Aims YouTube is a vast source of freely accessible user-uploaded medical information. To our knowledge no study has analysed the quality of parent-uploaded videos which depict illness in their children. We aimed to investigate the quality and quantity of videos representing two common conditions, croup and dehydration.
Method YouTube was searched using the search terms ‘croup+child’ and ‘dehydration+child’. The first 400 videos of each search were screened. Videos created by doctors or by educational institutions were excluded. The parent-uploaded videos were analysed using the validated Medical Video Rating Scale. Each video was separately evaluated for whether it represented a good clinical example of the condition featured.
Results Out of 38 ‘croup’ videos which met criteria, 15 were judged to be good clinical examples. Only 7 of these 15 ‘good clinical example’ videos were also of high technical quality. Out of 28 ‘dehydration’ videos which met the inclusion criteria, two were good clinical examples. One of these videos had good technical quality.
Conclusions There were very few videos of either condition which showed a good clinical example while also displaying high technical quality. It is extremely difficult and time consuming to isolate such examples from the mass of information available and therefore parents could be misled by apparently high technical quality videos which are not in fact good clinical examples. Healthcare professionals should be careful when discussing finding medical information on YouTube and consider creating repositories of good examples so they are able to direct parents towards more reputable resources.
- Medical Education
- General Paediatrics
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known on this topic
YouTube is the second most popular website globally, with over a billion users and 4 billion video views per day.
It is a constantly growing source of unmoderated user-uploaded medical information.
Information is lacking on the quality of parent-uploaded videos on YouTube demonstrating certain symptoms in children.
What this study adds
It is extremely difficult and time consuming to isolate high quality video examples of unwell children that would be suitable educational material.
Parents and laypeople could be misled by apparently high technical quality videos (which have good light, sound and picture quality) which are not in fact good clinical examples.
Healthcare professionals may advise parents that videos on YouTube featuring unwell children can be misleading, and should be able to direct them towards more reputable online resources.
YouTube is the second most popular website globally, with over a billion users and 4 billion video views per day.1 It is a constantly growing source of user-uploaded video content with freely available unmoderated and unregulated medical information that anyone can access.
Healthcare professionals have begun to use YouTube as a tool for patient education in certain areas,2 but the volume of professionally created content is far outweighed by content that is created by YouTube users who are not healthcare professionals. As videos are often not peer reviewed, there is a wide variation in the quality and reliability of videos available. Studies examining the type and quality of information available on YouTube for a variety of adult health conditions have generally found that a significant proportion of content is misleading.3–6 For some conditions, misleading videos have a higher number of hits than accurate videos.7
YouTube is increasingly being used to disseminate health information8 in a format that is familiar to those who have grown up with social media. This generation have begun to have children of their own, and as new parents they are increasingly likely to turn to social media when they need medical information.
As YouTube is such a popular video sharing website, parents may be using the site in order to find out more about a known (physician diagnosed) condition or to attempt to ‘self-diagnose’ by comparing their child's symptoms to videos of similar symptoms. Inaccurate YouTube videos which misrepresent a medical condition may mislead parents and feed their misconceptions around clinical signs and/or management of the condition.8 For healthcare professionals, knowledge of potential sources of misinformation could help to guide parents towards more appropriate sources of medical information.
It is also important to bear in mind that even healthcare professionals may find it difficult to judge the acuity of a case on video depending on its technical quality,9 and so non-medically trained parents would therefore find this even more challenging.
There has been limited study into YouTube videos of paediatric epilepsy,10–12 with very few other paediatric health conditions having been examined. No study to our knowledge has yet specifically analysed the quality of parent-uploaded videos demonstrating particular symptoms in children.
To describe the quantity, quality and popularity of parent-uploaded YouTube videos featuring unwell children with croup and dehydration, and to analyse their reliability as a potential source of medical information for other parents.
Two common paediatric conditions (croup and dehydration) were chosen for comparison, as the former has clearly visible and audible clinical signs—potentially lending itself well to video portrayal—while the latter may not be as easily represented on video.
