Integrating Behavioral Science and Design Thinking to Develop Mobile Health Interventions: Systematic Scoping Review

Background Mobile health (mHealth) interventions are increasingly being designed to facilitate health-related behavior change. Integrating insights from behavioral science and design science can help support the development of more effective mHealth interventions. Behavioral Design (BD) and Design Thinking (DT) have emerged as best practice approaches in their respective fields. Until now, little work has been done to examine how BD and DT can be integrated throughout the mHealth design process. Objective The aim of this scoping review was to map the evidence on how insights from BD and DT can be integrated to guide the design of mHealth interventions. The following questions were addressed: (1) what are the main characteristics of studies that integrate BD and DT during the mHealth design process? (2) what theories, models, and frameworks do design teams use during the mHealth design process? (3) what methods do design teams use to integrate BD and DT during the mHealth design process? and (4) what are key design challenges, implementation considerations, and future directions for integrating BD and DT during mHealth design? Methods This review followed the Joanna Briggs Institute reviewer manual and PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) checklist. Studies were identified from MEDLINE, PsycINFO, Embase, CINAHL, and JMIR by using search terms related to mHealth, BD, and DT. Included studies had to clearly describe their mHealth design process and how behavior change theories, models, frameworks, or techniques were incorporated. Two independent reviewers screened the studies for inclusion and completed the data extraction. A descriptive analysis was conducted. Results A total of 75 papers met the inclusion criteria. All studies were published between 2012 and 2021. Studies integrated BD and DT in notable ways, which can be referred to as “Behavioral Design Thinking.” Five steps were followed in Behavioral Design Thinking: (1) empathize with users and their behavior change needs, (2) define user and behavior change requirements, (3) ideate user-centered features and behavior change content, (4) prototype a user-centered solution that supports behavior change, and (5) test the solution against users’ needs and for its behavior change potential. The key challenges experienced during mHealth design included meaningfully engaging patient and public partners in the design process, translating evidence-based behavior change techniques into actual mHealth features, and planning for how to integrate the mHealth intervention into existing clinical systems. Conclusions Best practices from BD and DT can be integrated throughout the mHealth design process to ensure that mHealth interventions are purposefully developed to effectively engage users. Although this scoping review clarified how insights from BD and DT can be integrated during mHealth design, future research is needed to identify the most effective design approaches.


Behavioral Design Thinking
Integrating methods from Behavioral Design (BD) and Design Thinking (DT) to develop mHealth interventions that more effectively engage users Empathize with users (DT) and their behavior change needs (BD) Define user requirements (DT) and behavior change requirements (BD) Ideate user centred features (DT) and behavior change content (BD) Prototype a solution that is usercentred (DT) and operationalizes behavior change insights (BD) Test the solution against user needs (DT) and for its behavior change potential (BD) The Behavioral Design Thinking Approach is based on a review of 75 studies that combined insights from Behavioral Design (BD) and Design Thinking (DT) throughout their mHealth design process . The approaches and methods studies used to integrate BD and DT at each stage of the design process are described, along with examples.* *The five stages outlined below are not meant to be mutually exclusive or followed in a linear fashion. The five stages have been used to organize and make sense of the vast range of methods and approaches available to design teams. The methods and approaches listed are also not meant to be exhaustive of what can be utilized during Behavioral Design Thinking.

"User Requirements"
Define the user and behavior change requirements

"Behavioral Requirements"
Identify requirements within the project scope and objectives Use feasibility criteria to refine and select certain requirements Consult experts and patient/ public users to prioritize the requirements Rank requirements according to their (a) likelihood elicit behavior change, (b) alignment with current practices and guidelines, (c) adaptability to a digital interface, (d) acceptability to users, and (e) compatibility with data collection needs.

