Health Education Serious Games Targeting Health Care Providers, Patients, and Public Health Users: Scoping Review

Background: Serious educational games have shown effectiveness in improving various health outcomes. Previous reviews of health education games have focused on specific diseases, certain medical subjects, fixed target groups, or limited outcomes of interest. Given the recent surge in health game studies, a scoping review of health education games is needed to provide an updated overview of various aspects of such serious games. Objective: This study aimed to conduct a scoping review of the design and evaluation of serious educational games for health targeting health care providers, patients, and public (health) users. Methods: We identified 2313 studies using a unique combination of keywords in the PubMed and ScienceDirect databases. A total of 161 studies were included in this review after removing duplicates (n=55) and excluding studies not meeting our inclusion criteria (1917 based on title and abstract and 180 after reviewing the full text). The results were stratified based on games targeting health care providers, patients, and public users. Results: Most health education games were developed and evaluated in America (82/161, 50.9%) and Europe (64/161, 39.8%), with a considerable number of studies published after 2012. We discovered 58.4% (94/161) of studies aiming to improve knowledge learning and 41.6% (67/161) to


Introduction
Background Serious games have emerged as a promising educational technique across various domains [1,2]. Previous studies, including a survey study, have identified health care as one of the main targets of educational serious games [3][4][5]. In contrast to traditional educational techniques, the focus of serious games on health is partly derived from the fact that they provide individuals with a risk-free environment to practice high-stake tasks and experience unpredictable outcomes. Serious games also provided a unique educational platform to increase patient safety and reduce cost, which, in turn, has propelled the rapid development of new health education games [6,7].
User acceptance is key to the successful impact of educational serious games. Previous studies have assessed various user acceptance challenges of educational games [8,9]. These studies revealed that a wide range of users, including health care providers and medical students, accept serious games as a substantial and useful educational technique [10,11]. These studies also showed that clinical instructors consider serious games as an attractive and engaging educational tool [8,12]. Higher engagement is partly explained by the active learning tasks experienced by the users while interacting with an educational game [13,14].
Similar to other educational techniques, serious games require goal-relevant design, and their effectiveness should be methodologically evaluated [15]. Designing educational serious games requires multiple stages to ensure the engagement of potential end users in all phases of development, ranging from flowcharts and wireframes to multidimensional design and repeated user experience tests [16]. In addition, to increase the generalizability of educational serious games in improving learning objectives, they need to be rigorously evaluated across different user groups using various methods ranging from user studies to focus groups and clinical trials [17].
Several review studies have evaluated the design, development, and outcomes of serious games; however, only a few have focused on health education games [7,15,16]. One study conducted a systematic review of educational serious games for medical students and concluded that serious games should be evaluated before use in medical school curricula [17]. Another study conducted a meta-analysis of sex education serious games and concluded that serious games can be used effectively for promoting sexual health [18]. Moreover, another study evaluated the frequency and progression of health serious games across various domains, including clinical training, rehabilitation, and health education [5]. Considering the current rapid development of serious games, an updated scoping review focusing on health education serious games is lacking.

Study Objectives
This study offers an updated scoping review of health education games. The study reviews various aspects of the recent developments of educational health games designed for various user groups and provides a comprehensive review of the design characteristics and evaluation of such serious games. The study also addresses the gaps and weaknesses of the recent developments of health education games.

Overall Framework
The York framework was used to develop the general framework of this study [19]. We followed the following stages to guide the methodology of our search and analysis: (1) identifying the research question; (2) identifying relevant studies; (3) search strategy and study selection; (4) extracting information from the studies; and (5) collating, summarizing, and reporting the results.

Research Questions
Systematic reviews often focus on specific questions; however, scoping reviews explore questions with a broader scope [20]. In this review, the overarching aims targeted 3 aspects of health education serious games: general information, design specifications, and evaluation outcomes. Specific questions targeting each aspect of these health education games were discussed by a group of medical informatics experts (2 faculty members and 4 graduate students) and selected through a consensus process (Table 1).

Identifying Relevant Studies
The research team used an established framework [20] to develop an overall guideline used to identify if a study is considered relevant to the topic of the scoping review. The guide specified additional details about the articles on 3 perspectives of population, concept, and context ( Table 2). The research team then applied this guide to develop a detailed inclusion and exclusion criteria for the review process. Table 2. Overall guideline for the inclusion and exclusion of articles.

