Exploring the potential of mobile health interventions to address behavioural risk factors for the prevention of non-communicable diseases in Asian populations: a qualitative study

Abstract

Background

Changing lifestyle patterns over the last decades have seen growing numbers of people in Asia affected by non-communicable diseases and common mental health disorders, including diabetes, cancer, and/or depression. Interventions targeting healthy lifestyle behaviours through mobile technologies, including new approaches such as chatbots, may be an effective, low-cost approach to prevent these conditions. To ensure uptake and engagement with mobile health interventions, however, it is essential to understand the end-users’ perspectives on using such interventions. The aim of this study was to explore perceptions, barriers, and facilitators to the use of mobile health interventions for lifestyle behaviour change in Singapore.

Methods

Six virtual focus group discussions were conducted with a total of 34 participants (mean ± SD; aged 45 ± 3.6 years; 64.7% females). Focus group recordings were transcribed verbatim and analysed using an inductive thematic analysis approach, followed by deductive mapping according to perceptions, barriers, facilitators, mixed factors, or strategies.

Results

Five themes were identified: (i) holistic wellbeing is central to healthy living (i.e., the importance of both physical and mental health); (ii) encouraging uptake of a mobile health intervention is influenced by factors such as incentives and government backing; (iii) trying out a mobile health intervention is one thing, sticking to it long term is another and there are key factors, such as personalisation and ease of use that influence sustained engagement with mobile health interventions; (iv) perceptions of chatbots as a tool to support healthy lifestyle behaviour are influenced by previous negative experiences with chatbots, which might hamper uptake; and (v) sharing health-related data is OK, but with conditions such as clarity on who will have access to the data, how it will be stored, and for what purpose it will be used.

Conclusions

Findings highlight several factors that are relevant for the development and implementation of mobile health interventions in Singapore and other Asian countries. Recommendations include: (i) targeting holistic wellbeing, (ii) tailoring content to address environment-specific barriers, (iii) partnering with government and/or local (non-profit) institutions in the development and/or promotion of mobile health interventions, (iv) managing expectations regarding the use of incentives, and (iv) identifying potential alternatives or complementary approaches to the use of chatbots, particularly for mental health.

Introduction

Non-communicable diseases (NCDs) are non-infectious, chronic conditions that often require long-term treatment and care, impacting on patients’ functional capacities and quality of life [1]. NCDs, such as cardiovascular disease, diabetes, or cancer, are the leading causes of death and disability globally, with more than two-thirds of all annual deaths attributed to these conditions [2]. Common mental disorders (CMDs), such as depression or anxiety, also cause significant health burdens and are highly interconnected and co-morbid with NCDs [3, 4]. Risk factors for NCDs and CMDs are diverse, including a complex combination of genetic, socioeconomic, and behavioural factors [5, 6]. Lifestyle behaviours, including tobacco and alcohol use, unhealthy diets, physical inactivity, poor sleep hygiene, or ineffective emotional regulation strategies, are key modifiable risk factors for the prevention and management of NCDs [7] and CMDs [8].

The Southeast Asia region has experienced a rapid increase in NCD-related deaths [9]. For example, Singapore has undergone rapid urbanization and economic development in recent decades [10], resulting in widespread adoption of modern lifestyles characterized by low levels of physical activity, diets high in sugars and saturated fats, and demands on personal resources associated with high levels of stress. This has contributed to the high burden of NCDs [11] and CMDs [12] in the country. Several health promotion initiatives have been introduced to support a whole-of-nation effort in reducing the prevalence of chronic conditions and their burden on health services in Singapore [13, 14]. Modifying lifestyle behaviours at a nationwide scale is, however, a difficult endeavour.

Mobile health interventions may be an effective, low-cost, and scalable solution to support individuals in adopting and maintaining a healthy lifestyle [15,16,17,18]. Smartphone-based chatbots, in particular, are a novel and rapidly growing approach to digital health in which users interacts with computer systems (‘digital coaches’) that imitate natural conversation through text or voice, with the potential to improve user experience and engagement [19,20,21,22]. Such mobile interventions could be particularly effective in Singapore given the country’s high rate of smartphone ownership [23] and initiatives to promote digital adoption [24, 25].