Organization of Requirements Organization of Requirements mHealth Specific Considerations
Organize requirements by shared user beliefs & needs Create user stories, journeys and scenarios Translate user requirements into "knowledge seeking questions" for future ideation Organize requirements by behavioral determinants that need to be addressed (e.g., motivation) Select a behavioral model to guide future design Translate requirements into a logic model that clarifies relevant behavioral outcomes, objectives and determinants Organize requirements by their potential location in the mHealth application Clarify whether requirements will be structure or feature based (structures are the technical elements to deliver the features) Define user centred requirements by analyzing and sorting user experiences and preferences into key themes and sub-themes.
Make a "behavioral diagnosis" by analyzing empathy results, referring to constructs of a behavioral theory or model (e.g., COM-B).

Define the user and behavior change requirements
Examples from included studies [65,39,34,58]  Stage 2 of Ross et al's design process involved clearly defining the technical features and structure of the smartphone program. The selection of the smartphone program's technical features and structure was informed by findings from the formative study and a literature review (conducted by the research team) . The research team made sure to clearly define the technical characteristics of the intervention (eg, the platform used to deliver the strategies, personalization of the strategies) and the frequency and conditions of intervention delivery (eg, frequency of medication reminders) at this stage. Technology partners can help iteratively ideate and specify how behavioral and user requirements can be feasibility operationalized within an mHealth app (including design, functions, features and flows). Tools such as the BIT model, PSD framework and/or Nielsen's Heuristics can be used to guide the translation of behavioral and user requirements into technical elements for the program.

Technology-Partner Guidance "Behavioral Mapping"
A structured, evidence-based approach to identify behavior change content Researchers can refer to a "guiding principles" document to reflect on how user requirements can be made compatible with the target behavior, theory, and a digital format. Researchers can use their expertise to reflect on how behavior change insights may be integrated with existing programs, practices and norms, as well as with best evidence from mHealth design and clinical innovation.

Research-Led Ideation
Group discussion about most important behavior change insights and strategies (including application content dose, delivery, and organization) suitable for the target population, a mobile app and behavior change objectives. Brainstorm personalization and tailoring strategies to deliver specific behavior change, user-centred and clinically relevant content.

Multidisciplinary Stakeholder Workshops
Co-designing with patients/ the public can help ensure behavior change insights are user friendly. Getting patients/ the public to rank, prioritize and conceptualize how behavioral insights should be operationalized to meet their own needs. Utilize visual and creative methods to help users reflect on app ideas, content, and design. ·Utilize "user personas" to guide ideation, constantly checking if behavior change insights are accounted for (and to discuss their potential integration if not addressed) Utilize "solution inspiration" and "whiteboarding" to structure how the app could be built.

Patient and Public Engagement
Summary of approaches and methods from 75 studies  Methods to "Ideate"

user-centred features and behavior change content
Examples from included studies [47,27,49]  First, behavior change techniques (BCTs) suitable for targeting the selected determinants were identified using the Behavior Change Wheel and BCT Taxonomy. The most appropriate BCTs for each determinant were chosen, excluding those that were either unfeasible given the digital mode of delivery or irrelevant to the target behavior. Second, based on the system requirements and list of BCTs, ideas for the look, feel, and functionality of the app were developed by the steering group (health psychologists, a public health consultant, a project manager, two young people, and a design lead) and the Young People's Partnership Board (YPPB) (group of end users). Over a period of 3 months, the steering group, the YPPB, and the design agency worked together to develop practical strategies (what users of the intervention would see and do) to deliver BCTs. This was a creative and iterative process. All ideas were judged according to their feasibility and ability to effectively deliver the BCT. The ability of the proposed ideas to engage the priority population was discussed among the YPPB. The steering group and YPPB made the final decision about which practical strategies should be used based on the above criteria. After potential behavioral strategies were identified (from a literature review of behavior change strategies), a multidisciplinary team engaged in a series of group brainstorming sessions, also known as ideation, to conceive of creative ways in which these techniques might be implemented. Brainstorming centered around potential mobile app features and was inspired by a collection of highly rated apps on the App Store. A multidisciplinary team enabled the ideation of a wide range of diverse ideas taking behavioral theory, user insights, and product experience into account.

