The prevalence of chronic illnesses in children is on the rise, with up to a quarter of children in the general population having at least 1 chronic illness [1], including pediatric cancer. The situation is further complicated as up to 60% of chronically ill children have at least 1 co-occurring mental disorder [2], compared with the 10% to 20% prevalence in the general population [3,4]. Indeed, it has been well established that chronic illnesses predispose children for higher risk of developing mental disorders such as anxiety, depression, and behavioral problems [2,5].

Psychological support, although generally seen as an integral part of comprehensive and effective care [6,7], has not yet been unified across hospitals [8], leading to large differences in access to psychological care across regions. Differences in the availability and quality of psychological support increase the likelihood of leaving psychological problems unattended, which in turn may have long-term negative health effects and interfere with the treatment process of the underlying chronic illness [9]. Furthermore, untreated psychological problems may affect treatment compliance in the pediatric care setting [10,11] and may carry on into adulthood [12]. This attests to the need for the further understanding of psychological problems co-occurring with chronic illnesses, both at the level of risk factors and disease development but also in the search for effective novel intervention strategies. Timely and accessible evidence-based psychological support in the pediatric care setting might be a crucial factor in achieving desirable treatment outcomes.

The factors leading to mental burden among chronically ill patients are largely universal across different conditions, encompassing, for example, changes in daily routines, stressful states related to treatment procedures, and psychological uncertainty [2,13]. Relatedly, psychological problems are comparable between general and chronically ill populations, involving mostly symptoms associated with anxiety, depression, and behavioral problems [2,5,14]. Hence, various traditional intervention strategies aimed at reducing mental burden (eg, psychoeducational programs, solution-focused brief therapy, cognitive behavioral therapy [CBT], and mindfulness-based interventions) have been used interchangeably between chronically ill patients and patients without chronic illness [4,11,15-18].

In addition to traditional interventions, digital tools have been shown to be a promising avenue in delivering psychological support to patients. Currently, there are only a few mobile health (mHealth) apps for children with cancer that are publicly available, including: Pain Squad for pain management [19]; Re-Mission2 for patient empowerment [20]; and Super K where kids can fight cancer cells according to the game description [21]. However, these solutions have not integrated psychological care for patients and are primarily used for pain monitoring or empowering.

Rathbone et al [22] reviewed mHealth apps that used CBT principles for several psychological conditions and concluded that although apps can be effective tools in a health setting, the need for further studies that evaluate the effectiveness of various mHealth solutions is imminent [23]. Similarly, another recent review brought out the mental health app efficacy evaluation as a concern, showing that 38% of app store descriptions included phrases related to claims of effectiveness, whereas only less than 3% provided scientific evidence for such declarations [24]. Furthermore, previous findings also highlight the need to improve user engagement [25]. As an example of user engagement, the gamified version of smartCAT solution was found to be more effective than the nongamified version in delivering brief CBT treatment [26], therefore suggesting that gamification could be effective in achieving desired mHealth platform effectiveness targets. If the solution is effective, it can facilitate desired behavior changes [27] and ultimately lead to improved health outcomes.

Against this background, this study aims to assess the usability, acceptability, and feasibility of a gamified mHealth intervention, Triumf, both as a whole and by its individual constituent components. Furthermore, this paper explores the preliminary effectiveness findings of the mHealth game, Triumf, whereas further analysis with a pretest-posttest design and without randomization of the participants is published as a master thesis [28].

A newly developed digital health intervention, called Triumf, aims to reduce the negative psychological changes associated with chronic illness through an mHealth game. For clarity purposes, intervention or (mobile) game will also be used interchangeably to refer to the Triumf digital health intervention from hereon in. Importantly, the game has been designed in cooperation with pediatric patients with cancer, their parents, and care teams to determine illness-related burdening factors to develop a solution with maximum relevance to patients. Creating a game environment in collaboration with the key stakeholders has allowed building an intervention that approaches children in a way that is familiar to them. In addition, patients may be more empowered to comply with treatment if the intervention is delivered in a way that approaches the treatment process from a new angle [29]. By delivering care through a safe and familiar game environment, it is possible to fill in the gap between attractiveness and effectiveness, a common challenge in digital health solutions. Furthermore, as children in general are increasingly using their mobile devices for day-to day socializing and free-time activities, a new and still unused avenue for interacting, supporting, and educating pediatric care patients has been opened.

