The role of WhatsApp® in medical education; a scoping review and instructional design model

Background Technological advances have driven huge change in educational practices though concerns exist about a lack of evidence informing this change, in particular with social media-based medical education activities. The purpose of this study was to conduct a scoping review of WhatsApp use in medical education, narratively describing how it has been used and evaluated, and the theoretical considerations in relevant articles. Methods A modified 5-stage scoping review model was used. We performed 2 searches from February 2009 to February 2019 in EBSCO, SCOPUS, Web of Science, EMBASE, Medline PubMed and Google Scholar) using the term “WhatsApp” in all search fields. A 3-stage process for study selection was performed. Only original articles in English presenting original data about WhatsApp in medical education were included. The Kirkpatrick model of training evaluation was used to describe learning outcomes in included studies. Results Twenty-three articles were selected for review. Three strategies for WhatsApp use were apparent; primarily educational use with a pre-defined curriculum (n = 5), primarily educational use without a curriculum (n = 11), and primarily non-educational use (n = 7). Most of the educational studies used an online moderator and were in a local hospital or university department. Studies not primarily educational were national or international and seldom included an online moderator. All 5 studies with a pre-defined curriculum reported Kirkpatrick level 2 learner knowledge outcomes. A majority of the remaining studies only reported Kirkpatrick level 1 learner attitudes. Seven studies with 647 participants reported an improvement in learners’ knowledge following WhatsApp learning, though methodological weaknesses were apparent. Evidence for underlying learning theory considerations were scant throughout the studies. Conclusions WhatsApp is popular and convenient in medical education. Current published literature suggests it may also be effective as a medical learning tool. By combining the 3 strategies for WhatsApp use and the exploration-enactment-assessment integrated learning design framework, we propose an instant messenger design model for medical education. This may address the need for theory-driven instructional design in social media learning. Further research would clarify the role of WhatsApp and our design model in this area.


Background
Advances in information technology have driven huge changes in many aspects of human behaviour and communication. These changes have had considerable implications for educational practices. In particular, the last decade has seen widespread access to mobile internet devices (MIDs) which in turn have expanded educational opportunities outside the classroom setting [1]. Learners with a suitable MID and a link to the world wide web have ready access to a wide range of multimedia learning resources, collectively known as mobile learning (mLearning) [2].
MIDs enable access to two main resources for the medical learner; applications and social media (SM) networks. The former, an extensive list, include UpToDate®, Medscape®, peer-reviewed journals and numerous podcasts [3]. The latter includes wikis, online blogs, YouTube®, and instant messenger applications (IMAs) such as Facebook®, WhatsApp®, Twitter® and WeChat® [4]. IMAs, while not primarily educational in nature, share common features which can facilitate learning; group collaboration, peer communication independent of time and geographical location, and multimedia message sharing [5][6][7].
WhatsApp®, a free standalone IMA launched in 2009, has over 1 billion active users in 180 countries [8]. In December 2017, it was the most popular IMA in South America, India, Russia, Eastern Europe, the UK and Africa, and the second most popular in North America [9]. As a secure educational tool it uses two-way opt-in for all users, allows the monitoring of users' activity and message reading, and has end-to-end encryption [10]. It has some theoretical benefits over other IMAs; prior registration with a SM network is not required, and it is more favourable if internet bandwidth or speeds are poor [10].
The use of SM and IMAs as learning tools has met with resistance from some medical faculty members. While this in part relates to technical unfamiliarity, real concerns exist about professional implications of SM use [11] and the quality of evidence supporting their learning benefits [12]. One recent review of SM in medical education highlighted how the 13 included studies tended "to focus on evaluating the effective outcomes … as opposed to understanding any linkages between social media and performance outcomes"(p369) [13]. A more recent larger postgraduate education review drew similar conclusions [14]. A large majority of studies in these reviews evaluated Facebook® but contained little information about other media or IMAs.
A key concern therefore is that the advance of SM and IMA learning in medical education may be driven more by social behaviour and the high availability and low cost of technology rather than by empirical educational research or by theory-driven instructional design. What is the evidence that recent technology advances, and the learning that they have promoted, have brought about improvements in educational outcomes? Furthermore, if such evidence exists, does it have a sound basis in the principles of educational theory?
Accordingly, the objective of this study was to explore published literature, using a scoping review framework, to evaluate the role of WhatsApp®, a ubiquitous instant messaging application, as a medical learning tool, and to articulate the extent to which this literature has a foundation in educational theory.

