The study took place in Singapore and was comprised of an exploratory quantitative survey and qualitative study using in-depth interviews. For both the qualitative and quantitative component, we used purposive sampling from an existing community health study patient list. This list comprised participants from two neighborhoods in Singapore who participated in a community health study aiming to identify health needs of residents. Of those who consented to recontact, we recruited via telephone those aged 40 years and above with ASCVD or associated risk factors; Textbox 1 lists the complete inclusion and exclusion criteria. Once recruited, surveys and interviews were conducted in person at places of convenience to each participant. There was no overlap between recruitment for the qualitative and quantitative studies.

Research team members administered the survey instrument and conducted semistructured, in-depth interviews in the participant’s preferred language by staff fluent in that language. Both the survey instrument and interview guide were developed as part of the larger txt2heart collaboration, which we then adapted to the Singapore context (see Multimedia Appendix 1). Both covered topics such as patients’ sociodemographic characteristics, health care access, medication adherence, mobile phone technology usage (ie, ownership, access, and utilization), and interest in mHealth. We have included a consolidated criteria for reporting qualitative research (COREQ) checklist in Multimedia Appendix 2.

The quantitative exploratory component involved 100 participants who took part in either a face-to-face or phone survey as per their convenience and availability. The qualitative component involved face-to-face, semistructured, in-depth interviews with 20 participants. Out of the 20 participants interviewed, 19 agreed to audio recordings and 1 participant declined. For the latter, detailed field notes and an extensive memo were taken for inclusion in data analysis.

Informed consent for participation and recording was obtained before both the quantitative surveys and qualitative interviews using a participant consent form and information sheet. Participants could refuse to answer any of the questions and/or discontinue participation in the research at any time. Ethical approval was obtained from the Institutional Review Board at the National University of Singapore.

For the quantitative component, all data were analyzed using SPSS Statistics for Windows, version 24.0 (IBM Corp). Frequencies (n) and percentages (%) were used to summarize sociodemographic characteristics, the pattern of mobile phone technology usage, and interest in mHealth.

For the qualitative component, interviews were recorded and transcribed in full. Two research team members coded interviews using NVivo 11 software (QSR International), applying inductive approaches, thematic analysis, and techniques from the constant comparative method, where line-by-line analysis of early interviews is used on subsequent interviews to test preliminary assumptions [22,23]. Reviewers agreed on identified codes and themes, namely barriers and facilitators to adherence, mobile phone usage, attitudes and preferences, health information-seeking behaviors, and willingness to learn. Through these discussions, reviewers agreed that thematic saturation had been reached.

Using emergent themes from the analysis of the qualitative study, we developed a proforma to guide persona development (see Multimedia Appendix 3). The proforma captures participant demographics, socioeconomic factors, social supports, treatment adherence, health literacy and information seeking, and mobile phone usage; using standardized definitions (see Multimedia Appendix 3), each qualitative interview was analyzed against the proforma to bring forth the salient attributes of each persona.

This proforma enabled us to cluster participants into groups, taking into consideration low-to-high social support, treatment adherence habits, and mobile phone usage, as well as identification of common factors contributing to these. Using both qualitative and quantitative data analysis, we constructed personas that reflect behaviors and characteristics of the user population in Singapore with relevant codes identified by participant ID number to maintain confidentiality and exemplary quotes from the 20 qualitative interviews. While some studies opt to give personas names, faces, and individual characteristics, we have kept our personas as icons and titles in order to maintain the focus on broader trends in user groups rather than construction of individuals.

Participant characteristics from the qualitative study are outlined in Table 1. The quantitative component involved a total of 100 interviewer-administered surveys, inclusive of one incomplete survey that had missing data on mobile phone technology usage. Participant characteristics from the quantitative study are outlined in Table 2.

As persona creation must account for larger socioeconomic processes, it is important to take note of key health systems factors in Singapore. Health care in Singapore is generally accessible and affordable, with patients availing themselves of largely subsidized care offered in polyclinics (ie, government-subsidized general practice clinics), by private general practitioners, and even in tertiary care facilities for primary care and management of chronic conditions. Within primary care settings, it is a common practice for doctors to prescribe multiple months’ worth of medication to regular patients; this availability of medication is reported by our participants to facilitate their adherence. Further, the Singapore government subsidizes and offers various accessible schemes for those with chronic conditions; by and large, our participants reported using these schemes to fund their medications and most reported being able to afford their medications.

However, despite these systems-level facilitators to adherence, 65 out of 100 (65.0%) survey respondents reported suboptimal adherence to ASCVD medication as defined by the Medication Adherence Report Scale (MARS-5). Further, our qualitative findings highlighted different patient experiences, including self-titration of medications in response to symptoms and forgetfulness, as playing a role in suboptimal adherence.