Together with the cooperative input from the partners and previous research literature, the intervention has been designed to serve as a basis to deliver effective psychological support that is universally applicable across different chronic illnesses, while maintaining the platform-level flexibility to consider disease- and region-specific differences. Thus, the overall gameplay, mechanics, and setup of the game and its principles are also applicable to other chronic illnesses, such as diabetes and asthma. The aim of the intervention was to offer patients psychological support and information about their health condition and emotions that would (1) help children to better understand their new health condition, (2) offer children external support to promote internal motivation to better cope with the illness, (3) offer children cognitive challenge and distraction, but also activity-based learning of healthy behaviors, (4) profile and screen the psychological well-being of children and offer psychoeducation and coping techniques accordingly, and thereby (5) support the formation of better self-understanding and constructive health behaviors, such as physical activity and diet. In short, through the continuous screening and support of various aspects, the intervention seeks to identify, prevent, and lessen the potential psychological problems and support behavioral change.

Triumf intervention covers various aspects of comprehensive care, and the game user experience is personalized and dynamic. The game learns and adapts with the user and offers individual and targeted experience. The intervention follows a predetermined structure in onboarding where the players are guided through the game narrative. Further gameplay, that is, accessing the educational module, entertainment games, and other elements of the intervention, is determined by the in-game choices made by the player. Furthermore, provision of psychological support is dynamically dependent on the individual emotional state of the patient.

This study has 2 main objectives. First, to evaluate the usability, acceptability, and potential preliminary effectiveness of the intervention among pediatric cancer patients. Second, to assess the feasibility of the study protocol of administering the intervention without the randomization of patients.

Of note, the usability and acceptability assessments were collected both at the level of the whole game and at the level of constituent components. Furthermore, the preliminary assessment of possible beneficial effects on the well-being of pediatric patients with cancer was carried out.

To fulfill the main study aims, the following analytical steps were carried out: (1) assessment of the relations between participants’ game behavior, medical treatment, self-reported well-being, motivation and attitudes, sleep, and evaluations on the intervention, (2) assessment of the perceived usefulness, ease of use, and enjoyment of the intervention as a whole and at component level, (3) general evaluation of cooperation and process with hospitals, and (4) general evaluation of applicability of the game in clinical context.

All pediatric patients with cancer aged between 7 and 14 years with a new or recurrent diagnosis of cancer, diagnosed no more than 1 year ago, were eligible to participate and were invited to the study (Figure 3). Participants were recruited by their medical doctors from Tallinn Children’s Hospital and Tartu University Hospital within a 6-month period starting from June 2018. Of the invited 15 children, 10 agreed (10/15, 67%) to participate in the study, with 1 participant withdrawing during the intervention owing to unfamiliarity with the game interface and without willingness to familiarize oneself. Thus, 90% of participants completed the study, and the sample used for analysis consisted of 9 pediatric patients with cancer with the average age of 9.1 years (SD 1.5; range 7-12), including 4 girls (44%) and 7 with Estonian as their native language (77%). Together with the withdrawn patient, 6 patients in total declined or withdrew their participation, with a mean age of 10.5 years (SD 2.4; range 7-14), 2 of them being girls (33%) and four (67%) having Estonian as their native language. Patients had various reasons for declining participation, where not being interested in the study or in mobile games was reported most frequently (4/6, 67%). However, the sample used for the final analysis did not statistically differ from the patients who decided not to participate or withdrew from the study (Table 2).

Each child and their parent gave written informed consent to participate in the study. The Research Ethics Committee of the University of Tartu (decision 283/T-32) approved the study protocol. Recruiting and participation in the study processes did not have deviations from the study protocol.

The study protocol involved a predetermined 60-day intervention period with a suggested 10 min of gameplay per day. The game was available in both Estonian and Russian. Android-operating smartphones were prepared and provided for the study period by the research team, and all participants had the option to continue the intervention following study completion using their own Android devices. Updates to the game were provided every other week to improve user engagement through new content. Technical support and information about game updates were offered to participants primarily via email.

Medical doctors recruited the patients to the study based on the eligibility criteria. Subsequently, information about the treatment and a health status for each participant were obtained from the medical doctors. This information included a general evaluation of the participant’s medical status during the study period, diagnosis and time from the diagnosis, notification of medical treatments that could affect mood, affect cognitive functions, or cause fatigue, time spent at the hospital according to the treatment plan, and outside planned treatment. Time from diagnosis, treatment that could affect participant’s capabilities, and total time spent at the hospital and at home were explored in the analysis. There were no deviations from the initial study protocol.