Example 2: The development of Drink Less: an alcohol reduction smartphone app for excessive drinkers by Garnett et al., 2019 [27]
Garnett et al. translated specified intervention requirements into app modules. First, BCTs targeting salient behavior constructs were selected by the research team using a mapping approach guided by the BCT Taxonomy Version 1. Second, the research team used PowerPoint slides to describe how each intervention component could be translated into an app module (in terms of the text, graphics, and functionality) to the app developers. The intervention components were translated into potential app modules in close collaboration with expert and experienced app developers. Regular discussions between members of the research team and the developers were conducted to ensure that the description provided of the intervention components was translated in ways that were feasible for users (judged by the app developers) and met the initial specification (judged by the research team).

Behavioral Design Considerations
Co-Development of BCTs: Users and key stakeholders can help clarify what (content), when (frequency), how many (dose), who (target user) and which (medium type) BCTs should be delivered. Co-development can also help create more appropriate questions to personalize application content (to deliver unique behavioral, clinical and user-centred content).

Develop an Intervention Plan:
Developing an "Intervention Plan" can help clarify the reasons behind the ideated content (i.e., behavioral and usercentred linkages) which may help ensure BCTs and user preferences are not lost in translation during technological development.
Prototype a solution that is user-centred and supports health behavior change

Design Thinking Approach
Iterative prototyping, feedback and refinement (e.g., using sprint, scrum and agile methods)

Key Prototyping Tools Design Thinking Tools
Wire framing, paper prototyping, rapid mockups, system architecture diagrams, app flow charts, user stories, and use case scenarios.

Behavioral Design Considerations
BCT Flow Charts: Map out how the BCTs will be interacted within the proposed prototype.

User Stories organized by Target Behavior Changes:
Organize user stories by the target behavior changes to be achieved to ensure the solution meets relevant behavioral requirements.

Design Thinking Aids
Usability Heuristics, Eight Golden Rules, Human Interface Guidelines

Behavioral Design Considerations
Use the PSD Framework, BIT Model or Computer Tailoring Method: Using a behaviorally-informed technology development framework or model can help guide the translation of behavioral insights into operationalized features and functions within a prototype.

B D D T Prototype a solution that is user-centred and supports behavior change
Examples from included studies [69,16,64,68]

Example 2: Targeting Parents for Childhood Weight Management: Development of a Theory-Driven and User-Centered Healthy Eating App by Curtis et al., 2015 [16]
After deciding what BCTs to include, Curtis et al aimed to embed these BCTs into an actual app informed by user preferences. Consulting with the app company facilitated the process of how BCTs identified in previous steps could be meaningfully combined with findings on user preferences to create app features. Consultations with a software engineer led to the development of a "Flow Chart" to map specific BCTs to app features, showing the sequence of how BCTs would be delivered to parents. The diagram was further refined through parental feedback. With respect to enhancing user experience and engagement in the application, parents' preferences for app features were combined with suggestions from the app company. For example, progress bars, achievement badges, and points were identified as a way of providing feedback for parents on their children's eating behavior.

Example 3: The Design and Development of a Personalized Leisure Time Physical Activity Application Based on Behavior Change Theories, End-User Perceptions, and Principles From Empirical Data Mining by Sporrel et al., 2021 [64]
Sporrel et al. outlined that while some frameworks provide a step-wise instruction to develop an mHealth intervention (e.g., IDEAS framework) others provide guidance for specific steps such as the translation of behavioral theories to persuasive strategies delivered within an app (e.g., BIT model). Sporrel et al. used the BIT model to guide the practical implementation and design of the persuasive strategies in the app. Specifically, the BIT model helped to describe what intervention modules the user receives (e.g., notifications), how it relates to the persuasive strategies (e.g., reminders), when they have access to the module (e.g., under which condition), and how the modules look like (e.g., complexity and aesthetics). The team outlined these conditions by creating "User Stories" with stakeholders, behavioral scientists, computer scientists, and the app developer. The "user stories" served to translate the requirements into the app modules, as they explain the technical implementation and design characteristics of the selected persuasive strategy (e.g., "As a user, I want to restrict the amount of messages I receive during the week, so I do not get disturbed too frequently and get annoyed by the intervention.")

