YouTube as a source of patient information for Diabetes: Evaluation of Arabic content

Introduction

People diagnosed with diabetes mellitus (DM) has risen dramatically over the last two decades. More than 382 million cases throughout the world have been reported, with 90% of those diagnosed with type-2 diabetes mellitus. [1] In Arab countries, of the 35 million people diagnosed, North Africa and the Middle East leads the world in the highest rate of type-2 diabetes mellitus where one in 10 people are living with the disease. By 2035, the number of people with diabetes is expected to increase by 96.2%.[1]

The advent of social media becoming increasingly popular provides a quick and convenient source for patients looking for health-related news and information. Recent studies have shown that the Internet serves as a primary outlet for patients seeking medical-related knowledge and information, and the browsing habits of patients show that 80% of their online activity is seeking knowledge, medical support, and information .[2,3] YouTube serves as the second most popular website visited around the globe and in Arab countries.[4] Unfortunately, after an evaluation of its content, studies have concluded that YouTube consists of faulty and misleading medical-related content, with the majority of researchers agreeing that patients could be misled or misinformed about their health in accessing the site.[5–11] The Internet poses as a major source for diabetic patients to self-educate[12] and is an ever-growing outlet for health organizations and hospitals to inform patients on healthcare issues. The quality of those sources though is not congruous to the availability of these videos. The enormity in the number of educational videos uploaded to YouTube daily oftentimes makes these videos inaccessible to users.

If diabetes-related material could be presented on YouTube in the Arabic language, it could serve to educate Arabic patients and caregivers in a more meaningful and more widely- understood format. [13] To the best of our knowledge, no studies exist that assess Arabic content provided by YouTube on the subject of diabetes. This study aimed to analyze Arabic content on YouTube for accuracy and assessed the responses of viewers as it pertains to diabetes videos.

Methods

a YouTube search was performed on May 15, 2018 using the search term, “diabetes mellitus” in Arabic with the filter by relevance function. Excluded were non-Arabic videos, duplicate videos, and those without audio functions or unrelated to our focus. Cookies and the computer’s history were cleared to have no effect on the search results. Overall, 127,000 videos were found. The first 200 (20 videos/page, 10 pages) videos were retrieved as a sample with the assumption that most users would not venture beyond this point. Furthermore, similar sample sizes were used in previous work of qualitative methods assessing YouTube videos in comparison to our analysis.[14–16]

The video demographic assessment recorded likes, dislikes, upload dates and source information, video length, and number of views. To assess the number of mean views per day, a calculation was determined to divide the total amount of views by the number of days since the videos had been uploaded online. Classifications were assigned to identify the upload source. These included health care professionals (HCP), personal users, academic/hospital/professional institutions (PA), and other by using information from their YouTube profile in the “about” listing. Independent two researchers watched each individual video and made determinations regarding their scientific merit and accuracy (Alnughaymishi and Algeffari). A third independent researcher (Diabetologist Consultant) was used to resolve any disputes that arose between the two researchers on the content of the videos and served as the final decision maker (Almogbel). Three groups of classifications resulted from viewing the contents of the videos as 1) reliable, meaning scientifically accurate for DM; 2) misleading, which means the information contained scientifically inaccurate or unproven information; and 3) patient experiences, where the content pertained to the personal experiences of patients and did not provide DM medical knowledge. In the event a video content included two categories, reliable and misleading information, “misleading” was used as the default. The videos were further broken down into seven different categories of diet, complications, interview, education, management, symptoms, and general information.

Descriptive statistics were given for all variables. Interrater agreement was calculated using the k test. The kappa value was classified as poor agreement when <0.20, fair when 0.21–0.40, moderate when 0.41–0.60, good when 0.61– 0.80, and very good when >0.81.[17] Differences between groups were compared using one-way analysis of variance (ANOVA). A p value of less than 0.05 was considered as significant. All statistical analyses were done of SPSS version 20 (IBM Corporation).

Result

The initial search using the term “Diabetes Mellitus” in Arabic resulted in finding 127,000 videos. Out of first 200 (20 videos/page, 10 pages) videos, 129 was the final number included in the analysis after 71 videos excluded (Fig. 1).

Table 1 provides a list of characteristics for the 129 videos deemed eligible for the analysis. In terms of time, the videos ranged from 40 seconds to two hours long. The median runtime was determined to be 4.1 minutes. The median total number of views was 29,464. The range for number of views ran between 40 to 3,234,372.