International Business Machines SPSS version 20.0 [45] and R free software environment [46] were used for data analysis. Owing to this study’s exploratory nature, aiming to provide relevant research questions for further similar studies, we have reported tendencies as measure averages and used small group size–based statistical comparisons and correlations. Tendencies and statistically significant results based on our small sample provide valuable insights that could be meaningful and therefore could be more stringently tested in future studies that employ larger sample sizes.

In detail, a Pearson correlation was used to evaluate possible associations between age, total time of all mini games played, exploring around the city environment (total of collected items), helping city kids to maintain healthy behaviors (total of interactions), number of days when the intervention was used, treatment that could affect mood (in days), time spent at home (in days), time spent at the hospital (in days), time from diagnosis (in months), SDT average score, study period average depressive symptoms, study period average anxiety symptoms, study period average attention problems, study period average self-reported general health score, average sleep score, score of questions answered in the qualitative interview (questions answered yes divided by total number of questions answered), subscore of usability from the qualitative interview, and subscore of acceptability from the qualitative interview. Goodman and Kruskal gamma was used for exploring relations between sex and the aforementioned variables, as well as treatment complexity and noted variables. A paired sample (dependent) t test was used to evaluate group differences of continuous variables and a chi-squared test was used for evaluating differences in dichotomous variables. Covariates were not included in the analyses. A conventional cut-off point for 2-tailed significance (P<.05) was used.

Sample characteristics are presented in Table 2. The average time from diagnosis to recruiting was 5.3 months (SD 4.1, range 0-12 months). Participants received medical treatment that could potentially influence emotional state and cognitive functions on 12.7 days (SD 18.5, range 0-60), spent 17.2 days at the hospital (SD 15.8, range 0-37), and 42.8 days at home (SD 17.8, range 23-60). During the study period, 8 participants out of 9 (89%) received chemo or hormone therapy, and 1 participant received the last treatment right before the study period. Out of the 8 participants, 4 (44%) had a more complex treatment regimen (eg, presence of infection in addition to main diagnosis). The group of participants who had a more complex treatment versus the group who received regular treatment presented a statistically significant difference in total time spent playing mini games (gamma=−.70, P=.02), that is, participants with more complex treatment spent less time playing mini games. No systematic psychological support in the sample was received during the study period.

Within the intervention environment, participants were divided into 3 categories on the basis of their profile of existing problems and strengths, resulting in 8 participants falling into the good psychological functioning category and 1 participant into the need for psychological support in more than 1 area category. On the basis of this, the majority of participants (8/9, 89%) received psychological screening and support approximately 2 times a day, with the exception of 1 patient, who received support more often. The participant who was categorized into the greater psychological support group had more depressive and anxiety symptoms, the lowest SDT score, and gave the lowest scores during the qualitative interview, but spent less time playing the game compared with the other group.

General comparisons (at the level of average scores) between the beginning and the end of the study period showed that there were no statistically significant changes in the in-game self-reported depression, anxiety, and attention problem symptoms, although a very slight improvement in depressive symptoms (from 1.15 to 1.19) and anxiety symptoms (from 1.46 to 1.50) was observed, accompanied with a very slight decrease in attention problems (from 1.31 to 1.29). For self-reported general health problems, there was a statistically significant change resulting in less problems (t₇=−4.4; P=.003).

Quantified usability and acceptability evaluations are presented in Table 3. All participants evaluated that the game was easy to use (100%) and 7 out of 9 (78%) would play it again. About 78% (7 out of 9) concluded that the instructions of the game were clear and that they learned something from the intervention, whereas 56% (5 out of 9) thought that there was too much to read in the educational module.

Acceptability evaluation was based on the total of 8 participants’ answers. All of them (100%) concluded that they liked Triumf game activities and that the intervention contained beneficial information and that the intervention was trustworthy; 94% (7.5 out of 8) liked the visuals and the characters of the intervention.

