Games and Health Education for Diabetes Control: A Systematic Review with Meta-Analysis

Finding methods to improve people’s diabetes control and management is important to prevent its complications and maintain the quality of life. The aim of this review was to assess the effect of games on the blood glucose level (glycated hemoglobin (HbA1c)). A systematic review and meta-analysis were made. Pubmed, Scopus, and CINAHL databases were consulted in July of 2020. Ten studies were selected as a final sample, most of them being clinical trials using games to improve diabetes control. Half of the studies had samples between 8 and 14.9 years old and the other half between 57 and 65 years old. The studies informed about using applications/games for mobile phones, game consoles, and board games for diabetes education and management. The meta-analysis was performed with 4 studies showing a mean difference of 0.12 (CI 95% 0.57, 0.33) of HbA1c in favor of the intervention group with p > 0.05. Games are positive for diabetes health education and promoting healthier lifestyle, but their impact on HbA1c is low.


Introduction
Diabetes Mellitus (DM) is an endocrine disease, with a certain hereditary component, characterized by an increase in blood glucose. It can be classified into type 1 (DM1) or insulin-dependent diabetes, type 2 (DM2) or non-insulin-dependent diabetes mellitus, and others less frequent such as gestational diabetes and sub-categories of DM1 like Latent Autoimmune Diabetes in Adults (LADA) and Maturity Onset Diabetes of Youth (MODY) [1,2]. DM1 is characterized by autoimmune B-cell destruction, which generally leads to absolute insulin deficiency, and it usually occurs in children or young adults. The characteristic symptoms are: polyuria, polydipsia, and, approximately a third, present diabetic ketoacidosis. In DM1, there are no factors that

Information Sources and Search
The following databases were consulted: Pubmed, Scopus, and CINAHL. The search equations are indicated in Table 1. The search was completed on 31 July 2020.

Study Selection and Risk of Bias
The results obtained in the databases were first selected by reading the title and abstract. Then, the full text was read. After that, a critical reading was made and finally, a reverse search, reading the references of each study, was performed with the included studies. The selection process was completed independently by two members of the team, consulting a third person in case of disagreement. The critical reading was done with CASPE critical appraisal checklists. In order to assess the level of evidence and grade of recommendation of the included studies, the Oxford Center for Evidence-based Medicine (OCEBM) [21] recommendations were followed.

Data Items and Collection Process
A data collection notebook was used. The following variables about the characteristics of the sample were collected: year of publication, country of study, study design, sample, game or gamification characteristics, and main results in relation to glycated hemoglobin (HbA1c) and diabetes management.

Data Analysis and Summary of Results
All the studies included in the review were descriptively analyzed and categorized in 3 sections (Electronic and board games for diabetes knowledge acquisition, Electronic simulation games for solving problems and situations related to diabetes, and Electronic games for adherence to physical exercise in people with diabetes). The effect size of the intervention for the reduction of HbA1c (mean difference between control and intervention group with a confidence interval of 95%) was calculated with a random effect meta-analysis done with Review Manager 5.3. For the meta-analysis calculation, the necessary data from each study were the mean, standard deviation, and sample size of control and intervention group. Before the meta-analysis, a sensitivity analysis (to check that the effect size did not significantly change after eliminating each study from the meta-analysis) and an Egger test for publication bias were performed. I 2 test was used for heterogeneity analysis. The Cochrane handbook for systematic reviews of interventions was followed for the analyses [22].

Study Selection and Study Characteristics
The search showed 789 results that were reduced to 54 after eliminating duplicates and reading the title and abstract. Then, after reading the full texts, 44 studies were eliminated due to the fact that they did not include information about HbA1c, had no people with diabetes sample or were not using games or gamification process, leaving a final sample of n = 10 studies [23][24][25][26][27][28][29][30][31][32] (Figure 1). The 70% of the studies were randomized clinical trials, 10% cohort studies, 10% cases and controls studies, and 10% qualitative studies. Most of the studies were performed in the USA (70%) and others studies were performed in Canada, Switzerland, and France. The higher sample was n = 456 and the lower n = 20. The 40% of the studies were centered in people with DM1. The main characteristics of the included studies and the results about using games and gamification for diabetes are summarized in Table 2.  The 70% of the studies were randomized clinical trials, 10% cohort studies, 10% cases and controls studies, and 10% qualitative studies. Most of the studies were performed in the USA (70%) and others studies were performed in Canada, Switzerland, and France. The higher sample was n = 456 and the lower n = 20. The 40% of the studies were centered in people with DM1. The main characteristics of the included studies and the results about using games and gamification for diabetes are summarized in Table 2. Smartphone game following physical activity guidelines to increase physical activity behavior. The use of the game and reaching the activity goals make the person progress in the storyline of the game. Game for Ipad with a 12-min duration. The game simulates multiple real-world pressures from job and family for people with diabetes including excuses for non-adherent health habits. The patients have to deal with those situations.
After 3 months: -HbA1c in the intervention group significantly decreased from 10.7 to 9.6 but no comparison was done with the control group. - The adherence to medication, diet and exercise significantly increased but no comparison was done with the control group. 1a/A