Example 1: Using codesign to develop a culturally tailored, behavior change mHealth intervention for indigenous and other priority communities: A case study in New Zealand by Verbiest et al., 2018 [69]
Prototyping was an iterative process for Verbiest et al., with regular feedback and consensus meetings. The app developer took the high level of input from the community focus groups and worked with the multidisciplinary research team to create several wireframe prototypes of the mHealth tool. Incorporating BCTs into mHealth intervention was an iterative process founded on community based co-design. Through the co-design and early prototyping, it became evident that despite community partners having similar ideas with regard to BCTs, there were clear differences with regard to the content of them and how they would be delivered. Communities provided feedback on the wireframes and broad consensus was reached regarding content, features, and functionalities (i.e., not everything could be adopted due to reasons such as time and technology constraints). van Agteren et al argued that although the design process steps 1-3 resulted in the selection of evidence-based BCTs and practical applications to accomplish behavioral change, they did not provide sufficient guidance on the design of mHealth interventions. Using the Persuasive System Design (PSD) framework helped can Agteren guide the overall functional development of Kick.it. The PSD framework provides intervention designers with 28 design principles falling into four categories that aim to ensure that the final system is actually capable of supporting sustainable health behavior change.

B D D T
Test the solution against user needs and for its behavior change potential Examples from included studies [45,42,27,44,26]  Mathenjwa et al completed in-depth user interviews with participants who received the intervention. The interview questions explored why participants had not been linked to care after the HIV diagnosis and their views of the intervention and the impact it had on them. Interviews were thematically analyzed with themes guided by Self-Determination Theory (SDT), to understand the intervention's potential impact on autonomy, competence, and relatedness.

Example 4: A Personalized Physical Activity Coaching App for Breast Cancer Survivors: Design Process and Early Expert
Evaluationby Moneiro-Guerra et al., 2020 [44] An expert evaluation was performed with the mobile app rating scale (MARS) and the app behavior change scale (ABACUS) to assess the quality of the app and the potential for behavior change, respectively. The MARS examined app elements such as engagement, functionality, utility, aesthetics, and information. An overall functionality score out of 5 was derived using this scale. The ABACUS scale comprises 21 items and was used to examine the potential behavior change of the app in relation to goal setting, action planning, barrier identification, self-monitoring, and feedback. The final version of the intervention was formally mapped to Michie and colleagues latest BCT taxonomy in order to describe which BCTs were present. Every page was reviewed (and where a BCT was explicitly meant to be assigned). Morrison et al did this evaluation to provide a reliable record of the content of the behavior change intervention, and to confirm that they included the BCTs as planned.

Example 5: Design of a Mobile App for Nutrition Education (TreC-LifeStyle) and Formative Evaluation With Families of Overweight Children by Gabrielli et al., 2017 [26]
Gabrielli at al conducted a pilot test of their intervention, and assessed: (1) parents' knowledge of the Mediterranean Diet guidelines; (2) state of change for healthy nutrition-measured by an adapted version of the URICA-short-form scale, based on the 4 readiness to change states of the Transtheoretical Model of Change, and (3) intention to use technology for nutrition education. After 3 weeks of the app usage, any changes on the above factors were recorded and parents' perceived usability of the app was assessed (System Usability Scale Questionnaire). After 6 weeks of app usage, any changes in the factors above were recorded again as well as a more in depth review of the participants' experience with the mHealth intervention in a semi-structured interview.

Example 3: The development of Drink Less: an alcohol reduction smartphone app for excessive drinkers by Garnett et al., 2019 [27]
Garnett et al. conducted a usability testing study to explore user views toward the app in order to determine whether the BCTs were acceptable and feasible to users and how they might be improved. The usability study consisted of two parts: a think aloud study to understand users' first impressions and semistructured interviews to investigate user's impressions of using the intervention in naturalistic settings.