The qualitatively analyzed subjective free form feedback included both positive and negative feedback. Interviews showed that each participant liked the different modules of the game the most. Specifically mentioned by the participants were the features of the memory game (1/9, 11%), obstacle course (2/9, 22%), medical labyrinth mini game (1/9, 11%), collecting stars (29, 22%), helping citizens (1/9, 11%), the well-being questions (1/9, 11%), educational module (3/9, 33%), and characters in general (5/9, 56%). The game in general was liked by 67% (6/9) of the participants. Interviews indicated that all the study participants used educational module. From negative aspects, different kind of preferences regarding the intervention were presented, for example, possibilities to access more buildings and use collectables in more advanced ways. The commonly reported critique indicated that the game was perceived to be too short for the 60-day intervention and that it was more interesting to play the game in the beginning of the study period until approximately half of the study period. Parents in general evaluated that the study process was smooth, and they were given sufficient amount of information.

When offered to continue playing the game following the study period, on their personal Android device, 6 of the participants out of 9 (67%) wished to do so. The reasons for the 3 participants not continuing were lack of a personal Android device (1 participant, 33%) and not interested (2 participants, 67%).

To the authors’ best knowledge, this was the first study to evaluate the usability, acceptability, potential preliminary effectiveness, and feasibility of a personalized digital health intervention that uses a game environment to deliver psychoeducation, coping techniques, and treatment support. This approach allowed the researchers to use gamified, personalized therapeutics to offer comprehensive supportive care to pediatric patients. Previous studies have repeatedly stressed that it is highly important to integrate pediatric oncology-psychology research and standard of care [47], as well as the comprehensive supportive care of other chronic illnesses [6]. This study gave support to the prospect of delivering supportive care through a digital mHealth game in addition to traditional methodology.

The main findings of the study showed that the patients positively perceived the game, specifically with regard to their engagement, liking of the intervention, and learnings from it. The quantitative data showed that the mini games and the support module were used the most, whereas the qualitative findings also indicated the use of the educational module by all patients. However, on the basis of the qualitative interview evaluations, all patients expressed their own opinions about the most favorite parts, showing that all participants found something valuable to them which, in turn, offered more personal content. In general, our findings are supported by previous literature which has found mHealth solutions to be well suited for children, most of whom are savvy technology users [23].

Out of the specific modules we found, on the basis of consistent monitoring during the intervention, the well-being of patients improved when considering general health, but did not change significantly for depression, anxiety, and attention problems. The evidence of no statistically significant change in mental health aspects could be explained by the generally stable mental state of the patients during the study period. Consistent monitoring is crucial for prevention and early detection of psychological problems as, for example, the intensity of the treatment and the change of health may influence the mental well-being. Furthermore, it could be expected that mental distress is experienced, even when the overall health status has improved, as noted in previous findings showing that psychological problems could last into adulthood [12]. The evidence that children receiving more complex treatment played fewer mini games requires further investigation of the potential different game behavior in this group of patients.

We also found that lower anxiety levels resulted in higher intervention evaluations, which could indicate that feeling less anxious during the intervention, and the qualitative interview may influence the child’s view and opinion about the intervention. Furthermore, longer stays at the hospital resulted in less general health problems, which could indicate that receiving treatment at the hospital, and therefore receiving greater monitoring by medical personnel and timely adjustment in diagnosis, could result in less-reported general health problems. Unexpectedly, better sleep was only related to higher evaluations on the intervention but not with mental or physical health status. However, the importance of sleep should still be emphasized, and sleep quality should still be continuously monitored to observe whether sleep disruptions indicate short-term or long-term problems [48,49].

General attitudes toward health showed that a higher SDT score was related to more time from diagnosis, indicating more positive attitudes and motivation toward health. However, boys had lower motivation and attitudes compared with girls and thus may need more support to reach desired health outcomes. Focusing on an SDT-based approach could be suggested, as previous research indicates that components of SDT are associated with improved self-care among chronically ill patients, and thus positively related to treatment compliance, quality of life, and other health-related outcomes [50]. In addition, using general theories such as SDT as guiding theories for Triumf game development efforts, many of which having been widely used to explain the facilitation of motivated health behavior in wide variety of previous studies [51], enables us to adapt the intervention for different chronic conditions. SDT theory is also associated with the concept of mental toughness that refers to an individual’s capacity to be consistently successful in coping with difficult life circumstances [52], one of the goals of Triumf intervention.

We also found that higher evaluations on the usability of the intervention were related to higher age, which could indicate that it was easier for older patients to understand the guidelines and text in the game. The youngest patients gave feedback that they did not understand some of the words used in the educational module. A revision of the text has been included in updates following the study, although already now the text is presented in levels, which allows younger patients to access more simple explanations compared with older ones who are able to access further details. Boys and those who spent more time at home engaged more in the exploratory use of the game. This could mean that patients feel more secure in the hospital setting and at home look for more support or that boys are more curious about the possibilities of the game.

