This qualitative study was based on a prototype of the blended mHealth lifestyle intervention—The MIND-PRO app—built on insights from the HATICE and PRODEMOS trials [9,16-21]. To ensure usability and cultural relevance, we actively involved our target population by asking for their feedback on the prototype. In iterative cycles of feedback, the researchers and mHealth developers refined the intervention according to participants’ input. This co-design process incorporated culture-specific needs and preferences, aiming to align the intervention with its target population (Figure 1).

Initially, semistructured interviews were conducted to explore participants’ perspectives and requirements for a coach-supported lifestyle intervention. Among Turkish participants, we observed diverse views on culturally sensitive approaches. To gain a deeper understanding of these different perspectives, we organized a focus group to facilitate further discussion and refinement of the platform, ensuring that it was both user-friendly and culturally appropriate. The perspectives of Dutch citizens with low SES perspectives on mHealth interventions had already been qualitatively explored in a previous study [20] and tested in a randomized controlled trial for efficacy and implementation [16]. Therefore, in this study, we directly tested the MIND-PRO prototype in a focus group with this population. This study adhered to the COREQ (Consolidated Criteria for Reporting Qualitative research) checklist (Multimedia Appendix 1).

Participants were community-dwelling Dutch individuals with low SES or a Turkish or South Asian Surinamese background. We used convenience sampling by recruiting participants from community centers, with the help of key figures and the authors’ networks. For the Dutch or nonmigrant group, low SES was defined based on self-reported educational attainment, categorized as none, only primary education, or lower vocational or lower secondary education. A Turkish or South Asian Surinamese background was defined as having at least one parent born in Turkey or Suriname. Inclusion criteria were being 50-75 years old, owning a smartphone, and having at least one dementia risk factor (hypertension, dyslipidemia, diabetes mellitus, physical inactivity, smoking, depression, or being overweight). We acknowledge that referring to “individuals with Turkish or South Asian Surinamese backgrounds” may insufficiently represent the diversity within the group. However, specifying cultural background is essential to tailor the intervention effectively [22]. Similarly, using SES as a criterion may not fully capture individuals’ competencies and may be perceived as stigmatizing [23]. Our rationale for these definitions is that these groups are underrepresented in research, often leading to scarce and ineffective interventions targeting their specific needs.

The intervention design was guided by Bandura’s Social Cognitive Theory, which informed the goal-setting and self-tracking structure [24]. To enhance perceived ease-of-use and usefulness, we applied the frameworks of the Technology Acceptance Model [25] and Unified Theory of Acceptance and Use in Technology [25]. In the prototype, participants could set lifestyle change goals, record self-measurements, and monitor their progress, for example, on bodyweight, blood pressure, and physical activity duration and frequency. Progress was visually represented through graphs and color-coded diaries. Additionally, participants had the option to send messages to a coach, and the app featured a library of education materials specifically designed for this study (Figure 2).

Interviews and focus groups were conducted by six members of the research team: ARE (Dutch, female, MD), JEL (Dutch, female, MD), RR (Dutch, male, anthropologist, South Asian Surinamese migration background), RA (Dutch, female, master’s student, Ghanaian migration background), ES (Dutch, female, research coordinator, Turkish migration background), and MPH-B (Dutch, female, postdoctoral researcher). We worked with a multicultural team to promote alignment between interviewers and participants in terms of language and cultural background. Interviewing experience ranged from novice to highly experienced. Interviews and focus groups were conducted in Dutch or Turkish, with ES providing translation when necessary [26].

Topic lists (Multimedia Appendices 2 and 3) were developed by ARE and JL under the supervision of senior qualitative researchers (EPMC and MPH-B), drawing on the Technology Acceptance Model and Unified Theory of Acceptance and Use in Technology models [25]. ES contributed to ensuring accurate translation and clarification of concepts.