Despite their potential however, the effectiveness of many mobile health interventions has been hampered by low engagement and high user drop-out [26,27,28]. To develop engaging and effective interventions, it is important to understand the end-users’ perspectives regarding the target behaviour(s), and their perceived barriers and facilitators to mobile health interventions [29, 30]. This is highlighted in several intervention development frameworks whereby meaningful engagement with key stakeholders is conceived as an integral part of the intervention development process [31, 32]. While previous research has investigated the factors influencing the uptake and effectiveness of digital health interventions, the majority of the evidence comes from Western countries [22]. Given the effects of health interventions are often highly dependent on context [31], such findings might not apply to Asian populations. Moreover, previous studies in the literature tend to focus on clinical population subgroups [33] (e.g., cancer survivors [34]), who may have different views and perceptions compared to the general population.

Considering the aforementioned gaps in the literature, the present study aimed to: (i) explore perceptions of healthy lifestyle behaviours and mobile health interventions; and (ii) identify barriers and facilitators to using mobile interventions for health behaviour change, in Singapore.

Methods

Focus groups were conducted during July and August 2021 as part of a wider research project aimed at developing and evaluating a new smartphone-delivered lifestyle intervention for the prevention of NCDs and CMDs in Singapore [35]. Ethical approval was obtained from the Institutional Review Boards of the National University of Singapore (NUS-IRB-2021-232) and ETH Zurich (EK-2021-N-30). The Consolidated Criteria for Reporting Qualitative Research (COREQ; [36]) was used to guide reporting (supplementary file 1).

Sampling and recruitment

In line with recommendations [31, 32], this focus group study recruited potential end-users of a new smartphone-delivered lifestyle intervention to inform the first phases of the development process [35]. Our intended target end-users are middle-aged adults who have a higher risk of developing NCDs than younger adults and may be identified as such through the national health screening programme in Singapore [37]. Participants were thus eligible to join the focus groups if they were (i) middle-aged adults (35–55 years), (ii) English speakers, (iii) Singapore citizens or permanent residents, (iv) able to provide informed consent, (v) agreed to be audio recorded, and (vi) owned a smartphone.

Participants were recruited via mailing lists from the Singapore-ETH Centre (the institution where the research was carried out) and the Facilitators Network Singapore (a local company assisting with focus group facilitation), as well as online advertisements on the Singapore-ETH Centre website and social media. Those who were interested in participating were invited to access a link or scan a QR code that directed them to a participant information sheet. It provided details of the research team, explained the study purpose and procedure, and included a recruitment survey which collected sociodemographic information, contact details, self-rated confidence using apps and the internet, and previous experience using lifestyle apps. Recruitment surveys were reviewed for eligibility and a purposive sampling procedure was used to recruit participants, whereby those with indicators of lower socioeconomic status (SES) and those with prior experience using lifestyle apps were invited to participate first. Reasons for this sampling procedure were threefold: (i) people with lower SES are at greatest risk of developing NCDs but are among the least likely to engage with health interventions [38]; (ii) there are concerns that the digital divide could exacerbate health inequalities [39] and therefore it is important to understand the views of those with lower SES on the topic; (iii) to enrich the conversations and understand lived experiences rather than hypothetical scenarios. We aimed to conduct six focus group discussions with six participants (total n = 36), based on previous research suggesting six focus groups are sufficient to identify 90% of themes in a homogenous study population using a semi-structured discussion guide [40]. There were no prior relationships between the participants and the research team. However, some participants had taken part in previous focus group research moderated by the Facilitators Network Singapore.

Interview procedure

The questioning routes for all groups centred on perceptions, barriers, and facilitators to (i) healthy living and (ii) digital health interventions, including mobile lifestyle applications (apps) and chatbots. The content of the topic guide was developed by one researcher (JM) based upon the research team’s experience and the Theoretical Domains Framework, an integrative framework of behaviour change constructs that can be used to systematically understand behaviour and its determinants [41], as used elsewhere [42]. Questions and probes were developed to generate discussion within the TDF domains (e.g., “beliefs about consequences”, “intentions”, “knowledge”, “behavioural regulation”) and ensure the topic guide covered a wide range of factors that potentially influence users’ uptake and engagement with mobile health interventions (supplementary file 2 – topic guide). The first version of the topic guide was reviewed and piloted by the research team and the focus group moderators prior to commencing the focus group discussions. This allowed the facilitators to fully understand the goals of the research and make minor changes to the wording of the questions to ensure good understanding by the lay participants. The piloting process was not recorded or used in the final data analysis. The topic guide and questions were also reviewed and adapted iteratively following each focus group discussion (supplementary file 2 – topic guide).