Table 1: Descriptive results of entire sample (n=129).
	Mean	Median	Standard deviation	Min	Max
Days since upload	888	648	726.2	33	4004
Views	227620	29464	558295.9	40	3234372
Likes	2317	133	6582.2	0	37000
Dislikes	174	11	482.9	0	2800
Video duration (minutes)	9.4	4.1	15.7	0.4	121.4
views per day	484	29	1460.2	0	9827

In the classification for accuracy, the researchers showed a very good inter-rater agreement (K = 0.84). Six videos produced a dispute among the two researchers that resulted in resolution by virtue of discussion. Table 2 refers to the accuracy of video. Overall, more than half were considered reliable (n=78; 61%), 47 were determined misleading (36%), and 4 videos (3%) as patient experiences. Regarding total number of views, 16,358,514 viewed misleading videos, 12,482,917 viewed reliable videos, and patient experiences were viewed 521,536 times. The highest mean number of views belonged to misleading videos (348053± 611310.6), followed by reliable videos (160037± 528287.7), and patient experiences videos at 130384± 216865.6 (ANOVA, p=0.178). The mean number for “likes” of reliable videos was 1632 ± 6039.2, a lower amount than misleading videos at 3412.5 ± 7444.1, and patient experiences videos at 3070.3 ± 5953.6. Correspondingly, significant differences in the mean number of “dislikes” (ANOVA, p < 0.041) compared to lower numbers for reliable videos. Patient experiences videos regarding length were longer (17.5 ± 24.6 min) in comparison to reliable videos (7 ± 10.4 min) and misleading videos (12.0 ± 21.1min).

Table 2: Basic video descriptors by accuracy of contents.
Descriptor	Reliable (n=78)	Misleading (n=47)	Patient Experiences(n=4)	P
Total views	12482917	16358514	521536
Mean views (SD)	160037(528287.7)	348053(611310.6)	130384(216865.6)	0.178
Mean days since upload (SD)	901)706)	831.2(750.8)	1287.8(894.1)	0.469
Mean views per day (SD)	278 (1047.9)	856.3(1966.6)	131.5(221.7)	0.088
Mean video duration, min (SD)	7 (10.4)	12.0(21.1)	17.5(24.6)	0.169
Mean likes	1632(6039.2)	3412.5(7444.1)	3070.3(5953.6)	0.363
Mean Dislikes	92 (360.2)	320.3(632.6)	95.3(169.5)	0.041
Upload source
PI	8(10.3%)	1(2.1%)	0(0.0%)
HCP	6 (7.7%)	2(4.3%)	0(0.0%)
Personal	27 (34.6%)	33(70.2%)	1(25.0%)
Other	37(47.4%)	11(23.4%)	3(75.0%)
Video content
Education	19(24.4%)	1(2.1%)	0(0.0%)
Diet	16(20.5%)	2(4.3%)	0(0.0%)
Interview	2(2.6%)	1(2.1%)	2(50.0%)
Complications	10(12.8%)	1(2.1%)	0(0.0%)
Symptoms	9(11.5%)	2(4.3%)	0(0.0%)
Management	3(3.8%)	32(68.1%)	0(0.0%)
General Information	19(24.4%)	6(12.8%)	2(50.0%)

Personal users uploaded many of the videos (47.3%, n=61). 27.1%(n=35) of that videos contain information on the management of diabetes. Personal users mostly uploaded misleading videos (70.2%, n=33), which was followed by other users at 23.4%(n=11). Regarding misleading videos, the treatment of diabetes using alternative medicine comprised the largest group of users at 74.5%(n=30). Of those, 23 of the misleading videos promoted herbal medicine, 5 advocated hydro-thermal therapy, and two proposed Ruqya Al Shariah therapy. A trend in alternative medicine videos uploads increased over the previous two years, with 2017 figures at 37% (2017; n=11, 37%) and 2018 at 47% (n=14). The highest number of mean views belonged to alternative medi­cine at 402,483 times compared to other videos at 25,2000 times.

Discussion

Health education today comes from many resources. Internet videos are used frequently, with YouTube leading the pack. As of June 2018, YouTube consisted of over 5,000,000,000 videos and over 1,000,000,000 hours watched daily by users.[17] The usefulness of videos relative to any given field and the sources that are used the most have been the subjects of many studies. In this study, we assessed the accuracy and scientific proven material and information available on YouTube regarding diabetes. In spite of different entities increasingly using the Internet to relay medical information, Arabic information related to diabetes on YouTube has not been inspected for accuracy. From our results, we conclude that less than two-thirds of diabetes videos on YouTube provide reliable information. Reliable videos resulted in 60.5%, which is about what previous studies determined in evaluating YouTube videos on medical information in other fields. The range of reliable videos in these studies were between 54.9% to 63.6% including topics such as gallstone disease [18], kidney stones [19], on hypertension [20] ,and dialysis [21].

