Introduction

The Internet is a popular source of health care information. In the United States, an estimated 95% of adults used the Internet in 2024, and 58.5% of adults reported using the Internet to access health or medical information [1, 2]. In a 2017 study, Americans reported using a vast number of Internet sources to obtain health-related information. These sources included commercial websites (71.8%), search engines (11.6%), academically affiliated sites (11.1%), and government-sponsored sites (5.5%) [3].

Traditional sources of information include consulting medical professionals. However, online platforms have become increasingly popular for patient education due to their accessibility and visual appeal [4]. YouTube is the most visited social media site in America, with an estimated 83% of the population visiting this site [5]. This platform is becoming increasingly more popular for accessing health-related information. The vast number of uploaded videos allows users to cross-check multiple videos to subjectively determine which information is reliable. A 2024 study found that out of 3,000 participants, 87.6% watch health-related content (HRC) and 40% of the users watch HRC to decide whether to consult a doctor or adopt certain practices [6]. Although videos are mainly verbal, they also contain a large portion of written information. Thus, the concept of health literacy becomes essential. The National Institutes of Health (NIH), the U.S. Department of Health and Human Services (HHS), and the American Medical Association (AMA) all recommend that patient education materials be written at or below a sixth-grade reading level to enhance comprehension and usability among the general population. The recommendation comes from research demonstrating that materials exceeding this readability threshold can hinder understanding and patient engagement [7]. Patient education plays a pivotal role in improving health outcomes by empowering individuals to understand and actively participate in their care. When patients understand the cause of their disease, as well as treatment options, they can participate in shared decision-making and comply with physician recommendations. This contributes to better disease management, reduced complications, and improved quality of life [8].

Despite the increasing use of online resources, there is an array of both high-quality and low-quality health-related videos available for users to access. Video quality according to users may be evaluated using criteria such as credibility of the source and overall content usefulness [9]. As a result, users are left to decide which information is reliable. Previous studies on the educational accuracy of YouTube videos have been conducted on orthopedic conditions and topics such as hip and knee arthritis, lumbar discectomies, articular cartilage defects, and more [10–15]. Results from these studies indicate that videos uploaded to YouTube have poor educational content. To help combat this issue, YouTube Health was recently launched with the goal of helping users find health information from authoritative sources. Currently, it is unknown whether the quality of verified videos is higher than that of unverified videos. To address this problem, this study assessed the educational content quality and reliability of YouTube Health verified and unverified videos, specifically osteochondritis dissecans (OCD), because it is a common cause of joint pain among children and necessitates early diagnosis and management.

OCD is a form of osteonecrosis that primarily affects school-aged children and adolescents. The highest incidence occurs between the ages of 12 and 19, and males have a 2–4 times higher incidence than females. However, the exact incidence of OCD is unknown [14]. The prevalence of OCD is between 9.5 and 29 per 100,000. The pathophysiology of OCD begins with the disruption of epiphyseal vessels, which causes ischemia and necrosis at the site of injury. The disease is most commonly due to repetitive trauma, but can also occur with isolated injury. Patients typically present with symptoms of poorly localized knee pain that worsens with physical activity. Patients may also experience stiffness and a locking sensation. The initial assessment of OCD involves plain radiographs to locate the lesion and rule out other causes. MRI is the most valuable diagnostic imaging technique used to diagnose OCD when plain radiographs are negative, but the clinical suspicion remains high. The treatment has either a conservative or surgical approach, depending on the patient’s age and the severity of the disease. Nonoperative treatment is considered for patients without a displaced fragment, whereas operative treatment is recommended if conservative measures are ineffective for 3–6 months. The prognosis of OCD is generally more favorable for juvenile patients compared to adults. If left untreated, adults with OCD can develop arthritis. Similar studies on higher-volume orthopedic topics such as the anterior cruciate ligament, Achilles tendon, or knee osteoarthritis have been conducted, so the authors of this study chose OCD because, although less commonly discussed, it is still a clinically significant condition with limited patient education resources available online. Thus, it is essential to evaluate the accuracy and quality of the limited existing resources to ensure patients are adequately informed of the pathophysiology and treatment of the disease.

Materials and methods

This study utilized a cross-sectional observational design. The authors of this study searched the term “osteochondritis dissecans” on June 3, 2024, in the Midwestern United States. Videos were reviewed and graded independently by three medical students. A Google Chrome incognito tab was used to eliminate any confounding factors influencing the results. The first 50 videos based on the key term were evaluated; this method has been reported as a viable method of video selection because this method of evaluation has been used in previous peer-reviewed literature in orthopedic surgery [15]. Videos included for initial review were limited to the first 50 videos that populated after searching the term “osteochondritis dissecans”. Exclusion criteria included videos in non-English languages, audio-only soundtracks, and OCD videos not relating to humans. In these cases, the next video that did not violate exclusion criteria was used.

Results

52 videos were screened because two videos met the exclusion criteria. One video was excluded because it was about OCD in dogs, and the other video was excluded because it was audio-only. Of the 50 videos regarding OCD, the mean number of views was 6,537 ± 7,974. These videos had a cumulative total of 326,851 views, and video characteristics are listed in Table 2.

The most common information presented was disease-specific information (58%), followed by management techniques (24%), patient experience (10%), and advertisements (8%) (Figure 1). The most common source of videos was physicians (56%), followed by academic sources (18%), medical sources (14%), and commercial (12%) (Figure 2).

