This exploratory study aims to increase the understanding of the acute low back pain (ALBP) content available on YouTube (Google Inc) in 2020 and to establish the plausibility of using a simple checklist to facilitate the discovery of YouTube content consistent with current guidelines. Low back pain exerts a high economic and social burden across the globe. The 2018 Global Burden of Disease study suggested that low back pain was the leading cause of years lived with disability in most countries [1].

This paper focuses on ALBP. ALBP is commonly described as a new onset or exacerbation incident lasting less than 12 weeks and including sciatica [2,3]. Most people with ALBP have rapid improvements in pain and disability within several weeks. In most cases, the cause of low back pain cannot be identified, and most low back pain is therefore described as nonspecific [4,5]. However, pain and disability persist for a proportion of people. Up to 80% of people with ALBP may experience recurrence within 12 months [6,7]. The estimates of recovery from an episode of low back pain over 12 months range between 54% and 90% [8,9]. Differing definitions of recovery also complicate the epidemiology of ALBP. Stanton et al [10] described the heterogeneity of case definitions of acute exacerbations of preexisting back pain as further contributing to the diverse understanding of the scale and outcomes of ALBP.

Access to guideline-consistent effective care for ALBP is generally regarded as limited in the scientific literature. We have described access to ALBP care based on a framework developed by Aday and Anderson [11]. Aday and Anderson’s 4 dimensions of access are characteristics of the health delivery system, utilization of health services, characteristics of the population at risk, and consumer satisfaction. These are described below.

Most people manage low back pain with little assistance from health care providers. Estimates of the number of individuals that manage back pain independently, or with occasional formal health care, range from 50% to 70% [24,25]. In scientific literature, the terms self-management and self-care are often used interchangeably. Self-management is generally regarded as a clinician-guided collaborative intervention that enhances an individual’s capacity to monitor and manage their own physical and emotional responses and maintain their quality of life [26,27]. The related term self-care generally refers to actions and decisions taken independent of health providers [28]. Importantly, self-management and self-care are not passive processes. Rather, these processes involve active patient decision making, including symptom monitoring, goal setting, information search and interpretation, and self-efficacy [29,30]. In the case of ALBP, psychosocial status, including fear avoidance, self-efficacy [31,32], and catastrophizing [33], may contribute to the transition from an acute to a chronic condition. Therefore, access to psychosocial support is a consideration in ALBP self-management interventions.

Digital health interventions (DHIs) may improve access to guideline-consistent ALBP care by facilitating self-management decision making. DHIs include all interventions delivered via digital technologies that facilitate health behavior change [34]. This includes web search, social media, symptom checkers, apps, and telehealth. The reported access benefits of DHIs include reduced health system costs, patient waiting, travel time, and expenses [35] and improved patient-provider communication [36], health outcomes [37], scalability, and safety [38].

By 2020, searching the internet was regarded as a routine dimension of individual health self-management. However, access to web-based health information remained more limited among older and low-income people [39]. Researchers have noted further potential problems associated with the feasibility of implementing DHIs at scale. These feasibility problems include acceptability and demand, usability, real-world implementation, and integration with existing practices [40]. In the case of DHIs for ALBP, a 2017 systematic review found no evidence of positive clinical outcomes or cost-effectiveness [41].

Researchers have suggested that increased patient access to health information represents a profound change in the relationship between patients and providers [42]. Digitally enabled independent self-care by patients has not been universally welcomed by health providers [43,44]. Researchers have reported a consistent pattern of health care provider unease when managing patients informed by internet information.

YouTube is one of the world’s most popular websites. In 2020, YouTube was the second most visited website [45] and the second most popular social media network globally [46]. YouTube is commonly used as a source of instructional advice. In 2018, Pew Research reported that 86% of adults in the United States used YouTube for “figuring out how to do things they haven’t done before” [47]. It is the instructional use of YouTube that is relevant to this study.