Interviews and focus groups began with general discussions on lifestyle perceptions, past behavior changes, and experiences with health apps. The intervention prototype was then introduced with minimal guidance, allowing participants to explore it independently or in small groups while researchers observed usability. Open-ended questions explored key features: user-friendliness, goal setting, progress tracking, coaching, and educational materials. Participants provided input on necessary adjustments to better fit their needs and preferences. Identified themes related to culture and gender were tested across groups. The interview guide was iteratively refined throughout data collection (Multimedia Appendices 2 and 3).

Adjustments to the app were made in real time, allowing participants to immediately confirm whether their preferences were understood. Most interviews were conducted individually, though four small-group interviews (up to four participants) were held at participants’ request, primarily due to language barriers. All interviews and focus groups were audio-recorded, with two researchers present—one leading the discussion and the other taking field notes.

Data analysis was conducted iteratively, allowing interim findings to inform adaptations to the prototype intervention and interview guide. After data collection, we thematically analyzed the transcripts based on Braun and Clarke’s [27] six-phase framework. First, the four interviewers familiarized themselves with the data through listening to the recordings and reading the transcripts. Second, each interview was categorized and double-coded by JEL and RA using MAXQDA software (version 24; VERBI Software) and discussed with ARE and the broader team to discuss initial findings and identify knowledge gaps. Data from the focus groups were directly discussed within the team. In case of uncertainty, the most experienced qualitative researcher (EPMC) acted as an arbitrator. Third, key ideas and concepts were extracted to generate themes and subthemes. Fourth, these themes were collaboratively reviewed for consistency in interpretation. Fifth, the core research team (JEL, RA, RR, and ARE) refined theme names and definitions before developing a final narrative synthesis.

We received a waiver from the Medical Ethics Committee of the Amsterdam UMC (METC reference 2023.0578), as the study did not fall within the scope of the Medical Research involving Human Subjects Act (WMO Act) in the Netherlands. Informed consent was obtained from all participants before the start of each interview and focus group (Multimedia Appendix 4). Participants’ privacy was protected by deidentifying their data and replacing their names with confidential participant numbers. Participants were compensated for travel costs, where applicable. The intervention developed in this study will subsequently be evaluated for effectiveness and implementation in a randomized controlled trial (MIND PRO: ISRCTN registry: ISRCTN92928122) [28].

Between October 2023 and November 2024, we interviewed 23 participants with a median age of 59 (IQR 55-63) years, of whom 65% (n=15) were female. The sample consisted of 14 participants with a Turkish migration background (10 female and 4 male participants; median age 55, IQR 50-59 y) and 9 participants with a South Asian Surinamese migration background (5 female and 4 male participants; median age 62, IQR 60-66 y). In the Turkish population, it was more challenging to include men. The focus group with Dutch individuals with low-SES was organized in an area housing mainly factory workers with generally low educational levels. This focus group consisted of 3 women and 10 men with a median age of 73 (IQR 68-73) years. A second focus group was organized at a community center with 7 Turkish women aged between 50 and 63 years to explore and refine contrasting perspectives observed in the individual interviews and to facilitate group discussion. As findings from the individual interviews with the South Asian Surinamese population were largely consistent, an additional focus group was deemed unnecessary. Our main findings are summarized in Multimedia Appendix 5, with supporting quotes from all three groups. Specific adaptations made to the mHealth intervention based on our findings are presented in Table 1. In describing the results, we focus on the overall blended mHealth intervention, which combines self-management of health factors with guidance from a remote coach. The app used by participants served as the delivery platform for the intervention.