Electronic and Board Games for Diabetes Knowledge Acquisition
The use of mobile applications is very widespread, since the smartphone is an essential element in our day to day life. One study with children used an application that automatically received blood glucose levels through Bluetooth [23]. The app generated graphs of blood glucose level evolution and gamified the glucose control giving rewards when reaching healthy goals (keeping blood glucose level in range), such as free downloads of iTunes songs [23]. In a 12-week period, HbA1c did not change significantly, but improved the glycemic control of the sample and 88% of participants assured that they would use the app after the study since they found it useful for making decisions about their health [23].
Another study used the DiaSocial app for 14 weeks to promote adherence to physical exercise, improved nutrition, and adherence to drug therapy in veterans (mean age 65.4 years) with DM2. Training about the app was done in groups and the app gave points depending on people performance of healthy activities (maintaining blood glucose level, recording blood glucose level, adequate nutritional intake, minutes of exercise, and adherence to medication) [24]. Higher final scores were observed in people with greater reduction of HbA1c [24]. The mean HbA1c in the control group, which received a classic health education of recommendations and control, was 8.78 and in the intervention group, it was 8.06 [24].
Other authors tested two games as a method for acquiring knowledge about diabetes in a sample with 59 years old and DM2 [25]. The intervention group received a game about diabetes self-management and a brochure on civic standards. The control group received a game of civic standards and a brochure on diabetes self-management standards. In addition, two questions about diabetes self-management education were sent two days per week through a mobile application or email, and the correct answer and the explanation are shown automatically after the user's answer. Correct answer gave points to the person and, with the points, a ranking of the participants was elaborated. The intervention group, at 6 months, showed an improvement in HbA1c, but not statistically significant, while empowerment over the disease increased [25].
In the study by Kumar et al., 2004 [26], performed with children and adolescents, a motivational game was used in the intervention group and the participants had the Dia-BetNet software on their personal digital assistant. The software integrates outcomes for diabetes management (previous blood glucose levels, insulin doses, and carbohydrates consumed) and challenges participants to predict their next glucose level based on a graphical display [26]. The participants received points for playing the prediction and more points for its accuracy. The intervention increased the number of blood glucose checks per day, but did not provide a direct improvement in HbA1c, but a decrease in hyperglycemia [26].
Another study used board games to encourage communication between people with diabetes and, with it, the acquisition of knowledge [27]. The study observed how, after playing the board game, participants improved their HbA1c [27]. The intervention was based on different sessions where a "conversation" on a Conversation Map was "played". People took advantage of the experiences of others participants and also another maps added 4 different educational tools to the previous one: general description of diabetes, healthy eating, monitoring and use of their results, and natural course of diabetes. The intervention lasted 3 years and it demonstrated an improvement in HbA1c in the people who played compared to the people who received classical care. Game players decreased their HbA1c to 6.96 while the control group decreased to 8.27 [27].

Electronic Simulation Games for Solving Problems and Situations Related to Diabetes
Brown et al. [28] used a Super Nintendo Entertainment System ® video game. The game was called "Packy and Marlon" and was designed for young people, between 7 and 15 years, with DM1. The game simulated a diabetes camp where the protagonists had to maintain their health through specific behaviors (food choices, doing too little or too much exercise, and what to do according to the circumstances) [28]. This intervention did not significantly improve HbAc1, but knowledge and problem solving improved and it also decreased emergency room visits by 77% [28]. Another game simulation for problem solving in children and adolescents was "L'Affaire Birman" [29]. The protagonist of the game is a teenage boy named Alex who has flexible insulin therapy and the player, through his decisions, must maintain Alex's health. The intervention did not show a significant improvement in HbA1c, but it did improve knowledge about carbohydrate counting and insulin administration [29].
Another study used the simulation game "Patient Partner" that is designed for maintaining health through real life situations. Contrary to the previous two games, this was a mobile app focused on DM2. The use of this application decreased HbA1c and increased adherence to diet, exercise, and medication [30].

Electronic Games for Adherence to Physical Exercise in People with Diabetes
One study that used a smartphone game showed improvements in 24 weeks in the number of steps per day, but it did not influence a significant modification of HbAc1 [31]. Finally, a study that used a Wii Fit Plus sports game for 12 weeks 30 min a day improved adherence to physical activity and decreased HbAc1d from 7.1% to 6.8% [32].

Meta-Analysis Results
From the ten studies included in the review, only four had the necessary information (mean and standard deviation after the intervention for the control and the intervention group) for the meta-analysis; 50% of the studies had samples with DM2. The sample of the intervention group was n = 161 and for the control group, it was n = 149. The means differences in the percentage of HbA1c was −0.12 (95% confidence interval −0.57, 0.33) in favor of the intervention group, but without being statistically significant. The forestplot is shown in Figure 2. From the ten studies included in the review, only four had the necessary information (mean and standard deviation after the intervention for the control and the intervention group) for the metaanalysis; 50% of the studies had samples with DM2. The sample of the intervention group was n = 161 and for the control group, it was n = 149. The means differences in the percentage of HbA1c was −0.12 (95% confidence interval −0.57, 0.33) in favor of the intervention group, but without being statistically significant. The forestplot is shown in Figure 2.