Methods
We used a modified 5-stage model for scoping reviews proposed by Arksey and O'Malley [15,16]. These stages are (i) identifying research questions, (ii) identifying relevant articles, (iii) study selection, (iv) charting the data and (v) collating, summarising, and reporting the results. The purposes of the review were to define the nature of existing research into WhatsApp® for medical learning and to identify a focus for future research. In keeping with scoping review guidelines, we provided a description of each study but did not apply a quality assessment tool to each [16].

Identifying the research questions
The selected research questions were: (1) How has WhatsApp® been used as a learning tool in medical education? (2) How has WhatsApp® been evaluated as a learning tool in medical education? (3) What educational theoretical principles were evident in studies of What-sApp® as a learning tool in medical education?

Identifying relevant studies
The first literature search was performed across six databases (EBSCO, SCOPUS, Web of Science, EMBASE, Medline, and Google Scholar) from February 2009, when WhatsApp® was created, until July 2018. During manuscript rewriting, in February 2019, a second search across the same databases was performed. We used the search term "WhatsApp" applied to the text, title and abstract of all publications. Reference lists from included studies were also searched. Search results were collected, organized and shared between authors using Mendeley Reference Manager®.
Relevant studies were identified using a three stage process, which involved title and abstract screening, review of abstracts, and full-text review. The first 2 stages were done independently by each author and the final stage was done collectively by both authors. Article relevance was judged by the following criteria; (i) original articles, (ii) published in English, (iii) presenting unique data (original data presented in the study) (iv) describing the use of WhatsApp® as an educational tool in a medical setting.

Selecting studies for inclusion
A total of 2974 articles were identified on the first search from which 23 article were selected for review. Details of study inclusions and exclusions are shown in Fig. 1.

Charting the data
Appropriate study data were condensed in tabulated form for each study. Each author performed this step independently for all articles and a final table was compiled following collaborative discussion between the authors (Table 1).

