Participants in an EMD intervention pilot study described the importance of past experiences to current health beliefs and a sense of responsibility to achieve and maintain symptom control. Participants described increased awareness of their condition, in part attributed to EMD data. They were generally open to integrating EMD data with environmental, physiological and activity data, and saw ways in which such technologies could enhance asthma care.

Similar to findings in this study, previous investigators have noted heterogeneous responses to EMD use. Some adolescents have described feeling an increased sense of control and responsibility for their condition34. Others saw the monitoring as a sign their clinicians did not trust them35. Also similar are previous studies where adults and adolescents described behaviour and/or attitude change from the using EMDs (including “habit formation”), with some adults linking this to better asthma control34,36. Perceptions of how durable these changes would be were mixed35,36. These changes appeared to be somewhat linked to baseline attitudes and beliefs, and pre-existing adherence to or dislike of routines36.

Other studies have also described data facilitating conversations with healthcare teams34,36, a desire for data access to be limited34 and concerns that data would replace them being seen as a person35. Although reminders were not formally used in this study, some participants reported finding and using this function. The literature suggests variable acceptance of reminder functions with potential implications for their use34,35,36.

The association between a self-regulatory perspective (including its specific application, the necessity-concerns framework) and adherence in asthma has already been outlined. Participants discussed beliefs about asthma and its treatment, influenced by their varying experiences, that appeared to motivate or demotivate inhaler use in keeping with what is already known15,36,37. As has also been previously noted, there were elements in addition to the necessity-concerns framework which appeared to play a role16.

Participants in this study described being motivated to avoid recurrence of the kind of deterioration that had led to their recent exacerbations, highlighting a target population that could benefit from intervention. In a Protection-Motivation model of behaviour change, a ‘threat appraisal’ of susceptibility, severity, and fear are central to motivating intention to take on adaptive behaviours to address the threat38.

Importantly, Protection Motivation Theory notes a risk of maladaptive responses, including avoidance, denial and hopelessness, where an accurate threat appraisal has been made but there is low belief in treatment efficacy, high concern about treatment costs and low self-efficacy38. This may in part explain why some intervention participants demonstrated persistently poor adherence and why other narratives around poor ICS efficacy and low self-efficacy appeared to blunt intentions to adhere to treatment.

Previous work has similarly identified a subgroup of individuals with poor adherence which is resistant to intervention39,40,41. EMD-based interventions may aid identification of this group, facilitating sensitive exploration of underlying beliefs and adaptation of interventions. This could prove key for some in tipping the balance in favour of adherence-concordant beliefs and adaptive behaviours38,42,43.

Participants also highlighted the importance of cues which were visual, auditory and events-based, a finding also seen by Foster et al. in their study36. Participants perceived that factors which disrupt routines (e.g. shift work) negatively impacted their adherence. Habit theory proposes that, whilst beliefs inform initial motivation to begin a new behaviour, habit formation embeds behaviour change by rendering a new behaviour automatic. This allows new behaviours to be maintained long after both motivation and awareness have ended. For such automaticity to take place, a new behaviour must be learned in an enabling environment, a critical cue for action identified and a plan put in place to perform the desired action when cued44.

Overall, data from this study suggest that, in a selected post-exacerbation population, EMD interventions comprising clinician input have potential to influence beliefs and increase motivation for ICS adherence. They may also assist in identifying individuals who need more complex engagement around treatment efficacy and concerns, and self-efficacy. Finally, they may provide an opportunity to target habit formation as a means of embedding behaviour change.

EMD research in asthma has primarily focused on its potential to reduce adverse risk through improved adherence. Participants in this study however were curious about the potential for an integrated technology platform to inform lifestyle choices such as exercise and trigger avoidance. They wanted to integrate data with environmental data, physiological and activity markers, and validated symptoms to self-monitor and provide better information for shared decision-making with their clinical teams, a finding supported by previous work19.

Self-management is central to chronic disease care where clinician input is limited by time. The US Institute of Medicine suggests it comprises “confidence to deal with medical management, role management and emotional management” of a condition45. In asthma, supported self-management has been shown to be effective in improving outcomes46,47. Evidence suggests that self-management behaviours are most effectively influenced when clinicians take the time to engage with individuals37.

By providing clinicians with the same tailored information available to users, integrated with markers of modifiable factors and outcomes that matter to users, EMD-based interventions may provide a common language to increase engagement of users with their self-management and of clinicians with their patients. Personalisation has the potential to allow for tailored self-management interventions, including personalised asthma action plans informed by data, automated advice and access to mental health support using validated apps.

Ethical data use is key, with an expectation of data security, transparency over what data are being used for and of some level of control over data access. Without this, there is a risk of trust breaking down. This is particularly key when considering the implications of platform technologies that allow for the integration of commercial sensor data with potentially sensitive health data, and where development of automated interpretation is likely to involve algorithms which require training using existing data. Recent controversies highlight the importance of transparency48,49 in such circumstances.

Participants in this study were largely keen for EMD-based interventions to be delivered in primary care, enhancing rather than replacing their routine reviews. Given that EMDs’ key role is likely to be in supported self-management, this appears a natural choice. However, primary care services in many healthcare systems are already under pressure, meaning that careful thought is needed for implementation. This study suggests that, without training or allocated interpretation time, clinicians generally did not find EMD data helpful in informing management.

For monitoring interventions to be successful, individuals need assistance in processing and effectively utilising the data gained50. This is something that participants from this study actively expected, but will require clinician training and time for them to be able to interpret and use the data. Data outputs will therefore need to be presented in ways that are interpretable for users with asthma, as well as standardised and clinically useful for their clinicians. If data is to be from a platform source, automated integration and interpretation is likely to be required. This will need to add value and reduce clinician burden/time, for example by permitting remote monitoring and passive collection of inputs that form a core part of the asthma review.

In placing a spotlight on the expert perspectives of end-users at high risk of adverse events, this study places those most likely to benefit from digital interventions at its centre. The interview’s timing at final visit maximised participation by limiting inconvenience, also aiming to reduce recall bias. This may, however, have reduced the opportunity for participants to process and contextualise their experiences. Other limitations include the overwhelmingly female and Caucasian sampling. Participants volunteered for the pilot study, potentially self-selecting as a group more likely to be engaged in their self-management. This is likely reflected in the relatively high adherence rates seen. Findings are furthermore the perceptions and experiences of this unique group of individuals, who offer their own valuable insights and perspectives.

In conclusion, data from participants of a pilot interventional study supports model integrating beliefs and habit formation to achieve behaviour change. Participants expressed a willingness for a more integrated, platform-based approach to digital self-management, but were clear that they expected their data to be used ethically. This study finds a general optimism for the potential of inhaler technology to have both personal and wider impacts on self-management and on shared decision-making.