Designing for inclusion: applying behavioural frameworks in digital health

08 Apr 2026 12min read

Behavioural science methods are growing in popularity in health technology. Innovators increasingly acknowledge the importance of understanding the nuances of patient and healthcare professional (HCP) behaviours to facilitate adoption of supportive technology. It is especially relevant in digital health, where complex combinations of features must support diverse experiences without creating additional burden or anxiety. Considering how individuals learn, interact with technology and make decisions about their health is essential to implement tools that people can use consistently.

However, there are challenges associated with translating behavioural science into tangible design decisions and there is limited consensus on how to achieve this meaningfully in a non-academic context. So, how can a structured behaviour design framework be embedded in the design of digital health products to create engaging and equitable solutions?

Why behaviour design matters in digital health

Behaviour design involves the development of products that are informed by behavioural science principles, such as cognitive biases and mental heuristics and designed around interventions such as “nudges” or Behaviour Change Techniques (BCTs).

Digital health includes technologies that enhance healthcare pathways, from diagnostic apps and digital therapeutics to wearables and remote monitoring tools. These solutions aim to facilitate data collection and interpretation for healthcare decision-making. These interventions must strike a balance between improving existing practices and integrating within existing workflows. While digital interventions play a significant role in modern care pathways, their potential to improve outcomes is only realised when used regularly. The onus of responsibility is largely on product developers – solutions must be designed to fit seamlessly not only into healthcare pathways, but also everyday lives, inviting engagement without creating burden. However, if built on incomplete assumptions about users, they can struggle to sustain engagement.

This challenge of adherence can be tackled with behavioural science. There must be a strong understanding of the care pathway, but the analysis should go beyond workflow bottlenecks and address the human factors more broadly. A wealth of research in adherence and habit formation can be tailored to individual use cases. Behaviour design investigates specific barriers to adherence and systematically explores ways of addressing them.

Digital health solutions also have the potential to reduce healthcare inequalities by improving access to care, sharing knowledge and facilitating personalised treatment. However, they can also perpetuate existing gaps by overlooking the specific needs of different groups. The gender data gap, for example, is the result of systemic underrepresentation of data about women and their experiences. Applying biased data to product development and not specifically addressing women users’ needs, risks creating products that impose barriers to use or that are less impactful for women. Behaviour design applies a holistic view of all user experiences to address differences and barriers such as systemic biases and lack of access. These methods can be used to analyse underlying attitudes and systems behind biased outcomes and design for change.

Challenges of applying behaviour design

The practical application of behaviour design in digital health development is not without challenges:

Contextual complexity – Care pathways differ widely across conditions, regions and healthcare systems. An intervention that is relevant for one user group may be less impactful or feasible for another.
Variable approaches – Design teams may rely on anecdotal behavioural science insights rather than robust evidence from large-scale studies and systematic reviews. This can lead to inconsistent application of BCTs across solutions.
Resource constraints – Product teams often face tight timelines and budgets, making it challenging to conduct the research required to meaningfully embed behavioural science in development, especially for evaluating long-term engagement and efficacy in a randomised controlled trial (RCT).
Overgeneralisation risk – Without a conscientious approach to identifying individual differences, solutions may default to a ‘one-size- fits-all’ approach, overlooking the needs of underrepresented groups. These challenges highlight the need for a systematic approach that combines behavioural science theory with user-centred development, ensuring that solutions are grounded in evidence while respecting individual differences and distinct care pathways.

Applying a structured behaviour design framework

Knowledge of cognitive biases and behavioural economic principles is essential to guide and challenge design work iteratively, however further structure is needed to prevent teams from defaulting to familiar solutions that may be overly generalised. The behaviour design framework we highlight integrates behavioural science with UX design, research and agile testing methods (Figure 1).

It is made up of three core phases:

Foundations
Behavioural Science
Iterative Design.

Together, these represent a systematic series of activities to facilitate the balanced and tailored application of behavioural science.

Phase 1 - Foundations

The foundation of behaviour design is deep contextual understanding of the care pathway as well as the people experiencing it. This begins with exploratory research activities that illuminate care journeys, decision points and the driving factors and barriers influencing these key milestones.

At this stage, researchers should investigate and map the workflows where the solution will be implemented. The specific research questions will vary, but in general, the following should be addressed:

What are the current standards of care?
When and where are the touch-points between patients and HCPs?
What are patients and HCPs doing in between these touch points?
What kind of data is collected and utilised? What level of granularity is needed? What is considered clinically significant?
What tools are implemented?
What errors or misuses are observed with existing tools?
What costs and resources are associated with the pathway?
What differences can be observed between different segments of the target population (e.g.symptoms, ages, sexes, genders, socio-economic statuses, races, ethnicities and geographies)?
What pain-points do patients and HCPs experience?