We analyzed our survey data to report mobile phone ownership, access, use, and preferences in our target population, as well as to inform personas that reflect the technological context of a wider population in Singapore.

Among 100 surveyed participants, 99 (99.0%) fully completed the survey; 90 out of 99 (91%) participants owned a mobile phone, either an advance-feature or basic-feature mobile phone. Of those mobile phone owners, 70 out of 90 (78%) reported accessing their mobile phones in general at least once a day; the same proportion of patients—70 out of 90 (78%)—reported using SMS text messaging at least once a day. In general, participants predominantly used their phones for phone calls (87/90, 97%), SMS text messaging (60/90, 67%), and other text messaging services, such as WhatsApp (54/90, 60%).

Among mobile phone owners, 48 out of 90 (53%) indicated interest in receiving medication information via mobile phone. Among those interested in receiving medication information via mobile phone, SMS text messaging was the preferred mode of delivery (40/48, 83%).

These five personas present insight into our target user group for a treatment adherence intervention in Singapore. Based on these personas, we can see variability in multiple dimensions, including lifestyle and social factors, adherence factors, and mobile technology use and needs. These will be useful across the project process and will help to better refine the intervention implementation and dissemination strategy [11,24].

For example, within our population, The Socializer persona may provide an accessible group to initially focus on and engage with, as members of this group seek and disseminate health information via mobile phone. As this group has high social connectedness, they may also act as influencers and provide a platform for scale-up and buy-in from their social networks over time. Thus, initial message content design may be tailored to and tested with this group and may account for their belief in their abilities to self-manage and their perceived high health literacy.

Conversely, both The Busy Grandparent and The Hard-to-Reach personas more keenly seek information from familiar trusted sources as compared to The Socializer; thus, when considering how to onboard users, these groups may prefer to be introduced to the intervention through a trusted health care provider. This information helps clarify stakeholders to be involved in the dissemination and diffusion of the intervention.

From an implementation perspective, our personas show that users are not only using SMS text messaging, but also other messaging platforms such as WhatsApp. The feasibility of using such apps for an intervention may be worth exploring in the Singapore context. We also see how The Quiet Analog persona prefers to use their phone for calls only, thus the intervention may require a multiplatform approach in order to reach the widest cross-section of users.

Finally, The Newly Diagnosed persona points toward the utility of the intervention in a previously unconsidered group; this provides the opportunity to explore a modified or targeted intervention to onboard patients at the point of diagnosis or shortly thereafter. While the different personas provide insight into the different ways that users engage with mobile phone use, they should be used as a guide rather than a direct prescription of how an intervention should change, given that the context in which these personas operate do influence an intervention’s efficacy.

Our work explores the creation of biopsychosocial personas and how these may highlight different patterns of use and inform the design of mHealth interventions for adherence, as they assist in communicating the diversity in behaviors, preferences, and needs that can be represented through their use. Further, our study explores these personas in the unique context of Singapore.

Singapore has a rapidly aging population and increasing prevalence of chronic disease, which require innovative approaches to disease management [25]. With a high rate of mobile phone usage and digital penetration, as demonstrated by the 91% mobile phone ownership in our sample, Singapore is an ideal candidate for mHealth interventions. Therefore, our research provides a handle for technology designers to tackle behavioral variability in adherence among older adults.

Adherence behaviors are inherently complex and dependent on multiple factors, including the health system, socioeconomic factors, the medication itself, patient factors, as well as factors related to the condition being treated [26]. These factors were evident in our study sample, particularly with regard to health system factors, including ease of access and affordability of medications. Of note is the challenge in promoting adherence for asymptomatic chronic conditions such as hypertension and hyperlipidemia; indeed, our participants demonstrated self-titration and suboptimal adherence to these medications. Other studies have found similar findings, pointing to the challenge of promoting adherence to these medications and the need for innovative patient supports for adherence [27].

While mHealth shows promise in providing a far-reaching and cost-effective approach to facilitate adherence [28-30], caution is warranted in implementing these interventions indiscriminately to target populations. For example, in our study, participants reported a wide range of mobile phone usage behaviors, and even those who had high mobile use did not necessarily want to receive an mHealth intervention for their ASCVD medications. Patient-centered care seeks to provide appropriate services that address the unique needs and preferences of individuals; as such, mHealth approaches should adopt the user-centered design tenet of “know thy user” and use best practices in design to create appropriate and effective interventions in line with user needs and preferences [31]. For example, the majority of participants, when asked in the interview, reported not wanting an mHealth intervention for adherence; however, many reported liking an SMS text message appointment reminder. Thus, there is the potential to gain buy-in from these users as they are already using a similar service.