Description Items
The search strategy should not limit the educational groups targeted by the studies (eg, no limitations in terms of gender, age, health, or economic status)

Population
The search strategy will not discriminate about the underlying educational goal of the games and will be agnostic about the platform used to deliver the games (eg, no limitations on the learning techniques and no limits on software choices)

Concept
The search strategy will not limit the inclusion based on the affiliation of authors but will limit the papers to English articles published between January 1985 a and December 2018 Context a Previous systematic reviews identified no health education game studies published before 1985.

Search and Screening Strategy
The search strategy was guided by the overall inclusion and exclusion framework ( Table 2). After reviewing a handful of serious games articles, including review articles, and consulting informatics professionals, the study team developed a set of potential keywords to match the overall inclusion and exclusion criteria of the articles (Table 3). To accommodate a wider scope of articles, 3 sets of potential keywords were developed to cover gaming, health, and education domains and later were assessed for comprehensiveness while being converted into a unified and single keyword phrase.  The publication year of articles was limited from 1985 to 2018, and non-English articles were excluded. Only original studies published in a peer-reviewed venue were included. PubMed and ScienceDirect databases were used to perform the search.

Study Selection and Inclusion
As depicted in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagram in Figure 1 [21], the initial search retrieved 2313 articles from PubMed (n=1978) and ScienceDirect (n=335). After removing duplicates (n=55), 2258 articles were retained for screening. The search team used the inclusion and exclusion guidelines ( Table 2) to filter the articles. Screening the title and abstract of the articles resulted in the exclusion of 1917 articles. To harmonize the results and exclude exercise games, health games using hardware accessories (eg, sensors) or adopting commercial games that are specifically designed for exercise (with minimal educational content) were removed during the title and abstract screening. The remaining 341 articles were furthered screened after reading their full text, which resulted in 180 articles being excluded from the study. We specifically excluded studies assessing the negative aspects of violent games, as they are not considered health games. The risk of bias was not assessed, as this scoping review did not intend to systematically review and evaluate all effective interventions. In the end, 161 articles were determined to meet all inclusion and exclusion criteria and were fully reviewed in the data extraction phase (which further split the studies based on their target groups of health care providers, patients, and public users).

Data Extraction
A data extraction form was developed by the research team after the determination of the conforming variables with the research questions and the study's goal ( Table 4). The variables were categorized based on the study's 3 major aspects (ie, general information, design, and evaluation; Table 1). The data extraction form was shared with all study team members and was finalized after addressing all remaining questions and comments.
In total, 2 reviewers used the data extraction form to independently mine 9.9% (16/161) of the articles. The kappa coefficient score between the 2 reviewers was calculated at 76%. Both reviewers further discussed discrepancies internally and were trained once more to reach a high consensus rate (ie, reaching a kappa score >90% in the second round of coding). Two reviewers then extracted data from the rest of the articles. In the case of nonconformation between the reviewers, a third reviewer was consulted to reach consensus.

Data Collation and Analysis
Data extraction results were collected in 2 Microsoft Excel sheets managed by each of the reviewers. The Excel sheets were then merged to generate the final set of results. Excel functions were used to populate the summary statistics and perform a frequency analysis. To analyze the captured data, we applied a frequency analysis for all variables of interest and presented the results in various chart formats. To improve the interpretability of the results, we stratified all findings into the 3 user groups of health care providers, patients, and public users.

Overall Findings
The findings of this review were categorized based on health education serious game aspects and specific questions identified earlier in the review (Table 1). Significant study findings were grouped into the geographic distribution of the studies, publication year, type and goal of the studies, target groups, gameplay duration and repetitions, intervention specs, length of the follow-up period, and the use of multidisciplinary teams.

Type of Study
In this review, articles were categorized into 2 classes of interventional (77.6%) and observational (22.4%) studies.
Among the interventional studies, 44 were randomized clinical trials, whereas 81 of the studies were quasi-experimental ( Figure  4). A few studies have assessed the educational health games using a qualitative approach (eg, design protocols and surveys).

Target User Groups
We studied 3 specifications of the target groups: age groups, user groups (eg, health care provider, patients, and public users), and medical condition (specific to patients).