YouTube has been widely researched as a source of health information. Although researchers have generally described the potential of YouTube as a clinical and patient education resource, they have also noted the heterogeneity of evaluation approaches and variable quality of clinical information. We identified 3 broad research approaches for evaluating YouTube health videos. These approaches are (1) metadata characteristics, (2) information quality review, and (3) expert clinician review. In practice, most YouTube research has incorporated two or more of these approaches.

First, researchers have described the metadata characteristics of YouTube videos exclusive of the content. A systematic review by Sampson et al [48] described that the most common video characteristics included in studies were the number of views, video length, likes, date posted, and language of the video. Similarly, a 2018 systematic review found number of views, video duration, and likes and dislikes to be the top 3 characteristics reported in evaluations of YouTube [49]. Other researchers [50] have suggested that view counts were the second most frequently cited concept in assessing quality on YouTube.

Second, reviews have reviewed the information quality of videos using validated instruments. These instruments are generic health information quality assessment tools, and they commonly describe the credibility of sources and information contained within publications. Commonly used instruments for evaluating the quality of health information incorporated in YouTube videos include DISCERN [51], Brief DISCERN [52], Patient Education Materials Assessment Tool [53], Health on the Net Foundation Code of Conduct [54], Flesch-Kincaid reading level [55], and guidelines from the Journal of the American Medical Association [56].

A third approach involves the evaluation of video content by expert clinicians. Expert clinicians have generally described the potential of YouTube as a self-management resource in positive terms [57]. Similarly, clinician reviewers have consistently noted concerns regarding the discovery of accurate clinical content. Clinician reviewers noted specific concerns about the selection of appropriate search terms [58], including the influences of algorithms [59], and video popularity (views) [60,61] as particular challenges in discovering accurate content.

In summary, the evaluation of YouTube videos is an underdeveloped area. Although there are several approaches to evaluating YouTube content, these are all time consuming, relatively complex, and thus unsuitable for shared care discussions in a clinical setting. Similarly, many patients turn to the internet for self-management advice, although they lack simple cues to identify the content that is consistent with current clinical guidelines.

This study was conducted to increase the understanding of YouTube information about ALBP in advance of potential feasibility and clinical studies. This exploratory study aims to (1) increase the understanding of the ALBP content available on YouTube in 2020 and (2) establish the plausibility of using a simple checklist to facilitate the discovery of YouTube content consistent with current ALBP guidelines. We defined plausibility as “a scenario is one that fits prior knowledge well, with many different sources of corroboration, without complexity of explanation, and with minimal conjecture” [62].

We examined the following 4 research questions: how was the data set defined?, what are the metadata characteristics of the videos in the data set?, what is the information quality of ALBP YouTube videos?, and what are the characteristics of the YouTube data set based on an ALBP self-management checklist?

This study incorporates 3 approaches to evaluate YouTube health content. These approaches are (1) analysis of metadata characteristics, (2) analysis of information quality, and (3) expert clinician review. In this study, we extended these approaches by classifying YouTube content based on author’s professional discipline and substituted a simple checklist for clinician expertise to analyze ALBP self-management content.

We used 2 simple checklists to analyze the YouTube content within a defined data set. Checklists have been widely used in medicine to reduce costs and improve health outcomes [63]. Coding of all items in the data set was conducted by the 2 authors of this study. Intercoder reliability for all coding results was reviewed using Krippendorf alpha [64] (Multimedia Appendix 1).

This is an infodemiology study. Eysenbach described infodemiology as “the science of distribution and determinants of information” [65]. Infodemiology studies have primarily examined public health and policy issues [66]. In contrast, this study examined patterns of YouTube clinical information, with the ultimate aim of developing a novel clinical intervention for the self-management of ALBP.