We performed identical internet searches via servers based in the UK and the Netherlands separately for each condition on 6 October 2015. Search terms were ‘croup+child OR croup+baby’ for croup, and ‘dehydration+child OR dehydration+baby’ for dehydration. Research ethics approval was not required as all information is in the public domain.
The first 400 videos identified in each search were explored (20 webpages of search results). Videos created by doctors or by educational institutions were excluded. Videos which clearly had been uploaded by parents or caregivers were included for evaluation. The current number of views, comments, likes and dislikes for each video was documented.
The videos which met the inclusion criteria were analysed independently by four individual paediatricians (a junior clinician and senior clinician based in the UK, and a junior clinician and senior clinician based in the Netherlands). Each individual scored each video for its technical quality using the first part of the validated Medical Video Rating System (MVRS).11 A score of either 1 or 0 was given for each attribute: light, sound, angle, duration and resolution. This gave a total possible high score of 5 for technical quality.
We combined the second and third parts of the MVRS (‘diagnostic accuracy’ and ‘effectiveness as a clinical example’) so that each clinician evaluated these attributes and gave a ‘yes’ or ‘no’ opinion on whether they would recommend the video overall as a useful resource for a non-medically trained parent whose child had the condition.
The MVRS is a subjective rating scale. If differing overall scores were given by the four assessors for a particular video, a separate senior clinician reviewed the video and cast the deciding vote on the score.
For each condition, the search identified a similar number of videos (over 12 000). The videos automatically displayed at the top of the search results were those with the highest view counts. There was no association between number of views and the quality of a video as a clinical example.
The search identified over 12 400 videos (table 1). From the first 400 videos in the search results, 40 croup videos met the inclusion criteria. Two of these were later removed by their creators before all members of the study team were able to view them. This left 38 videos which underwent full analysis.
The number of views per video ranged widely from 142 to over 1.2 million. Ten of the videos had over 50 000 views. The median number of views was 3501.
Out of 38 videos, 15 were judged to be a good clinical example. Only seven of these 15 were also highly rated for technical quality (MVRS 4 or 5). Three of the seven videos which were of high clinical quality, while also being good technical examples, had received over 50 000 views.
Of the 38 videos, 15 were of high technical quality. Of these, nine were judged to be poor clinical examples. Five videos which were of high technical quality but low clinical quality had over 50 000 views.
Nine videos were of low technical quality (MVRS score 1 or 2), although three of these were judged to be good clinical examples. The most viewed of these ‘poor technical quality, good clinical example’ videos was seen 23 000 times and the least viewed was seen 298 times.
The search identified over 13 000 videos (table 2). From the first 400 videos in the search results, 28 dehydration videos met the inclusion criteria.
The number of views per video ranged widely from 100 to over 1.8 million. Four videos had been viewed over 50 000 times. The median number of views was 1542.
Out of the 28 videos, two were judged to be good clinical examples. One video identified had the characteristics of being a good clinical example while also being of high technical quality. This video had received over 1.8 million views.
Five videos out of the 28 were of high technical quality (MVRS score 4 or 5). Four of these high technical quality videos were very poor clinical examples (scoring 1 or 2), although one of them had been viewed over 159 000 times.
There were three videos of low technical quality (MVRS score 1 or 2). None of these was a good clinical example although one had been viewed over 15 000 times.
In a comparison of YouTube videos of croup and dehydration uploaded by parents, we identified a similar number of search results for each condition. However, the videos available were mostly of poor technical or clinical quality. In particular, technical quality and clinical relevance were not related. For example, only four of the 15 ‘good clinical example’ croup videos were rated as having high technical quality. Conversely, three out of the nine ‘low technical quality’ croup videos were found to be good clinical examples.
The non-medical viewer of a poor technical quality video (ie, a video with poor lighting, bad angle or inadequate sound) may be less likely to persist with viewing it, and falsely ascribe a negative judgement regarding its utility as an informative piece of media.11 Equally, a video with high production values may appear more ‘trustworthy’ than it is—its high technical quality masking an insufficiency of useful medical information. There are inherent risks in parents searching for unfiltered and unregulated video cases. While we only reviewed two clinical examples, it is likely our findings would be similar for other medical conditions.