Participants were generally open to mHealth interventions and recognized the importance of a healthy lifestyle. However, most associated this primarily with overall well-being rather than the specific link between an unhealthy lifestyle and dementia risk in later life. User-friendliness of the platform was deemed essential in all three groups, and could be improved through minimal text, visual aids (photos or videos), adjustable font sizes, clear navigation between pages to prevent disorientation, and an intuitive design resembling familiar apps (Q3 and Q4 in Multimedia Appendix 5). While all participants owned a smartphone, many of them—especially those with a low SES—felt less competent with modern technologies compared to younger users, emphasizing the need for clear instructions on app use (Q4 in Multimedia Appendix 5). To improve the comprehensibility of the information in the app, participants emphasized the use of simple, easy-to-understand language. Specifically, participants with a Turkish background mentioned the importance of providing a Turkish translation. For them, it was not just about understanding the language but also about the emotional connection it created (Q1 in Multimedia Appendix 5). Many participants with a Turkish and South Asian Surinamese background emphasized the importance of cultural congruence (Q2 in Multimedia Appendix 5). They expressed a preference for culturally sensitive educational materials and a coach familiar with their cultural background, to better accommodate culture-specific needs and preferences.

This study explored the views and preferences of Turkish and South Asian Surinamese individuals, as well as Dutch individuals with low SES, regarding a blended mHealth intervention. Using the MIND-PRO app as a prototype, we assessed its usability and acceptability across these populations.

Participants recognized the importance of a healthy lifestyle for disease prevention but found it challenging to sustain long-term behavior changes. They indicated that the MIND-PRO intervention could assist them in making lifestyle changes and serve as a personalized guide with tailored information to keep them motivated and engaged. In the context of individuals with low SES and migration background who often have a shorter educational attainment, participants appreciated the app’s simple and familiar interface, which made it easy to navigate. Most participants identified at least one lifestyle goal they wished to achieve, highlighting the importance of making gradual, manageable changes. They regarded the app as a reliable source of information for healthy lifestyle education. Although tracking progress was seen as useful, some participants felt uncomfortable comparing their progress with others. To build trust, a face-to-face meeting with the coach and personalized advice were suggested, as remote coaching was unfamiliar to many. The ability to make independent choices within the intervention was particularly important, as participants valued autonomy in their decision-making. Social support played a significant role in behavior change, with encouragement from partners and peers viewed as beneficial, whereas family obligations were sometimes seen as a barrier. The expressed need for cultural tailoring varied. While some valued adaptation in dietary and physical activity guidance, Turkish participants felt their cultural diversity already facilitated adaptability to Dutch customs. In contrast, South Asian Surinamese participants remained more attached to traditional dietary patterns as part of their cultural identity. Across groups, there was a preference for culturally aware coaching, tailored language options, and personalized physical activity guidance.

mHealth interventions have demonstrated their potential in lifestyle modification for dementia prevention [29], and can be suitable for ethnically diverse populations if barriers related to accessibility and usability are addressed [12]. Nevertheless, most studies on mHealth interventions for lifestyle modification and healthy aging show only modest effects and have high attrition rates [13-15,20,29]. Strengthening engagement strategies is essential to use the full potential of mHealth for lifestyle modification.

Self-Determination Theory suggests that behavior change is more sustainable when it is driven by intrinsic motivation rather than external pressure [30]. This principle of choice and self-governance is incorporated into our intervention, allowing users to customize the intervention by setting personal goals and self-tracking their progress to enhance the feeling of autonomy. A remote coach can further support motivation by using techniques such as motivational interviewing to identify goals aligned with each user’s needs and values [31]. Positive reinforcement from a coach can improve usability, adherence, and health outcomes [32]. Furthermore, self-tracking is a well-established strategy to improve adherence, which this intervention supports through simple graphs and color-coded feedback in progress journals [33].

In addition to coaching, social interaction with peers is often mentioned as a valuable tool for behavior change, such as by enabling users to share their progress [34]. The study participants had mixed views on this: while some valued privacy and felt discouraged by others’ success, others found social support and a sense of healthy competition motivating. This contrast is supported by a systematic review on social features in mHealth interventions, which reported varying preferences for social involvement [35].