We examined the metadata characteristics of each video in the final data set. First, we coded the 202 unique videos in the final data set according to the author’s stated disciplinary affiliation. We used the following 6 disciplinary categories: chiropractor, fitness, medical doctor, physiotherapist, yoga, and other (including osteopaths and massage therapists; Table 1 and Multimedia Appendix 4). Researchers have identified relationships between author’s disciplines and user assessments of source credibility. The assessment of web-based source credibility is generally based on rapid evaluation of multiple content features, including visual design [72], trustworthiness and expertise of the source [73], and social cues such as likes and comments [74]. Source credibility is a dimension of user engagement with video content and thus relevant to this study. Second, we coded each video according to the 3 content categories: education, real-time exercise, or real-time treatment. These categories were derived from the videos in the data set. Third, we incorporated YTDT data, including length of video, number of views, and YouTube channel name. Through this approach, we were able to describe the characteristics of the final data set by the author’s professional discipline.

We used a modified Brief DISCERN instrument to assess the information quality of the final data set for this study. The full DISCERN instrument has been widely used in YouTube research [75]. The 6-question Brief DISCERN was designed to be a simpler version of the full DISCERN for patient and clinician use [76]. The Brief DISCERN has been used to evaluate the quality of web-based health content [77]. However, we were not able to identify previous peer-reviewed research using the Brief DISCERN for YouTube analysis.

To analyze information quality, we first modified the Brief DISCERN instrument. We added an ALBP self-management–specific codebook to the original 6 items in the Brief DISCERN instrument (Multimedia Appendix 5). Second, we coded each video in the final data set using our modified Brief DISCERN instrument. Each video was coded as yes or no only. Third, we organized results by authors’ disciplines. Thus, we used the unvalidated modified Brief DISCERN to analyze the information quality of videos in the final data set by the authors’ discipline.

We analyzed the final data set based on the checklist of ALBP self-management strategies that we developed for this study. We included checklist items that an individual patient may reasonably be expected to independently implement as part of a self-management intervention for ALBP. In contrast to generic information quality YouTube evaluation tools such as DISCERN, this checklist incorporates specific ALBP self-management guideline-consistent items.

First, we developed a codebook for analyzing the data set using the ALBP checklist. The ALBP checklist was based on self-management items described in the Lancet 2018 ALBP guidelines [12] (Multimedia Appendix 6). This includes maintaining physical activity, education, identification of red flags, analgesia, and reassurance. Second, we coded each video in the data set using the ALBP self-management checklist. Third, we analyzed the ALBP checklist results by authors’ disciplines. In summary, by examining the characteristics, information quality, and self-management content, we aimed to determine whether it was plausible that a checklist for YouTube video assessment may facilitate self-management of ALBP.

We identified clear differences based on the author’s discipline in the ALBP videos in our data set. We found that the videos authored by each discipline strongly featured a specific intervention domain such as education, treatment, or exercise. Using a checklist, we found that the videos authored by physicians were consistently coded with the highest ALBP self-management content scores relative to all other disciplines. We suggest that a checklist may facilitate the discovery of guideline-consistent ALBP YouTube content.

We compared the YTDT results with the Australian and US website results to determine the validity of the YTDT raw data set. We found that the US raw data set obtained via YTDT matched with the Australian and US YouTube website results obtained via anonymous sign in and VPN through Chromebook from New York and Sydney. After the removal of duplicates, we identified 202 unique videos. These 202 videos became our final data set (Multimedia Appendix 8). The final data set represented popular videos likely to be displayed to YouTube searchers for information on back pain in Australia and the United States in March 2020.

We had several notable findings from our analysis of the characteristics of the final data set. Videos published by mainstream health providers (physicians and physiotherapists) were more common in results filtered by search relevance than in most viewed categories (Table 1 and Multimedia Appendix 4). However, overall, ALBP videos published by other providers (chiropractors, fitness, yoga, and other categories) were viewed more often than mainstream health provider videos (Textbox 1 and Multimedia Appendix 7). Overall, chiropractic videos were the most viewed discipline in our final data set. We found that each discipline predominantly produced videos in a specific domain. For example, medical authors primarily published education videos, whereas chiropractors published primarily real-time treatment videos (Table 2 and Multimedia Appendix 8). The most viewed video in our data set featured real-time chiropractic treatment (Multimedia Appendix 9). This video scored poorly on the modified Brief DISCERN and ALBP checklists. User comments suggested that this video was commonly viewed for the purposes of sexual gratification. In summary, we identified clear differences in the ALBP videos in our data set based on the author’s discipline. The disciplinary background of the ALBP video author appears to be a noteworthy consideration in selecting guideline-consistent YouTube videos appropriate for facilitating the self-management of ALBP.