Collating, summarising and reporting the results
After data tabulation, we adopted a narrative approach to summarising and reporting the data, informed by our 3 research questions. We used consensus statements to guide the description of study design [39]. The Kirkpatrick Model of Training Evaluation was used as a framework for describing the learning outcomes in each study [40].
Sixteen (69.6%) of the twenty-three studies had a prospective design. Three used random allocation of participants to WhatsApp® or control groups [23,27,37]. Five studies used participants as their own controls, adopting a pre−/post-intervention design [21,30,32,36,38]. The fifteen remaining studies had a single arm design, two of which collected mainly qualitative data [17,31].
How has WhatsApp® been used as a learning tool in medical education?
Sixteen studies (69.6%) used WhatsApp® groups solely for educational purposes with a learning period from 2 days to 2 years (median duration 20 weeks).  All but one of these groups were moderated by a facilitator and most (13/16; 81.3%) were conducted in a local university or hospital setting. Seven used WhatsApp® in a blending learning setting, combining it with non-eLearning strategies [17,27,31,[34][35][36][37]. Only five of  The seven remaining studies described WhatsApp® groups that included non-educational discourse [10,20,22,25,28,29,32]. This included sharing the clinical aspects of patient care, organisational and scheduling information, emotional support and social messages. Only one of these studies had a designated moderator [29] and a majority (4/7; 57.1%) occurred at a national or international level.
How has WhatsApp® been evaluated as a learning tool in medical education?
We grouped the methods of evaluating WhatsApp® into three categories; technical/logistical aspects of the medium; learner/learning activity during discussions; and educational outcomes of WhatsApp® interventions.
Technical disadvantages cited were the necessity for internet access and compatible hardware devices, and poor image quality [17,26,34]. Technical advantages over other social media platforms (e.g. Facebook®) included easier image upload, quicker access and message posting, and the low cost and ease of use [23,26,31]. Several studies noted the high investment required by faculty to maintain the group discussions [24,[34][35][36]38] and to prevent learner disengagement over time [30,36].
Seven studies assessing a change in knowledge reported a benefit from WhatsApp® discussions but each study had flaws limiting its conclusions. Three used a pre−/post-intervention assessment tool and showed an improvement in learner knowledge but did not include a control group [21,36,38]. The remaining four studies had a control group, comparing blended learning using WhatsApp® with traditional teaching. Of these, three studies demonstrated improved knowledge in the What-sApp® groups but omitted baseline pre-intervention Table 2 Kirkpatrick level 1 learning outcomes from studies included in the scoping review Positive Convenient and efficient method of learning and solving difficult clinical problems [10,19,23,29,32] Enables learning by numerous means; -By revision [17,31] -By Q&A problem solving strategy [17] -By preplanned curriculum [23,30] or by adapting to an evolving curriculum [31] -By using multimedia tools to explain complex concepts [31] -By teacher-learner and learner-learner model [23] -By learning in a legitimate, collaborative, social, online group space [23,25,31] -By deconstructing hierarchy, reducing inhibitions and encouraging active involvement by all grades of learner [21,26,31] -By obtaining links to relevant learning material [10,23,26] Enables assessment; -Formative assessment within discussions [21,24,30] -Summative assessment tool, especially as a method for measuring learner engagement/participation in discussions [17,31] Negative Intrusiveness and interference with routine clinical work [24,31] Large volume of learning material can impede learning [23,25] Concerns about breaching patient confidentiality [24,26] Effective learning depends on "completion" of a discussion topic which does not always happen [26] testing [27,32,37]. The final study compared 2 months of WhatsApp® learning with didactic lectures, using a control group and pre−/post-intervention testing [23]. Significant improvements in learner knowledge were reported in each group but not between groups.
What educational theoretical principles were evident in studies of WhatsApp® as a learning tool in medical education?
Five of the twenty-three studies articulated a theoretical basis for learningeither eLearning theory [41] or mobile learning [42] which guided the research design [23,27,31,34,35]. Two studies used their findings to subsequently suggest a theory informing learning in WhatsApp® groups; andragogy [34,43] and peer-assisted learning [28]. Notwithstanding, there was indirect evidence of technology-rich orientations throughout many of the other studies, in particular cognitive theory of multimedia learning [44] and Harasim's theory of online collaboration [45]. Some non-technological theories also bridged numerous studies. Several studies identified the importance of group learning in WhatsApp® users [18, 22-25, 28, 31, 32], reflecting influences such as an online community of practice [46], and social learning theory [47]. Motivational theory was also evident, in particular the ARCS model [48], whereby the convenience of WhatsApp® facilitated learner attention, the subject matter was relevant, learners were confident in the non-hierarchical environment and learner satisfaction was apparent in several of the studies' results [10,21,31,32].
Cognitive load theory [49] was relevant to studies where the high volume of learning material was thought to impede learning [23][24][25]. The user-friendly, familiar platform minimised extraneous cognitive load, prioritising the germane load of the online learning activities. Constructivism was a key theoretical construct in studies demonstrating learning built upon learners' contributions rather than on student-facilitator dynamics [10,18,22,28,31].