The videos determined to be misleading – many of which associated with alternative medicine, endorsed known myths and inaccurate information about the management of diabetes. Interestingly, those videos earned more YouTube views than reliable videos. The data pertaining to these investigations and studies point to a contradiction in results; nevertheless, most researchers agree that websites such as YouTube play a critical role in educating and informing the public.[19,22,23] This inclination also applies to YouTube videos that pertain to medical-related issues.[18,20,21] The popularity of videos being used as a source of information cannot be denied, regardless of factors leading to contributions in the number of views or the content. The study presented here highlights that many of the misleading videos include alternative medicine information. Regardless of the lack of evidence or effectiveness alternative medicine provides, these videos were viewed widely all over the world. The reported prevalence in the use of alternative therapies for diabetes was pointed out in Kamel et al. as 64% in Saudi Arabia.[24] Given the popularity, YouTube users in Saudi Arabia are more likely to view alternative medicine videos classified as misleading.

In another study referencing cervical cancer, Adhikari et al.[25] found that most of the videos were being uploaded by physicians. Additionally, Gunes et al. [7] found that videos on varicose veins were mostly being uploaded by professional health providers. In this study, a very small amount of content was attributed to hospital or academic professionals. The impact of what healthcare professionals contribute on YouTube is quite small subsequently, and possibly intentional given the amount of erroneous information on this platform. These results show consistency with other research, and suggest that the impact is minor when healthcare professionals and organizations become involved in the manufacturing of YouTube content on health-related matters.[26] Personal uploads and news sources were uploaded more than professional or academic sources. As such, personal stories, alternative medicine information, or entities selling or advertising services on the management of diabetes were shown to be more popular in uploads from independent users. This warrants discussion.

The limitations of our research include first, the ever-changing environment of Internet media in relation to the date and time of performing searches. Second, this study used keyword searches on YouTube (diabetes mellitus in the Arabic language) using the first 20 pages (20 videos/page) as its sample. The videos were then sorted by relevance, a default setting on YouTube’s platform that might have affected the results due to advertisements. Although this could factor in, the results showed consistency compared to other research on health-related content and information. It also points to an unlikely event that missing variables have any true effect on results.

Conclusion :To conclude, this study supports that YouTube videos in the Arabic are not very helpful. Overall, reliable content videos received less views than those determined to be misleading. Professional institutions, health care providers, and academic sources need more encouragement to develop and upload videos on the management and evidence-based treatment of diabetes. The popularity of the YouTube platform warrants further efforts to increase the accuracy and promote science-based information on diabetes.

Figure 1: Flow diagram showing search strategy:

127000 videos were retrieved using the search term “Diabetes Mellitus “ in Arabic

71 videos excluded:

8 Non-Arabic

14 Duplicate videos

42 Without audio

7 Not medical

126800 videos excluded when limited to first ten pages

129 videos included for analysis

200 videos for analysis

References

[1]. International Diabetes Federation. IDF Diabetes Atlas, 6th edition. Brussels, Belgium: International Diabetes Federation, 2013. Available from: URL: http://www.idf.org/diabetesatlas

[2]. John Powell, Nadia Inglis, Jennifer Ronnie, Shirley Large, The characteristics and motivations of online health information seekers: cross-sectional survey and qualitative interview study, PubMed entry [Internet]. [cited 2018 Nov 25]. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21345783

[3]. Kapil Chalil Madathil , A Joy Rivera-Rodriguez , Joel S Greenstein , Anand K Gramopadhye , Healthcare information on YouTube: A systematic review ,PubMed entry [Internet]. [cited 2018 Nov 25]. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24670899

[4]. Alexa - Top Sites for Countries [Internet]. [cited 2018 Nov 25]. Available from: https://www.alexa.com/topsites/countries

[5]. Janak Adhikari, Priyadarshani Sharma 1, Lubina Arjyal , Dipesh Uprety ,YouTube as a Source of Information on Cervical Cancer. - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/27213142

[6]. Neetika Garg , Anand Venkatraman, Ambarish Pandey, Nilay Kumar ,YouTube as a source of information on dialysis: a content analysis. - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/25641264

[7]. Tevfik Güneş, Mustafa Serinken, İhsan Alur, Halil Beydilli, et al. ,YouTube as a source of information on varicose veins , SAGE Journals, 2016 [Internet]. [cited 2018 Oct 30]. Available from: http://journals.sagepub.com/doi/abs/10.1177/0268355515596894?journalCode=phla

[8]. Jun Suh Lee , Ho Seok Seo , Tae Ho Hong, YouTube as a source of patient information on gallstone disease, PubMed entry [Internet]. [cited 2018 Oct 30]. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24744597

