An Emotional Bias Modification for Children With Attention-Deficit/Hyperactivity Disorder: Co-design Study

Introduction

Background

Attention-deficit/hyperactivity disorder (ADHD) is one of the common neurodevelopment disorders. Children with ADHD typically display symptoms such as hyperactivity, impulsivity, and inattention [1]. These individuals have inherent underlying difficulties with emotional regulation [2]. The advances in experimental psychology have led to better understanding of emotional biases and methods by which these biases could be modified [3]. Previous studies have reported that emotional biases are most prevalent among those with the combined subtype of ADHD. Children with this subtype tend to have difficulties in comprehending the emotional states of others and recognizing facial emotions and emotional cues [2]. Previous studies have described how the visual probe or dot probe task is applied in the assessment and potentially in modifying these biases. In the visual probe task, individuals are presented with 2 stimuli on the screen simultaneously in the assessment phase. The presented stimuli are images showing different emotional cues, for example, an angry face or a neutral expression. The set of stimuli would disappear, and a probe would replace either stimulus. Individuals are required to indicate the position of the probe as rapidly as possible. For purposes of assessment, individuals may have positive emotional biases when they respond more readily to probes that replace the emotional stimulus instead of the neutral stimulus. For bias modification, the contingency of which stimulus is being replaced could be altered in a way that is similar to the application of the visual probe task for other psychiatric disorders [3]. Although there have been previous studies examining the effectiveness of cognitive bias modification for anxiety and depression among children and adolescents with other psychiatric disorders, for example, anxiety and depression [4,5], pioneering studies have been conducted in exploring emotional biases in the same population. In a recent study by Cremone et al [3], the authors reported the existence of emotional biases among children and adolescents with ADHD, and the amount of sleep affected the magnitude of the underlying emotional biases.

Although the visual probe task appears to be a relatively simple task to administer for the measurement of underlying emotional biases, and potentially for modifying biases, it is not without its inherent limitations. The numerous repeats that an individual needs to complete make such an intervention laborious. In recent years, serious games and gamification technologies have been considered for conventional psychological interventions [6,7]. It is believed that the use of these technologies would enhance engagement in the short and long term, promote self-empowerment, and improve existing skills [6]. As highlighted by Zhang et al [8], gamification, when applied to cognitive bias modification interventions, could help to reduce the repetitiveness of the game and increase motivation to train. As evident from previous study by Zhang et al [9], which involved conducting a series of participatory action research workshops involving health care professionals and patient service users, gamification could be used to address the limitations of the visual probe task and enhance the conventional task.

More recently, regarding ADHD, there have been further studies examining the evidence base of existing games for individuals with ADHD and the potential of gamification for serious games targeting individuals with ADHD. Jiang et al [10] conducted a scoping review of the currently available mobile games on several databases, and they reported that most of the existing games were focused on managing individuals with symptoms and less so on symptoms and diagnosis. Although the studies have shown an improvement in the performance of the children across the interventions, these studies were limited in several aspects, such as the inclusion of a small sample size and lack of a control group for comparison, and these factors influence the overall effectiveness of the study [10]. Nonetheless, the 19 studies highlighted the potential of considering a serious game approach when conceptualizing new interventions, and they also revealed that the most common techniques were that of having participants to respond to various cues, remembering details, or making association between different entities [10]. In another recent study by Sujar et al [11], the authors, having synthesized the evidence from literature, reported that future games that are designed for individuals with ADHD ought to consider the following aspects: the underlying mechanics of the game ought to be based on some form of cognitive exercise and therapeutic strategies that may be helpful include having levels of difficulties, a motivational element, time constraint for the task one has to undertake, and some form of reinforcement [11]. The findings reported by Sujar et al [11] are consistent with the aims and objectives of this study, which, as elaborated subsequently, aimed to seek the perspectives of health care professionals as an integral step in co-designing an intervention to modify emotional biases in children.