Age Groups
We used educational age ranges [22] to analyze age groups (  After combining the age groups into larger age bins for games targeting patients and public users, the age bin of 0 to 18 years showed the highest number of studies (n=65), compared with 18 to 65 years (n=49) and 65 years and older (n=3, Figure 6). Logically, all games designed for health care providers fell within the age range of 18 to 65 years (not shown in the figure).

User Groups
Overall, the user groups were divided into 3 general categories (

Medical Conditions
Of all reviewed studies, 42 (

Duration of Gameplay
Duration of gameplay (ie, time spent on interacting with the game) has been reported in 101 studies (101/161, 62.7%). Figure   7 depicts the distribution of gameplay duration among these 101 studies. The most common gameplay duration was 30 to 45 min (n=31) and the least was less than 15 min (n=13).

Number of Game Repetitions
Of all 161 studies, 61 health games (37.9%) mentioned the number of times the game can be repeatedly played (Figure 8). Among these studies, 2 general categories of repetitiveness were observed: (1) articles that limited the number of times a game can be played (n=58) and (2) articles that set no limitations for the number of repeats and users were allowed to have an interaction with the game with unlimited repetitions during the intervention time (n=3; a subset of >35 bar in Figure 8).
Among the articles of the first category, games with less than 5 repetitions had the highest number of articles among health care provider and public user groups. Games designed for patients had 20 to 25 as the highest number of repetitions ( Figure  8).

Duration of Intervention
The duration of intervention was mentioned in 94 of the reviewed studies (Figure 9). The duration of intervention varied between 1 week and 8 years. Most studies (34/94, 36%) had a time range of less than 1 month.

Time Between Intervention and Posttest
The time between intervention and posttest was reported in 59 (59/161, 36%) of the reviewed studies ( Figure 10). Two categories of time between intervention and posttest were seen: (1) conducting a test right after the intervention (48/59, 81%) and (2) conducting a test after at least 1 day is passed from the intervention (11/59, 18%). The time ranged between 1 and 12 weeks among the latter group of studies.

Follow-Up Duration
Follow-up duration was collected by a few of the studies (32/161, 19%; Figure 11). Among these studies, 97% had a follow-up duration of 1 week to 6 months. Only 1 study reported a 2-year follow-up period.

Multidisciplinary Teams
Although the use of multidisciplinary teams to design health games is strongly recommended, only 42% (68/161) of the reviewed games either explicitly mentioned the use of such teams or implicitly mentioned the involvement of such experts (eg, instructional, clinical, and user experience) in the development and assessment of the games (Figure 12). Analyzing the use of multidisciplinary teams over the years of publications did not show any significant trends.

Principal Findings
Serious games are increasingly recommended as effective techniques to improve health education [23][24][25]. Multiple studies have assessed educational serious games in different fields of health, ranging from preventative screening to management of chronic diseases [9,[26][27][28]. Over the last decade, a growing number of these studies have measured the efficiency and effectiveness of serious health education games using randomized trials of patients and clinicians [29][30][31][32][33][34][35]. Our review provides an updated scoping review of the underlying patterns and gaps of studies assessing the value of serious games in improving health education for health care providers, patients, and public users.

Serious Game Development, Target Groups, and Topics
Our results confirm the concentration of educational health game development in developed countries of North America, Europe, and Asia, thus lacking the opportunity to target educational needs of low-income countries by adopting contextualized or localized serious games. Given the increased availability of smart electronic devices and the penetration of the internet in developing countries [36], more research is needed to develop and assess the impact of serious games targeting specific health educational needs of such populations (eg, infectious diseases) [37]. Given the peak of health game development in recent years (between 2015 and 2018) and the potential higher commercialization of such solutions, focusing on emerging topics of developing countries can benefit larger populations in need [38].
Our results show that more than one-fourth (26.1%) of the health education games focused on higher education needs and targeted medical students and staff. The fact that most of these studies are using health games as an intervention and a considerable number of them have assessed their effectiveness using experimental designs (including randomized trials) provides a growing opportunity to assess the effectiveness of such serious games. It can be anticipated that medical schools (especially in North America and developed countries of Europe and Asia) will gradually incorporate such interactive solutions (ie, serious games) in their common educational curriculum in the near future. In addition, given the low cost of technology needed to use such games in medical and nursing schools, open source or free versions of such serious games can tremendously help reducing educational disparities in clinical sciences among low-resource countries.
An interesting trend revealed by our results is the gradual move from developing disease-specific serious educational games (eg, educational games for diabetes) to targeting broader public health topics (eg, safety and nutrition). Given the impact of health topics such as safety and nutrition on population-level outcomes, the future of educational health games may entail a larger coverage of the general healthy population rather than patients with specific diseases (ie, helping to bridge pubic and population health outcomes [39,40]). The wider target groups of end users can potentially translate into increased market opportunities for educational health games as well as sustainable commercialization over the long term.

