.0, P <0.05) and in actionability of health information (F = 10.7, P <0.01) across the five creator types.

Table 5.

Evaluation of Video Reliability and Validity Using DISCERN and PEMAT-A/V Assessments Across Creator Types

Scoring	Mean ± SD	F	P
DISCERN	27.51 ± 10.13	30.1	<0.001
PEMAT-A/V understandability	58.11 ± 29.79	4	<0.05
PEMAT-A/V actionability	33.33 ± 38.96	10.7	<0.01

Discussion

With the escalating prevalence of diabetes in the United States, there is a heightened incidence of diabetes-related complications, including the dermatological conditions of NL, GA, and NUs. These conditions vary in prevalence and severity, necessitating a spectrum of treatment modalities ranging from topical therapies to more invasive interventions such as surgical procedures or, in severe cases, amputations. This study delved into the evaluation of TikTok content pertaining to these dermatological manifestations of diabetes, underscoring the crucial need to assess the quality and reliability of health information disseminated through social media platforms for the benefit of patient education and support.

The analysis encompassed 166 TikTok videos, revealing significant disparities in content quality based on creator type. General users constituted the majority of creators (70.5%), with HCPs and health coaches contributing 22.2 and 6.4% of videos, respectively. The study found that videos by HCPs generally provided more reliable and actionable health information compared with videos by general users (P <0.001). The understandability of information significantly differed across creator types (P <0.05). These findings suggest the variability in health content on TikTok and highlight the need within the DOC for validated information sources.

This study followed a systematic approach using the Apify platform to collect and analyze videos, ensuring a comprehensive sample of relevant content. Using established tools such as DISCERN and PEMAT-A/V allowed for objective evaluation of content quality, understandability, and actionability. Moreover, the study’s focus on dermatological conditions associated how these specific health issues are addressed on social media platforms.

Future research should explore ways to enhance understanding and use of social media for health communication in diabetes. Specifically, investigating the use of longitudinal trends in TikTok health content related to diabetes and dermatological conditions will show how content evolves over time. This technique will also quantify the dissemination of accurate health information and the prevalence of misinformation. Additionally, investigating the impact of health education campaigns will assess the effectiveness of such campaigns on TikTok, whether initiated by HCPs or health care organizations. Such research into improving knowledge and behaviors related to diabetes and dermatological conditions could also tie into user engagement and behavior, which can be explored by assessing factors influencing interaction rates, such as video length, format (e.g., personal anecdote vs. educational video), and creator credibility. Future research should also examine other diabetes comorbidities and additional social media platforms.

This retrospective analysis emphasizes the need for accurate and accessible health information on dermatological manifestations of diabetes on platforms such as TikTok. Disparities in content quality across different creator types emphasize the importance of integrating evidence-based guidelines into social media health discourse. Moving forward, collaborative efforts among HCPs, social media platforms, and the DOC are essential to enhance public health literacy and empower individuals affected by diabetes-related skin conditions to make informed decisions about their health and well-being.

Limitations

The authors originally discussed including acanthosis nigricans (AN) in the initial compilation of dermatological hashtags. However, AN is more commonly known and resulted in generating numerous videos. We opted to include only NL, GA, and NU, which are not as common as AN and therefore generated a more manageable number of videos.

Other limitations include the reliance on publicly available videos, which may not fully represent all TikTok content. The study also focused on English-language videos, potentially excluding valuable non-English resources. Furthermore, the cross-sectional nature of the study limits causal inference, and the subjective nature of content evaluation introduces potential biases.

Conclusion

With the increase in diabetes prevalence, there also has been an increase in complications, including those affecting the skin. TikTok and other social media platforms may be used by the DOC to gather information. There is a need within the DOC for higher-quality information from reliable sources regarding dermatological conditions associated with diabetes.