Objectives

There has also been increased recognition of the potential of participatory action research methods and how it could help to make the conceptualized interventions more relevant [12]. Therefore, the aim of this study was to use such a method in refining the conventional emotional bias visual probe task for children and adolescents. The primary aim of this study was to collate health care professionals’ perspectives on the limitations of the existing visual probe task and to determine if gamification elements were viable to be incorporated into an emotional bias modification task. We sought to answer the following research questions: (1) What were health care professionals’ perspectives on the conventional emotional bias task? (2) What were health care professionals’ perspectives on existing gaming interventions for ADHD? (3) Would gamification be appropriate, and if appropriate, what strategies could be used?

Methods

Study Design

Principles of participatory action research were used for this study. A co-design workshop was conducted, and relevant key stakeholders (ie, health care professionals) were invited. Only health care professionals were invited, as they had knowledge of the psychiatric conditions and were best able to advise how the task could be modified, while adhering to the evidence base.

Study Setting

Health care professionals from the Department of Development Psychiatry, Institute of Mental Health, Singapore, were invited to participate. A diverse group of health care professionals was invited, and it included both psychiatrists and psychologists. We had originally planned to recruit occupational therapists also, but none of them expressed an interest in participating in the study.

Sample Size

On the basis of our previous protocol [13], we planned to recruit 8 participants. We managed to recruit 50% (4/8) of the projected participants for the focus group. However, there was a reduction in the number of participants recruited, mainly in part owing to the ongoing COVID-19 pandemic and the difficulties for individuals to commit to research studies as they have other clinical or administrative roles.

Details of the Co-design Workshops

Owing to the current COVID-19 pandemic and local governmental restrictions, we applied for the conduct of the workshop via the web. Therefore, all participants (4/4, 100%) who have had expressed interest were contacted separately to sign the informed consent form. The participants also had to complete a baseline demographic questionnaire individually. This questionnaire collated information regarding age, sex, and years of experience in treating children and adolescents with psychiatric disorders. Upon completion of the questionnaire, the principal investigators liaised with each of the participants separately to identify a common time for conducting the web-based workshop.

The workshop was subsequently conducted on May 19, 2021, via the web. Both principal investigators facilitated the workshop and recorded the field notes. All participants (4/4, 100%) were informed of the study’s rationale and the session’s specific objectives. Participants were also informed that their responses and comments were confidential and that the session would be audio-recorded. All participants (4/4, 100%) were given participant numbers to identify themselves, to provide further anonymity. First, participants were shown an example of the visual probe task paradigm that has been traditionally used for cognitive bias modification. The example that participants viewed was based on the previous study by Zhang et al [9]. In that study, the specific nature of cognitive bias modification was that of attention bias modification, and it was applied to individuals with addictive disorders. Then, participants were shown how the traditional visual probe task paradigm is modified to become an emotional bias task paradigm (by modifying the visual cues presented). Then, they were asked to identify limitations and suggest potential methods to overcome some of the identified limitations regarding the emotional bias modification intervention. Subsequently, participants were shown examples of existing gaming interventions in published literature and commercial stores. Figure 1 provides an overview of some of the games that were shown to the participants.

They were asked to comment on the advantages and limitations of these interventions. Finally, participants were shown a list of gamification techniques, and each of the techniques was explained to them. The list of gamification techniques was based on the previous study by Hoffman et al [14]. The list of gamification techniques is described in Table 1, and this has also been published in the previous study by Zhang et al [9]. They were asked if the inclusion of such techniques would be appropriate.

‎View this figure

Data Analyses

The descriptive data (demographics) were summarized as means and SDs. The audio recording obtained from the workshop was transcribed verbatim. One of the principal investigators, MZ, performed the first transcription and developed the coding frame. To ensure the reliability of the coding frame that was adopted, two authors (MZ and RV) reviewed the transcripts and discussed the coding frame. This ensured that the process of intercoder consensus was adhered to. Codes that were identified were classified into categories and then reorganized into themes. The themes that were generated were subsequently reviewed and further refined. The underlying methodology used was in accordance with the previous recommendations by Braun and Clarke [15]. NVivo (version 12.0; QSR International) [16] was used for thematic analysis.

Ethics Approval

This study was approved by the ethical review board of the Nanyang Technological University Singapore (approval number IRB-2020-03-058). Informed consent was obtained from all the participants, and they were also informed that they could withdraw from the study at any time.