There is no one solution for digital approaches and users must be understood and contextualized. This is important when attempting to bridge the digital divide and provide mHealth interventions to aging populations who are not “digital natives” [32-34]. For example, our personas highlight the multiple usability issues reported by this population, including wanting content in languages other than English and physical barriers to use such as poor eyesight. We also highlight users who use their phones in specific ways, such as only for phone calls to family or for a wide range of activities such as games and videos, but who feel empowered to manage their own medications. To this end, health researchers must strive to consider a diversity of older adult user behaviors, including user motivations, attitudes, and preferences, in order to design effective and appropriate interventions for adherence to ASCVD medications.

Personas allow a middle path to highlighting the diversity of the user base in a manageable and actionable way. While it may not be feasible to design and implement mHealth interventions for the preferences or adherence patterns of each individual, as evidence has shown, it is also unwise to design for a user group that has been reduced to a homogenous typology, such as adherent/nonadherent or elderly [35]. Personas can then be used to begin discussions on appropriateness during the intervention design phase, enabling the creation of more robust user journeys by prompting discourse on how a given persona may react to a certain task, prompt, or call to action, and by providing reminders of who is being designed for throughout the research process [24,36]. As noted in our personas, a few groups expressed low interest in using their phone to assist in their medication taking. We include these personas as it is important to consider why certain groups may not want to take up an mHealth intervention and to move toward either strategies to gain their buy-in or different interventions to support them.

As the use of mHealth becomes increasingly popular, health researchers will need to adopt design thinking approaches and research methods, such as persona creation, in order to better provide contextually appropriate health interventions [13,24]. For example, a study by LeRouge et al explored the creation and use of personas in the Chinese Aged Diabetic Assistant app and showed how devoting time to the creation of personas in the early stages of an mHealth or eHealth (electronic health) intervention can help to shape design requirements, influence functional requirements, and guide the choice of appropriate channels for implementation and diffusion resulting in a contextually appropriate and acceptable intervention [24]. Thus, personas are an important bridging step to facilitate design discussions and allow important data on user experiences and perception to inform and shape the design process of mHealth interventions. Indeed, user-centered design aligns with the ethos of people-centered care; therefore, methods that better enable program designers and implementers to frame programs with diverse user groups in mind are an important piece to creating meaningful and effective mHealth services to promote adherence [37].

A strength of our study is the multidisciplinary project team and multi-method study design, which provides greater exploration of nuanced personal experiences and perspectives impacting adherence and preferences. Further, our study includes participants from multiple ethnic backgrounds, giving perspectives in both English, Mandarin, and Chinese dialects. Finally, to our knowledge, this is the first study of its kind in Singapore to construct personas to inform the design of an mHealth intervention for adherence among ASCVD patients. Further, the results of this study may be of use to inform designers and implementers of mHealth interventions in other Asian urban centers with high mobile phone use.

A limitation is that our study excluded individuals with disabilities that prevented them from participating in verbal interviews. Thus, we are missing the views of persons who are hard of hearing or with other disabilities who may benefit from an mHealth intervention for adherence. More research is needed in designing mHealth interventions and creating personas that reflect the unique needs of individuals living with disabilities. Also, our research focused primarily on mobile phone use for the purpose of an SMS text messaging intervention; with increasing mobile phone penetration, it would be worthwhile to explore in greater depth the use and feasibility of other messaging services (eg, WhatsApp and Line) for mHealth interventions. Further, as The Newly Diagnosed persona exemplifies, there is a need to explore interventions for those who are newly diagnosed to support them in achieving optimal adherence.

Future research in this area should include greater exploration of user journeys among diverse elderly groups to better understand how and when is best to deliver mHealth interventions. These studies should leverage additional user experience research techniques, including user acceptance and usability testing, as well as traditional public health study designs, such as longitudinal studies with routine follow-up, to determine the impact of messaging to different groups. Unlike persona use in other fields such as marketing [38], there is currently a dearth of robust health research on outcomes and evaluations of interventions that use personas. There is also a need for more multi- and mixed-method, user-centered research that synthesizes evidence on user behaviors and preferences for mHealth interventions for adherence among older persons. Finally, there is a need for greater collaboration between health researchers and those working in digital and creative fields to develop robust and feasible interventions that leverage innovative techniques and best practices.

As public health and medicine move toward wider adoption of mHealth, it is increasingly important that we pay due consideration to user-centered design techniques, and design thinking in order to create meaningful, patient-centered interventions for adherence to medications, as well as to evaluative if these methods result in more effective interventions. Our study leverages multi-method data to explore context, user behaviors, and preferences, and to put forth personas that typify a range of user group experiences to inform the design and creation of mHealth interventions to facilitate adherence to ASCVD medication.