There was a stark contrast between the number of useful clinical examples for croup and the number of useful clinical examples for dehydration. Analysis of the videos that matched the inclusion criteria revealed that there were only two useful clinical examples of dehydration compared to 15 useful clinical examples of croup.
This reflects the fact that croup is a condition with clear visual and audible signs that can be demonstrated on video. Many of the results for ‘dehydration’ identified by our search were ‘amusing’ videos of otherwise well children unexpectedly vomiting—irrelevant to somebody searching for medical information regarding the clinical condition of dehydration.
It was beyond the scope of this study to examine the reasons behind parents' choice to upload videos of their unwell children. Only a minority of uploaders reported on reasons for uploading (given in the text description accompanying some videos). Whatever the reason behind uploading, however, the number of views of the videos is striking—the most popular videos have been viewed well over a million times.
As well as the motive for uploading, it would therefore also be valuable to understand more about who is searching for and viewing these videos, and their reasons why. It would be useful to study laypersons' reactions to these same videos to determine whether they are able to distinguish between high quality technical examples and low quality clincial examples.8
The strengths of this study include the use of an objective video rating scale to analyse technical aspects of the videos, and the fact that each video was independently analysed by four clinicians of varying clinical experience but all specialising in paediatrics. The study team was also equally split between the UK and the Netherlands and the search performed in each country identified the same videos, demonstrating that the same online content is available in the UK as in mainland Europe, supporting generalisation of our results to a wider European population.
It is not possible on YouTube to determine what proportion of video views come from any one area of the world. Further areas of study might include contacting the content creator/uploader to ask about their motivation for uploading, and analysis of comments below popular videos to gain understanding of viewers' reactions and motivations for viewing.
The healthcare professional or educator searching YouTube for resources will be able to find excellent examples of clinical signs if they are willing to spend time searching, although many of the videos with useful clinical signs are in fact of low technical quality. Videos of high technical quality which also display useful clinical signs are rare, and we found more of these videos for croup—a condition which has clear audible and visual signs compared with dehydration.
The sheer amount of video material on YouTube means that many low quality parent-uploaded videos are present within search results, and it is extremely time consuming and difficult to isolate any useful videos. The difficulties non-medically trained parents or carers will have in identifying such resources would be significantly greater.8 ,13
As healthcare professionals, we should be conscious of social media trends and maintain a broad awareness of the range of medical information which parents and carers may be accessing, and realise that their perceptions may be informed by inaccurate sources.
The fact that it is very difficult to find useful clinical examples on YouTube means that healthcare professionals should warn parents that material found on YouTube relating to their child's condition may be misleading. It would be beneficial if healthcare professionals were able to recommend alternative reputable internet resources if parents wish to access information in this way. These recommended reputable resources will vary from country to country and between institutions. There may be a role for paediatric professional societies to address this issue and offer a collection of suggested viewing material for parents.
It is currently unclear in most cases why parents choose to upload videos of their unwell children. If this were better understood, it could lead to collaborative projects between parents and healthcare professionals to create relevant, accurate and informative content in a format that may be preferred by healthcare students and parents alike.
There were very few videos of either condition which had high technical quality and were also good clinical examples. It is extremely difficult and time consuming to isolate such examples from the mass of information available.
Parents and laypeople could be misled by apparently high technical quality videos (which have good light, sound and picture quality), which are not in fact good clinical examples. Healthcare professionals should be aware of the content available on YouTube so that they can caution parents that videos featuring unwell children may be misleading. Healthcare professionals should also be able to guide parents towards more reputable sources if they wish to search for child health-related videos online.
Twitter Follow Damian Roland @damian_roland
Contributors DR had the initial idea for the research. KK and DMvL gathered and analysed data. KK prepared the abstract and manuscript. DR, HAM and RO were all equally involved as supervisors providing advice and guidance at all stages.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.