Game Design and Learning Outcomes
Educational games primarily aim to increase awareness and knowledge among the players. Nonetheless, the ultimate goal of educational health games should include a behavioral change in the end users, thus producing a lasting effect. Developing such complex health education serious games, however, is challenging as it requires the participation of multidisciplinary team members to address various game play perspectives, ranging from principles of design to psychology of behavioral change. In this review, the evaluation of the studies revealed that the use of multidisciplinary teams to design health education games is strongly recommended, but it is accomplished occasionally by game developers. Indeed, according to our evaluation, almost half of the studies were deprived from a multidisciplinary team. Continuous collaborations among the members of the multidisciplinary expert team in addressing various aspects of game design and development are strongly recommended to improve the educational experience of the users and potentially improve the impact of health education games.
Most studies reviewed in this study had a short duration of game play with minimal repetitions and limited follow-up periods. In addition, 6 studies did not report an effective intervention and could not achieve desired educational outcomes. These limitations can be attributed partly to the lack of expert user experience designers participating in the study teams [41]. Furthermore, although factors, such as time of intervention, time between intervention and posttest, and duration of follow-up, are critical in achieving long-term knowledge gains, generalized evidence on what factors with what frequency and length works for what age-range is still lacking. More work is needed to established common design guidelines on how to engage different user groups with best game specifications for health education for both patients and medical staff groups.

Comparison With Prior Work
Previous scoping reviews of health games included all types of serious games, ranging from educational to behavioral change and exercise-focused games [4,5]. Several reviews have predominantly focused on chronic diseases [42][43][44], whereas other reviews have focused on specific subgroups of educational games, such as games targeting clinical staff and students [10,12,14,45] or patients with distinct medical conditions or diseases (eg, diabetes, asthma, or obesity) [46][47][48][49][50][51][52][53]. Owing to either inclusivity (all health games) or exclusivity (specific educational games) of the previous reviews, these reviews do not reveal the overall trends in educational health games across various target users (ie, health care providers, patients, and public health users). This study provides a fresh review on the latest developments and trends in serious games targeting health education, regardless of the target population or observed clinical context.

Limitations
Our work had multiple limitations. First, we specifically narrowed our search on educational serious games for health. Our review should not be considered a systematic review of all types of health games. Second, our review did not conduct a statistical analysis to measure significant differences among various target groups (eg, measuring significant differences between patients and public health users). Third, we did not perform a meta-analysis, as the studies did not report enough details about their population to perform such analysis. In addition, the outcomes of interest varied greatly, and a meta-analysis was not appropriate for our review. Fourth, we only included English publications, thus unavoidably excluded studies that might be conducted and published in other developed countries not using English as their primary scientific language. Fifth, the variability of reviewed studies limited our review to encode generic specifications, such as publication date, study location, and target groups. More research is needed to tease out the details reported by various subgroups of these studies (eg, some studies reported exact outcomes of interest). Finally, we used PubMed and ScienceDirect as our search engines; however, other search engines might provide additional studies not indexed in the aforementioned databases.

Future Work
Future work should include a larger systematic review, including additional search engines and, perhaps, focusing on the educational games designed for either patients or clinical providers. Additional work on reviewing educational outcomes of interest among these studies and the effectiveness of the health games in achieving them is also needed.

Conclusions
Serious games are increasingly used for health education. This study offers an updated scoping review of the studies assessing the value of serious games in improving health education. Most educational health games are still developed in high-income countries, but a surge of new games focusing on healthy behaviors (eg, nutrition and safety) has been observed in the last 5 years. Game developers need to use multidisciplinary teams to improve the design of the serious games in keeping the end users engaged for a longer interaction and potentially more effective educational health outcomes. Investment in serious games, as a low-cost educational tool for both patients and medical providers, can potentially help to fill the gap for health education in developing countries.