Discussion
In reviewing published literature on the role of What-sApp® in medical education, we have shown that, in line with its widespread use as an instant messaging tool, WhatsApp® has been evaluated in numerous subspecialties in both undergraduate and postgraduate settings. Notwithstanding the design decisions, the risks of bias and scant theoretical foundations, a total of sixteen studies described its use primarily for educational purposes, of which seven reported, in a total of 647 learners, an improvement in learner knowledge, and one reported a change in learner behaviour. Therefore, while our findings highlight the convenience, efficiency, versatility and popularity of WhatsApp®, they also suggest that it may be an effective educational tool. The main finding of our review however is that there is a need for well-designed rigorous educational research with strong theoretical foundations to more clearly define the role and benefits of learning with an IMA.
Does it matter that an online platform such as What-sApp®a social phenomenon that is cheap and popular is of any real educational benefit? Perhaps the answer depends on the purpose for which a WhatsApp® group discussion is designed. Medical educators should ideally use learning resources and instructional design principles which have a theoretical basis and have demonstrable learning benefits. Conversely, health professionals reaching out to other like-minded colleagues and peers to share clinical and learning resources, in a local, national or international setting are not bound by such rigorous educational standards; current evidence strongly suggests that WhatsApp® is a suitable resource for their purposes and that further research in this area is not warranted.
Although all of the included articles used WhatsApp® in a similar manner, of more importance were the individual study design decisions about how instant messaging could drive learning. In some studies, WhatsApp® provided an online space for healthcare staff to share experiences, opinions and resources [10,22,25,28,29], and to offer professional or emotional support to likeminded participants. These groups did not have a primary educational agenda, though educational elements were perceived throughout the discussions. Dedicated facilitators were not used, groups usually had national or international representation, all enrolled postgraduate users, and the duration of discussions were long, usually beyond 1 year. Educational assessment was limited to user attitudes.
Five other studies used WhatsApp® as a primary education tool with a pre-defined learning curriculum [23,27,30,36,37]. All groups had a dedicated faculty moderator, had a finite duration (2 days to 5 months), were mainly (4/5; 80%) in a local institutional setting and for undergraduate (4/5; 80%) learners. All five studies assessed Kirkpatrick level 2 outcomes, and notwithstanding some methodological flaws, all showed an improvement in learner knowledge or confidence following WhatsApp® learning.
Between these 2 groups were eleven studies using WhatsApp® as an educational tool but without a formal learning curriculum. In these studies, WhatsApp® discussion occurred on an ongoing basis (up to 2 years), with impromptu learning opportunities, stimulated by available clinical cases. Most (7/11; 63.6%) were in a postgraduate setting and most (9/11; 81.2%) were within a local institution or department. Most of these studies (9/11; 81.2%) assessed only learner attitudes, perhaps reflecting the flexible and ad hoc nature of this learning strategy.
The objectives of these three strategies are quite different; a safe online space for postgraduate peer discussions; discrete learning modules designed around the IMA; a continuous online learning environment driven by topical clinical cases. Guided by these 3 strategies, we propose a design model of IMA learning, drawing from Dabbagh and Bannan-Ritland's exploration-enactmentevaluation learning design framework for online education [50], in turn informed by socio-cultural and constructivist theories (Fig. 2) [51]. We propose that this stands distinct from less specific models of technology enhanced learning, eLearning or mobile learning. Our model may be a useful resource for educators and/or healthcare professionals planning to use an IMA in their practice. It may also help to fill the theoretical vacuum apparent in many of the educational studies reported in our review, addressing the truism that well-designed educational research should have a strong learning theory foundation [52].
Our findings add to existing literature in this field. In common with our findings, a recent review of 29 studies evaluating social media in graduate medical education identified a majority of descriptive studies with pre −/post-intervention assessment, Kirkpatrick level 1 and 2 outcomes, and "institutional-specific surveys" [14]. Their search however did not include studies evaluating WhatsApp®. A further review [13] of social media in Fig. 2 Proposed design model of instant messenger learning in medical education. Adapted from Dabbagh [50] undergraduate and postgraduate medical education identified 13 articles evaluating Facebook®, YouTube® and Twitter®, but not WhatsApp®. SM use showed "no correlation with student performance" (p374) and studies lacked "rigorous programmatic evaluation" (p374). In a review of the educational impact of Facebook®, Pander and Pinilla noted, in 16 studies, a preference for ongoing local learning rather than for curriculum-driven activities and "no conclusive evidence on the impact of the use of Facebook … on higher clinical competency levels and on patient-outcomes" [53] (p7). A very recent systematic review evaluating mobile hand-held devices for health professions described social media learning as an "unusual example of mobile devices supporting learning" [1] (p132). Our study therefore echoes and complements the findings of previous related literature, while strengthening the case for using IMAs in medical education and advancing a design theory for instant messenger learning.
Our review has potential limitations. It is possible that we have omitted relevant publications. Notwithstanding this, our sensitive search term, independent author searching, the updated second search strategy and the large number of identified articles when compared with other related reviews [54,55] suggest a comprehensive coverage in our search results. Our conclusions and inferences are drawn from a heterogenous group of educational studies with inherent design flaws and with limited theoretical bases. This raises concerns about the generalisability and credibility of the included quantitative and qualitative data respectively. Nonetheless, our findings suggest there is mounting evidence supporting the use of IMAs in medical education. Our proposed design model may help medical educators adopt a more formal approach to incorporating IMAs into their daily practice.

Conclusion
In conclusion, our review of WhatsApp® brings into focus the educational benefits of instant messaging and the strategies that have been used to employ this system in the medical setting. Our findings and the accompanying design model may provide a theoretical and practical framework for those planning to use IMAs in their educational practice. Well-designed research is warranted to further evaluate the role of IMAs in medical education but also to explore the utility of our design model to improve practice in this area.