Data Management

No participant-related identifiers were captured on the hard-copy questionnaires. All the completed hard-copy forms and informed consent forms were stored in a secured facility. The audio recordings of the workshop were transferred to a local secured computer for storage, and the original recording was deleted from the recording device. The password of the local computer was changed frequently, and only the principal investigators were able to access the files. All the records and audio recordings will be maintained for a period of 6 years following the completion of the study. All participants were provided an inconvenience fee for their time and effort in participating in the study.

Results

Demographics

A total of 4 health care professionals consented and participated in the workshop, which was conducted on May 19, 2021. Of the 4 health care professionals, 1 (25%) was a child and adolescent psychiatrist and the remaining 3 (75%) were psychologists. The mean age was 43.5 (SD 5.74) years. Of the 4 participants, 1 (25%) was a man and the remaining 3 (75%) were women. The mean years of experience was 14.5 (SD 4.43) years, ranging from 10 to 20 years. Table 2 provides an overview of the baseline demographics of the participants.

Phase 1

In phase 1, the participants identified several limitations regarding the nature of the conventional visual probe task used to assess emotional biases. These limitations were related to the nature of the task parameters, included stimulus set, and other factors that could affect the accuracy of responding to the task. Participants also made recommendations regarding how the task could be improved. Textbox 1 provides a summary of the verbatim comments of the participants for each of the identified themes.

Phase 2

In this phase, participants were asked to provide their perspectives about the existing ADHD games. Several themes arose from the discussion. Participants highlighted that the games needed to be age-appropriate and clearly explain the specific skill sets that were targeted. Participants also raised concerns regarding the scientific evidence base for some of the presented games. Participants also highlighted issues with existing games regarding their novelty and stated that the consideration of novelty was important when implementing games for individuals with ADHD.

Textbox 2 provides a summary of the verbatim comments of the participants for each of the identified themes.

Phase 3

In this phase, the facilitator explained to the participants more about the common gamification strategies that have been used. Then, participants were asked to select the most appropriate gamification strategies that could be applied to the conventional emotional bias modification task. Strategies such as economic and performance-based gamification were deemed to be more appropriate.

Textbox 3 provides a summary of the gamification techniques that the participants have selected and their justifications for their selection.

Discussion

Principal Findings

This is the first study that has explored the use of participatory action research as a research design to refine the conventional emotional bias modification task. This study complements the ongoing research that has been highlighted in the Introduction section, which demonstrated the presence of emotional biases among children and adolescents with ADHD. In our study involving health care professionals, several limitations were identified regarding the conventional emotional bias intervention. These included the nature of the task parameters, included stimulus set, and factors that could have an impact on the accuracy of responding to the task. After examining the existing ADHD games, participants raised concerns about the evidence base of some of the apps. They articulated that any developed ADHD game ought to clearly identify the specific skill set that was targeted. When health care professionals were eventually asked to select gamification strategies that could improve the conventional intervention, they preferred economic and performance-based gamification approaches.

Some of the identified limitations of the existing conventional visual probe paradigm used for the assessment of emotional biases are congruent with the findings of previous studies. In our study, health care professionals identified limitations pertaining to the task parameters, particularly that which pertains to the speed of presentation of the stimulus set. This is congruent with the previous co-design workshop, by Zhang et al [9], involving health care professionals and patient participants, who similarly reported that the presentation of the stimulus was very rapid. In that study, the participants recommended a lengthy stimulus presentation time to allow them to process the set of stimulus images [9]. On the basis of the review by Zhang et al [9] about the paradigms for the visual probe task as applied for addictive disorders, it would be of important to have stimulus presented for both short and long stimulus intervals. This allows for assessing both attentional processes, namely, initial orientation and delayed disengagement. This should be considered when developing the task paradigm for assessing emotional biases among adolescents. Another limitation that was identified pertains to the nature of the images presented. Health care professionals recommended for standardization in the size of the images, whether images are presented with or without borders, and how much these presented images contrasted against the background. These identified limitations are crucial and need to be considered when designing the next iteration of apps that assess emotional biases.