The usability and acceptability results indicated high usability in general, and very high acceptability, with only too much reading being specifically brought out with regard to usability. In previous literature, a shortened text has been used [44], or alternatively, a more engaging way of presenting information could be implemented. The usability and acceptability findings are in accordance with previous literature that has highlighted that gamification is more engaging for users than intervention without gamification [26] and that using mHealth solutions is a valuable resource to deliver psychological techniques [22]. In general, the choice to collect feedback by constituent components of the game proved to be a valuable source of information, informing future studies about the presence of both strengths and areas of improvement at the same time.

The study protocol was generally feasible and was followed without deviations throughout the study period, although some amendments are necessary. Feedback from parents and medical doctors did not bring any modifications to the study protocol from the co-operation and communication perspective. As the questions of qualitative interview appeared difficult to understand for younger participants, a revision is needed. A simplification of the wording or more optimal amount of questions should be considered. In addition, transforming the oral interview format into a digital form could be more optimal for larger sample sizes. On the basis of the refusal and completion rates and necessary changes to the study protocol and the intervention itself, it could be concluded that the protocol is feasible for the randomized controlled trial, with minor modifications, and the game could be considered applicable in a clinical setting.

This study has several major strengths. First, the rationale behind the intervention has been clearly supported by previous research, which is central in understanding the included elements and therefore allows for the selection and design of gamified components that are potentially most effective [53]. Second, the intervention was built in a digital environment only, which indicates players’ engagement without external social encouragement and creates possibility to evaluate the cost-effectiveness of digital tools [53]. Moreover, a novel technological methodology to monitor and support patients in a timely manner was used. This interventional algorithm took individual responses and tailored the proposed support components accordingly. This study was conducted in 2 different hospitals, allowing to evaluate its effect independently of the hospital treatment context. Furthermore, it was observed that patients liked different modules of the game the most, suggesting that the personalized way of delivering psychological support is a preferred method, as we showed that all participants were able to use beneficial components of their liking. Taken together, our work informs future studies and contributes toward the development of effective mHealth interventions by giving the empirical evaluation of delivering psychological support in a health-focused digital game environment.

Several patients and their parents stated that the game was interesting for approximately half of the study period (30 days), followed by a decrease in engagement. A decrease in the intensity of use of mHealth solutions over time has been recognized [23], although this could be related to the intervention design, especially from the motivation and engagement perspective. On the basis of the above, the intervention period could be shortened in the randomized controlled study or the game could undergo significant updates throughout the 60-day intervention period to keep patients engaged with the intervention (a process that is already in place to an extensive degree). It is possible that a shorter intervention time may be sufficient to induce behavioral change as habit formation time is very individual [54]. For example, studies on improvement of physical activity using health apps have shown that shorter interventions are more effective, although findings on intervention effects over time are still scarce [55]. Taken together, a change to the study protocol includes the reconsideration of the length of the intervention.

As this study involved patients aged 7 to 12 years, future studies are needed to evaluate the game among other age groups (eg, 5-7 years and 12-16 years). The game was evaluated in the context of pediatric cancer; thus, the intervention could also be evaluated among children with other chronic illnesses.

In addition, it was found that the 1 individual who needed psychological support the most ended up using the intervention the least. There have been observations that more serious psychological problems may interfere with engagement with the game [26]. Considering that digital intervention studies have been conducted on patients with mild to moderate symptoms [53], it should be investigated further. To continue, amendments to the game should be made to accommodate the needs of those individuals who need psychological support the most. Thus, it was concluded that the intervention should be delivered in 2 steps. During the initial stage, only educational module and fun components of the game would be accessible. Subsequently, psychological intervention would follow. Through this 2-step approach, new information would be given in different stages of the intervention. This might result in higher compliance through reducing the initial load of information. The abovementioned modifications to the intervention would be implemented before the next study.

Although data on general health attitudes were collected, it was proposed that these findings would need reassessment, as reported general health problem results might have been dependent on whether prompted questions were about general health behaviors or related to the current health situation. Separate measures of the current health situation and general perceived health condition and behaviors should be considered.

In conclusion, this study shows that delivering comprehensive supportive care through a game environment to pediatric patients is a feasible intervention strategy and is accepted by the patients and applicable in clinical context. This study showed that a game environment is a safe and engaging way of collecting real-time comprehensive data that can be used for personalized support.