Culturally tailoring the intervention was deemed important by some participants, and previous studies have shown that tailored web-based interventions enhance patient engagement and are more effective at facilitating lasting behavior change [15]. However, cultural tailoring assumes that ethnic groups are homogeneous and share similar core values, whereas individuals bring their own unique social, cultural, and historical experiences. Both groups expressed the need for tailored physical activity guidance suited to individual abilities and skills. Facilitators for physical activity included exercising with peers and receiving suggestions for activities that could be done at home. To address these cultural nuances, the intervention will feature culturally sensitive content on topics such as language, diet, physical activity, and health goals, shared on users’ timelines to support lifestyle education (Figure 2). Such tailored content, aligned with cultural norms, beliefs, and practices, is easier to integrate into daily life [36], increases trust in the intervention, and improves users’ self-efficacy [15]. We aimed to balance cultural tailoring with personalization through features like goal setting, self-tracking, and personalized coaching. This dual approach acknowledges multiple cultural identities while remaining sensitive to individuals’ unique needs [37].

A key strength of this study is its focus on diverse populations, particularly those from culturally diverse backgrounds or low SES, who are often underrepresented in preventive health care. Many mHealth interventions primarily target younger, highly educated users, potentially excluding individuals with lower health literacy, lower socioeconomic backgrounds, migration backgrounds, or older adults, thereby exacerbating health inequalities [38-40]. This highlights the need for inclusive interventions tailored to these groups [40]. Another strength is the use of a functional prototype, enabling iterative refinements based on direct participant feedback. Finally, the study benefited from a diverse research team with backgrounds in medicine, anthropology, mHealth, and public health, including members with Turkish, South Asian Surinamese, and Dutch backgrounds. This diversity enhanced cultural sensitivity and facilitated theoretical triangulation.

The study also has limitations. First, although a bilingual researcher (ES) facilitated communication with Turkish-speaking participants, the use of interpreters and translation may have led to a subtle loss of nuance or introduced bias, particularly given the cross-cultural nature of the discussions. However, as ES was part of the study team, she was able to indicate during the Turkish-language interviews which information was relevant and aligned with the aims of data collection. In addition, we engaged in regular reflection with her on culturally specific concepts that were challenging to translate. Second, the gender imbalance—characterized by underrepresentation of men among Turkish participants and overrepresentation of men in the focus groups with Dutch individuals with a low SES—limited our ability to assess gender-specific preferences. However, a similar prototype was previously tested extensively among Dutch individuals with low SES, where no significant gender differences in user preference and needs were observed [20]. Third, this intervention was tailored to the culturally specific preferences of Turkish and South Asian Surinamese populations in the Netherlands, as well as to the Dutch host population with a low SES. The transferability of our findings regarding culturally sensitive content to other contexts or populations is, therefore, limited. Future studies developing mHealth lifestyle interventions in other countries or for different target groups should explore and integrate the cultural needs of those populations. This study could offer a useful guide to which components may require adaptation, such as language, culturally sensitive educational content (eg, on physical activity and nutrition), and culturally aware coaching. At the same time, several of the core functionalities we identified—such as goal setting, self-tracking, accessible educational materials, and remote coaching—may be broadly applicable across diverse settings.

This study highlights the importance of tailoring mHealth interventions for lifestyle modification to reduce dementia risk factors in migrants and individuals with low SES. To be effective, interventions should integrate cultural preferences, provide positive and nonobligatory content, and respect users’ autonomy while ensuring that information is relatable and easy to understand. A crucial factor for success is the active involvement of end users in the development process, allowing for the identification and integration of context-specific needs and preferences. Notable cultural differences emerged: South Asian Surinamese participants placed strong value on traditional dietary patterns, whereas Turkish participants prioritized language accessibility, preferring an app that ranged from general guidance to culturally tailored support. These findings underscore the need for a balance between cultural adaptation and personalization, ensuring inclusivity while addressing individual variation in preferences and needs.