[9]. Mustafa Serinken,Cenker Eken, Fikret Erdemir, Hayri Eliçabuk, etal., The reliability of national videos related to the kidney stones on YouTube , PubMed Central Link [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4791085/

[10]. Serpil Yaylaci , Mustafa Serinken, Cenker Eken, Ozgur Karcioglu, et al ., Are YouTube videos accurate and reliable on basic life support and cardiopulmonary resuscitation? , PubMed entry [Internet]. [cited 2018 Oct 30]. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25168312

[11]. Nilay Kumar , Ambarish Pandey , Anand Venkatraman , Neetika Garg, Are video sharing web sites a useful source of information on hypertension? - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/25064770

[12]. Pereira K, Phillips B, Johnson C, Vorderstrasse A. Internet delivered diabetes self-management education: a review. Diabetes Technol Ther. PubMed - NCBI [Internet] [2015 Jan;17](1):55–63. Available from: https://pubmed.ncbi.nlm.nih.gov/25238257/

[13]. Alverdy JC. A Video Is Worth a Thousand Words. JAMA Surg. , 2016 Jun 1;151(6):e160476–e160476. Available from: https://jamanetwork.com/journals/jamasurgery/article-abstract/2512055

[14]. Chou W-YS, Hunt Y, Folkers A, Augustson E. Cancer Survivorship in the Age of YouTube and Social Media: A Narrative Analysis. J Med Internet Res [Internet]. 2011 Jan 17 [cited 2018 Oct 31];13(1). Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3221357/

[15]. Gao X, Hamzah SH, Yiu CKY, McGrath C, King NM. Dental fear and anxiety in children and adolescents: qualitative study using YouTube. J Med Internet Res. [2013 Feb 22];15(2):e29. Available from: https://pubmed.ncbi.nlm.nih.gov/23435094/

[16]. Morahan-Martin JM. How internet users find, evaluate, and use online health information: a cross-cultural review. Cyberpsychol Behav. [2004 Oct];7(5):497–510. Available from: https://pubmed.ncbi.nlm.nih.gov/15667044/

[17]. YouTube by the Numbers (2018): Stats, Demographics & Fun Facts, omnicore [Internet]. 2018 [cited 2018 Oct 30]. Available from: https://www.omnicoreagency.com/youtube-statistics/

[18]. Lee JS, Seo HS, Hong TH. YouTube as a source of patient information on gallstone disease. World J Gastroenterol. 2014 Apr 14;20(14):4066–70. Available from: https://pubmed.ncbi.nlm.nih.gov/24744597/

[19]. Serinken M, Eken C, Erdemir F, Eliçabuk H, Başer A. The reliability of national videos related to the kidney stones on YouTube. Turk J Urol. 2016 Mar;42(1):7–11. Available from : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4791085/

[20]. Nilay Kumar , Ambarish Pandey , Anand Venkatraman , Neetika Garg, Are video sharing web sites a useful source of information on hypertension? - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/25064770

[21]. Neetika Garg , Anand Venkatraman, Ambarish Pandey, Nilay Kumar, YouTube as a source of information on dialysis: a content analysis. - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/25641264

[22]. Yaylaci S, Serinken M, Eken C, Karcioglu O, Yilmaz A, Elicabuk H, et al. Are YouTube videos accurate and reliable on basic life support and cardiopulmonary resuscitation?, Emerg Med Australas. 2014 Oct;26(5):474–7. Available from: https://pubmed.ncbi.nlm.nih.gov/25168312/

[23]. Liu KY, Haukoos JS, Sasson C. Availability and quality of cardiopulmonary resuscitation information for Spanish-speaking population on the Internet. Resuscitation. 2014 Jan;85(1):131–7. Available from: https://pubmed.ncbi.nlm.nih.gov/24036407/

[24]. Kamel FO, Magadmi RM, Hagras MM, Magadmi B, AlAhmad RA. Knowledge, attitude, and beliefs toward traditional herbal medicine use among diabetics in Jeddah Saudi Arabia. Complement Ther Clin Pract. 2017 Nov;29:207–12. Available from: https://pubmed.ncbi.nlm.nih.gov/29122263/

[25]. Janak Adhikari , Priyadarshani Sharma , Lubina Arjyal , Dipesh Uprety, YouTube as a Source of Information on Cervical Cancer. - PubMed - NCBI [Internet]. [cited 2018 Oct 30]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/27213142

[26]. Thackeray R, Neiger BL, Smith AK, Van Wagenen SB. Adoption and use of social media among public health departments. BMC Public Health. 2012 Mar 26;12(1):242. Available from: https://bmcpublichealth.biomedcentral.com/articles/10.1186/1471-2458-12-242