Discussion
The use of game and gamification processes reduces HbA1c level, but it does not show significant differences with the control group. Mobile applications linked to games and gamification processes with educational health content are an advance for people with diabetes, but their use does not make an effective difference with other usual interventions. In the case of DM1, when debuting at an early age, an effective method of communication and transmission of knowledge is by playing [33]. This can help to learn the administration of subcutaneous insulin to children [34] and other things like proper nutrition [35]. Although it seems that the games are more oriented to children and adolescents, there are also games for adults that are effective [36].
Even though various studies demonstrate the effectiveness of using games through mobile applications for health education in people with diabetes [37,38], some authors said that technology can be a barrier for some populations [39]. Research in this field should continue to advance in order to make the use of mobile applications easier and more practical [40], since, in some cases, patients may not learn the basic concepts needed for good self-control of the disease [18,41]. Moreover, health professional training is required to recommend and apply these technologies, knowing that they are safe and reliable [42]. The use of innovative therapies such as the use of games in people with diabetes is a very useful tool in clinical practice [43] for problems' solutions, since it improves adherence and minimizes complications [44].
In addition, the use of games in mobile applications has recently been implemented. These facilitate access to people with diabetes and are an educational tool that can improve HbA1c levels

Discussion
The use of game and gamification processes reduces HbA1c level, but it does not show significant differences with the control group. Mobile applications linked to games and gamification processes with educational health content are an advance for people with diabetes, but their use does not make an effective difference with other usual interventions. In the case of DM1, when debuting at an early age, an effective method of communication and transmission of knowledge is by playing [33]. This can help to learn the administration of subcutaneous insulin to children [34] and other things like proper nutrition [35]. Although it seems that the games are more oriented to children and adolescents, there are also games for adults that are effective [36].
Even though various studies demonstrate the effectiveness of using games through mobile applications for health education in people with diabetes [37,38], some authors said that technology can be a barrier for some populations [39]. Research in this field should continue to advance in order to make the use of mobile applications easier and more practical [40], since, in some cases, patients may not learn the basic concepts needed for good self-control of the disease [18,41]. Moreover, health professional training is required to recommend and apply these technologies, knowing that they are safe and reliable [42]. The use of innovative therapies such as the use of games in people with diabetes is a very useful tool in clinical practice [43] for problems' solutions, since it improves adherence and minimizes complications [44].
In addition, the use of games in mobile applications has recently been implemented. These facilitate access to people with diabetes and are an educational tool that can improve HbA1c levels [45]. Health education provides many benefits to patients [46]. Nonetheless, it is true that it requires a responsibility on the part of the patients [47] and that a standardization and simplification of the different applications is necessary, since there is a decrease in communication with health professionals [48].
Making a person adhere to the recommendations and treatments that professionals give is not easy. It has been observed that patients follow medical prescriptions better if they understand the treatment. Several authors affirm that the integration of technology-based exercise programs can have a positive effect on the adherence of the diabetic patient, since they produce an increase in enjoyment and make it easier to perform these programs [49,50]. Patients think that everything is important and these applications help them make decisions [51]. The use of these methods encourages the user to carry out physical activity, producing favorable results in the medium-long term and leading to an improvement in balance or flexibility [52] as well as an improvement at a psychological level, reducing the levels of depression [53,54]. However, there are authors who state that practicing physical exercise through the use of electronic applications or games only produces psychological benefits and does not improve adherence [55]. Likewise, other authors say that they do not observe this type of benefits, and affirm that adherence is greater when performed outdoors [56]. In turn, some authors claim that these methods are beneficial for exercising continuously from home. Decreasing barriers such as displacement, the weather, or going to activities where there are many people, can generate some insecurity in some people and this option solves this problem [57].
The study has some limitations. Although the number of games and gamification processes for diabetes is growing, the number of studies analyzing its effect on HbA1c is low. Most of the studies come from the USA and all of them are from the Western countries, so the results should be taken into account with caution in countries with different culture and life style. Future research should analyze which factors promote the attractiveness of a game to create more enjoyable and desirable games. Additionally, in clinical practice, the implementation of the games with better results would be of interest for long term follow-up studies. Finally, the use of behavioral sciences [58] to identify the key information for the design and development of games for diabetes would be of interest for developing better games.

Conclusions
Games and gamification processes are beneficial for education in people with diabetes and to promote adherence to healthy lifestyle habits. However, their impact on glycated hemoglobin (HbA1c) does not appear to be clinically relevant. It is necessary to investigate into this area to find the mechanism to create more effective games for diabetes management and glycemic control.

Conflicts of Interest:
The authors declare no conflict of interest.