Depending on the use case, different approaches can be pursued to collect this information. While regional standards provide a useful reference (e.g. NICE guidelines), primary insights from representative stakeholders are essential. Ideally, this should include a combination of in-depth interviews, surveys and ethnographic research for a strong foundation of data.

Interviews provide deeper, qualitative insights on the challenges and individual experiences in a workflow, tackling the challenges associated with designing for complex care pathways. Survey questions facilitate larger sample sizes, comparable responses and quantitative data to complement qualitative insights without overgeneralising and ethnographic research allows for the observation of details that may not be explicitly volunteered by participants.

There are likely assumptions that are taken for granted by people who are very familiar with these workflows. A neutral observer can help challenge these and identify opportunities for innovation.

When the relevant workflows are established, the next step is to collect more detailed insights about the users. To ensure a comprehensive approach, the following factors should be specifically investigated upfront. These insights form the basis for applying the COM-B model, developing a holistic perspective on the factors driving behaviours in the care pathway:

Capabilities: What capabilities are needed in this pathway? Are there physical constraints, cognitive barriers or knowledge gaps that could influence decisions?
Opportunities: What environments, tools and resources are needed? How accessible are they? What stigmas and cultural norms are present?
Motivations: What beliefs, habits and emotions influence how people behave?

Collecting these insights through user research requires a journey-driven approach, whereby participants talk through or demonstrate their experiences end-to-end, allowing researchers to make connections between events, circumstances and perceptions. In research interviews, asking directly what motivates someone or what they would find helpful yields reflective responses, meaning it may not reveal automatic habitual influences or core needs. Solution-agnostic and open questioning is more likely to uncover deeper insight, especially when combined with other methodologies (e.g. ethnographic research) where investigators can observe the behaviours that are not overtly acknowledged.

Phase 2 – Behavioural science

Identifying and understanding behaviours

Once the care pathway is mapped, key behaviours can be identified based on data representing the entire journey. At this stage, the focus should be on behaviours that are most relevant to the intended outcomes of the digital solution or target behaviours. For example, if a remote monitoring solution is designed to reduce the number of in-person healthcare visits needed, the target behaviours are those surrounding the decision to set up in-person appointments, as well as taking action to monitor oneself at home. These can be translated into specific, measurable behaviours to encourage with the intervention. The approach should be specific to the care pathway and the intended healthcare or system outcomes.

The target behaviours can be analysed based on findings from the exploratory research. The COM-B model of behaviour from UCL is one of several models that can be applied at this stage. In this framework, the COM-B model works well by facilitating a systematic approach to not only the analysis of target behaviours, but also the identification of evidence-based interventions. It encourages distinct analyses of individual behaviours, which fosters more tailored solutions. With this model, each behaviour is characterised by its influencing barriers and drivers, which can include those related to capability, opportunity and motivation. Identifying the relevant factors across these three categories for each behaviour ensures that designers attend to holistic contexts of use and acknowledge how these may differ across user groups.

Identifying interventions

The COM-B model is part of a broader approach to behaviour change intervention design, the Behaviour Change Wheel (BCW), which includes a taxonomy of behaviour change techniques and a guide on where they have shown success in other research. BCTs such as goal-setting, framing and biofeedback have been studied in various use cases and have tackled specific barriers defined in the COM-B model. With this theoretical framework and the review of previous research, design teams can identify the techniques that have shown promise in analogous situations (e.g. similar patient populations, related behavioural barriers). The techniques can serve as helpful prompts to de-risk the design process by ensuring evidence is guiding the selection of key features and architectures.

Translating and designing around interventions

The next step in the framework is conceptualising how selected BCTs translate into tangible design principles, UX strategies and product features. For example, the BCT of biofeedback might be recommended as a way of targeting a lack of awareness about one’s own health that prevents them from seeking the support needed. This technique can then be explored collaboratively within the broader context of technical constraints, to design a solution involving sharing wearable device data back to the user at regular intervals. Using techniques as guides for feature design in this way can help reduce over-reliance on a team’s previous experiences.

Phase 3 - Iterative design

This process is grounded in theory and evidence, but should also be sensitive to the lived experiences of specific patient groups. It is critical to test and iterate behavioural interventions in context. Doing so through ongoing user research ensures that solutions resonate in practical settings with representative users.

To collect actionable feedback to inform iterations, testing can incorporate qualitative feedback and task performance data to identify potential usability challenges as well as user perspectives. Rapid prototyping techniques, such as code-free mock-ups of digital interfaces, can ensure that research is done repeatedly at low cost, before investing in larger RCTs with a more mature solution. For example, data presentation styles, use of language, proposed frequency of use, number of use steps, etc. can all be investigated in early-stage user research. Evidence from agile testing combined with insights from theoretical models can help build confidence in the solution.

Conclusion

Applied behavioural science has significant potential to improve the effectiveness and inclusivity of digital health solutions. By embedding a systematic behaviour design framework into product development, teams can optimise patient engagement, de-risk development and tackle disparities.