One of the concerns raised by health care professionals about existing ADHD games was whether these games were based on validated frameworks and adhered to the evidence base. In our workshop, we presented participants with examples of ADHD games that are commercially available, some of which have been previously evaluated. In a recent review by Penuelas-Calvo et al [17], they examined video games for the assessment and intervention of individuals with ADHD. A total of 22 papers were identified, and they reported that the existing tools were influential in establishing whether individuals do have attentional issues as compared with the control. Unfortunately, the review was published after we conducted our focus group. Otherwise, the data obtained from the review would have helped to advance the discussion. Another concern highlighted during our focus group was the need for games to be specific in terms of identifying the skill sets they sought to target. In the examples we shared, it appeared that some of the commercially available apps were specific in terms of what they wanted to develop. Penuelas-Calvo et al [17], in their examination of the literature, reported that most of their identified video games have been developed based on previously validated tasks, for example, the performance task by Corner and the go–no-go task. Regarding video games that were interventional in nature, most of them focused on cognitive training, such as improving executive functioning, attention or working memory, reaction time, cognitive flexibility, or motor ability. The responses obtained from our focus group and the findings by Penuelas-Calvo et al [17] further highlight the importance of computerized interventions on previously validated evidence-based tasks. It is also important to be mindful that when adapting the task to a computerized intervention or even a video game, the mechanism of the conventional task should not be altered.

Our participants also reported various gamification strategies that may benefit our emotional bias assessment task. Previous study has justified the importance of considering gamification strategies, given that they help with the improvement of engagement rates and reduction of dropouts [6]. The perceptions of our team of participants in the co-design workshop indicated that repetitive activities in the game with increased difficulty levels would potentially affect the performance, continuity, interest, and novelty of the game. Novelty, user-friendliness, and interactive game design concepts should be considered, so that the ADHD games can be made engaging effectively. One of the critical perspectives shared during the workshop was the need for clear instructions and directions in the gameplay. Otherwise, it is likely that children with ADHD may be frustrated with the game. Regarding the gamification strategies suggested, leader boards, digital rewards, and real-world prizes could engage the users and allow them to play the intervention multiple times. Our participants also advocated the consideration of economic approaches, given that they offer individuals a tangible, immediate reward. Despite these suggestions, we need to acknowledge that one of the major limitations of our study was that we have not considered the perspectives of children themselves and those of their caregivers. It is important to understand the perspectives of children, so that the intervention could be personalized to their needs. Although we agree with the previous recommendations by Penuelas-Calvo et al [17] that health care professionals should collaborate with computer engineers, we feel that apart from collaborating with an engineer to ensure a high-quality app that resembles the quality of commercial apps, it is far more critical to ensure that the app is personalized to the needs of the patients.

The main strength of this study was the use of co-design methods in the refinement of an existing evidence-based paradigm for the assessment of emotional biases. This helped to ensure that the eventual design is based on evidence but meets the potential needs of end users. This study also presented examples of existing ADHD apps to participants. In doing so, we were able to allow the participants to have a better in-depth understanding of existing apps and identify issues and limitations with existing apps. Despite these strengths, our study had several limitations. Although we initially planned to recruit a total of 8 participants, we managed to eventually recruit only 4 (50%) participants. The COVID-19 pandemic affected our ability to recruit participants, as they may have to attend to other clinical needs. In addition, although we had previously planned for a physical workshop, the COVID-19 pandemic prevented the execution of such a workshop owing to infection risks. Thus, we were limited to conducting a web-based workshop, which may have resulted in challenges among participants in responding. It would be ideal to also obtain insights from patient participants or their caregivers.

Conclusions

This study has managed to elucidate health care professionals’ perspectives on the refinement of a conventional emotional bias intervention for individuals with ADHD. Taken together, gamification strategies could be applied to conventional emotional bias interventions. The findings from this study have implications on the subsequent studies seeking the personalization and gamification of such apps. Although we have sampled the perspectives of health care providers, it remains necessary to discuss these perspectives with the intended sample group as the next step, to ensure that the future conceptualized app will be consistent with their needs.