We used a modified Brief DISCERN checklist to examine the information quality of ALBP videos in the data set. We examined the information quality for each discipline separately (ie chiropractic, fitness, medicine, physiotherapy, yoga, and other categories). The number of videos in the final data set varied by discipline. We therefore displayed information quality results by each modified Brief DISCERN item as a percentage of the number of yes responses to that item (Table 3 and Multimedia Appendix 10). For example, in the medicine discipline, 75% (n=15) of videos were coded yes in response to Question 3 of the modified Brief DISCERN. Question 3 refers to videos featuring a biologically plausible mainstream explanation of the mechanism of treatment. In the case of the medicine category, 75% (n=15) of the videos were coded yes for providing a biologically plausible mainstream explanation of the mechanism of treatment action.

Overall, we found that videos categorized as medicine were consistently coded with higher scores than all other disciplines. These higher scores indicated that medically authored videos had the highest information quality. In contrast, videos from fitness and other disciplines were consistently coded with the lowest information quality scores.

We used the ALBP self-management checklist to examine the content for each discipline in the final data set. Overall, we found that medically authored videos were coded with consistently higher scores for self-management content than all other disciplines (Table 4 and Multimedia Appendix 11). Chiropractic and fitness videos were consistently coded with the lowest scores for ALBP self-management. Overall, we found that the ALBP self-management checklist may be more sensitive than the modified Brief DISCERN instrument in assisting researchers in identifying differences in self-management content among disciplines and among individual videos.

In summary, we identified clear differences in the ALBP videos in our data set based on the author’s discipline. The author’s discipline appeared to be a determinant of the number of views, information quality, and ALBP self-management content of the videos in the data set.

We identified considerable variability in the guideline concordance of ALBP self-management content on YouTube. We found that the video author’s discipline is an indicator of the provision of guideline-consistent information. We suggest that the access to guideline-consistent ALBP content may be improved by referring to the author’s discipline. Furthermore, we suggest that a checklist used with YouTube videos may facilitate the discovery of guideline-consistent ALBP self-management content. We have described the implications of our findings under the following categories: access and discovery, discipline-specific discovery, and self-management.

We identified several limitations in this study.

Further research is needed to establish the feasibility of checklists and YouTube content discovery for self-management of ALBP in specific clinical contexts. We propose 2 directions for future research to extend this exploratory research into clinical practice. First, a future research program could refine the questions of feasibility, including cost-effectiveness and clinical utility. We suggest that the next stage of research on the use of YouTube for self-management of ALBP could focus on establishing feasibility in a specific clinical context, such as low acuity low back pain interventions by paramedics. Second, during 2020, COVID-19 accelerated the demand and supply for telehealth across the globe [97]. This has also accelerated the experimentation with novel clinical approaches. In light of the rapid uptake of telehealth during 2020, further research into the feasibility of incorporating YouTube with telehealth self-management may be warranted.

Individuals are increasingly using YouTube to self-diagnose and self-manage health conditions, including ALBP. However, the results returned by YouTube in response to searches for back pain content were highly variable. This exploratory study aims to increase the understanding of the ALBP content available on YouTube in 2020 and to establish the plausibility of using a simple checklist to facilitate the discovery of YouTube content that is consistent with current management guidelines. We suggest that a simple checklist may facilitate the discovery of guideline-consistent ALBP self-management content on YouTube. Further research may identify the clinical contexts in which the use of an ALBP checklist with YouTube is feasible.