Display full size

Figure 1. Relative frequencies of video content for OCD-related YouTube videos. OCD: osteochondritis dissecans.

Display full size

Figure 2. Relative frequencies of video uploader sources for OCD-related YouTube videos. OCD: osteochondritis dissecans.

Discussion

The first 50 videos had a cumulative total of 326,851 views, a mean of 6,537 views, and a range of 7 to 40,528 views when the search was conducted. Previous studies on the quality of YouTube content on orthopedic content, such as hip and knee arthritis, lumbar discectomies, and articular cartilage defects, were viewed more times and had a higher average number of views per video [17]. This suggests that there is less content regarding OCD on YouTube compared to more mainstream orthopedic conditions, such as posterior cruciate ligament injuries and Achilles tendon injuries. This may be attributed to several factors. There are limited diagnostic, prognostic, and therapeutic studies regarding OCD. OCD is a relatively uncommon disease, and the appropriate treatment is largely unknown. Thus, there may be limited educational content available for the public because of the lack of sufficient evidence.

The authors of this study found that the accuracy and reliability of YouTube videos regarding osteochondritis is low. The mean JAMA benchmark criteria score was 2.28 out of 4, which suggests the reliability of individual YouTube videos is moderate to low. The mean GQS was 2.60 out of 5. This suggests that the nonspecific educational content quality of YouTube videos is low, and viewers are not likely to be adequately informed about OCD based on the video. The mean OCDSS was 5.02 out of 11. This suggests that the specific educational content in individual videos is not comprehensive to inform viewers of the pathology, symptoms, and treatment options. These findings of low-quality YouTube videos regarding orthopedic conditions are similar to previous studies [18]. McMahon et al. [18] found that the mean JAMA benchmark score was 2.69, the mean GQS was 2.64, and the mean Achilles tendon specific score (ATSS) was 4.66.

Although physician and academic sources, as well as disease-specific videos, appeared to produce higher scores, these associations were only statistically significant in univariate analyses. Specifically, the uploader source was significantly associated with higher JAMA scores, and the content category was significantly associated with higher GQS and OCDSS. However, once we adjusted for other variables in multivariable regression, these differences no longer remained significant. This suggests that factors such as video length and verification status are more important predictors of reliability and quality. Longer videos consistently achieved higher scores across all three evaluation systems. Video popularity, as measured by views, likes, comments, and video power index, was not significantly associated with any scoring system. The average number of views per video was high, but this was driven up by a few highly popular videos. The median view count was lower, reflecting that most videos had relatively low to moderate engagement. The disconnect between popularity and quality highlights a potential risk for patients, as highly viewed videos are not necessarily the most reliable or comprehensive sources of information. It is reasonable to assume users who search for health content on YouTube rarely scroll past the first few pages of videos that populate when they enter their search term. Users look for relevant videos based on characteristics such as title, viewer engagement, or content. The results of this study suggest that users should watch videos based on video length and YouTube Health verification status rather than looking at the title or who uploaded the video.

When comparing YouTube Health verified versus unverified videos, verified videos demonstrated slightly higher mean JAMA, GQS, and OCDSS scores. However, these differences were not statistically significant in unadjusted analyses. Engagement metrics including views, likes, comments, and video power index, also did not differ significantly between groups. Notably, verified videos were significantly shorter in duration than unverified videos. In multivariable regression, YouTube Health verification was a statistically significant independent predictor of higher JAMA reliability scores, but it did not significantly predict GQS or OCDSS. Verification status was not randomly distributed across the dataset. Verified videos were disproportionately uploaded by academic sources. While this discrepancy exists, it is a positive sign that YouTube Health is verifying videos from trustworthy sources, such as Stanford Medicine, Scottish Rite Hospital, and board-certified physicians. Likewise, verification was more common in patient experience and disease-specific information videos. These findings suggest that YouTube Health verification may serve as a meaningful marker of reliability, particularly for features such as authorship, disclosure, and transparency. However, verification alone does not guarantee higher overall quality or content comprehensiveness. In conclusion, these results indicate that while verification is a step toward improving the reliability of HRC on YouTube, additional measures may be required to enhance educational quality and comprehensiveness for patients.

For viewers to be accurately educated through YouTube, we believe videos should address points addressed by the American Academy of Orthopaedic Surgeons because this data is peer-reviewed. This is reflected in the OCDSS we created. Since scores suggest low-quality educational content and reliability, we recommend physicians to counsel patients on the low-quality YouTube videos regarding OCD and provide patients with accurate information and enable patients to take more control over their care.

While YouTube continues to be one of the most frequently accessed platforms for obtaining health-related information through videos, the landscape of online health education has expanded with the development of AI-powered conversational models and specialized web-based platforms. YouTube primarily provides user-generated video content, and the quality of information depends on the expertise of the content creator. In contrast, AI tools such as ChatGPT, Gemini, and Perplexity are interactive and generate real-time responses to user queries. These platforms utilize large language models to synthesize relevant information and provide readable health explanations. However, concerns remain regarding content reliability, accuracy, and source transparency. A recent study comparing AI systems found that although responses were easily readable, there was variability in accuracy and reliability [7]. Prior analyses of traditional online resources revealed that web-based education materials vary in quality and reliability. This demonstrates the ongoing challenge of ensuring online resources that are available for anyone to access are accurate and reliable.