Focus groups were conducted online using a videoconferencing platform (Zoom Video Communications, Inc., San Jose, CA). Discussions were facilitated by two professional local moderators from the Facilitators Network Singapore, one male (lead) and one female (supporting), who were independent from the research team and had no personal interest in the research topic. The lead moderator explained the guidelines, set ground rules, and introduced the topics for discussion. Two scribes (AA and RK) summarised participants’ comments using the chat function and took field notes during the discussions. One lead researcher (JM) attended all focus group discussions to assist the moderator if needed but made no contribution to the discussion. The questioning was semi-structured and open-ended to encourage discussion, with probes used to solicit additional information when required. Focus groups were 90 min in duration, with 15 min allocated to introductions, guidelines, ground rules, and an ice breaker, 70 min allocated to the main discussion, and 5 min to close the session. Participants received a S$30 e-voucher as reimbursement for their time. Focus group discussions were discontinued when data saturation was achieved (no new themes were identified) based on an iterative review of the field notes following each discussion and agreement among the researchers.

Data analysis

Focus group data were analysed using a two-step approach. First, relevant themes were identified through inductive thematic analysis [43, 44], following six iterative stages: data familiarization, code generation, theme development, review of candidate themes, theme refinement, and write up. Second, sub-themes identified were categorised as perceptions, barriers, facilitators, mixed factors, or strategies by using a deductive, direct content analysis approach [45]. Epistemologically, analyses were grounded in an essentialist/realism paradigm [46]. Based on this perspective, qualitative methods aim to provide low-inference and straightforward descriptions of the phenomenon of interest, using language that is close to the collected data (i.e., within the surface or explicit meaning of the participants’ comments, rather than at the interpretative or latent level). The Braun & Clarke [43] checklist criteria for the conduct and reporting of thematic analysis was carefully followed throughout the study (supplementary file 3 – thematic analysis checklist).

In stage 1 of thematic analysis (data familiarisation), focus groups interviews were initially transcribed automatically by the Zoom platform. Two members of the research team (BF and JN) reviewed the transcripts, together with the recordings, and corrected errors. Any difficulties in transcribing certain phrases were reviewed by two other researchers (SZ and JM). Comments from participants that had been written in the meeting chat were incorporated into the transcript. Transcripts were then uploaded to Atlas.ti V.9 to facilitate the analysis. BF and JN engaged with the focus group transcripts by reading and noting down initial thoughts and observations.

During stage 2 (code generation), BF and JN generated initial codes independently for the entire sample of focus groups transcripts. In thematic analysis, codes are short labels that represent important features of the data relevant to answering the research questions [43]. For example, comments such as ‘make the app as user friendly as possible’ or ‘the user interface has to be very, very simple’ were grouped under the code ‘user friendliness’. These initial codes by BF and JN were collated, mapped for similarities, and discussed as a group in weekly meetings. During these meetings, three members of the research team (AS, JM, OC) adopted the role of ‘critical friends’ [47], that is, reviewing candidate codes / themes and offering points for reflection and alternative explanations.

In stage 3 (theme development), the final list of codes generated by BF and JN were reviewed by AS, JM, and OC, who made suggestions on whether some codes could be merged or separated and developed an initial grouping of codes independently. BF used the team inputs to implement further changes to the list of codes and grouped the codes together into a tentative set of overarching themes and sub-themes. For the purposes of this study, themes and sub-themes were understood as a collection of similar codes that provides detail about the participants’ views on healthy lifestyle and mobile health interventions.

Stage 4 (review of candidate themes) involved a critical analysis of BF’s final list of themes and sub-themes by AS, JM, and OC, to examine whether they told a convincing story of the data (one that answer the research questions) as well as revisiting the raw data under each theme and sub-theme to ensure coherence and consistency. The whole team discussed and agreed on any revisions to the final list of themes (e.g., renaming or rearranging some of the sub-themes).

In stage 5 (theme refinement), names and descriptions for each theme / sub-theme were written up by BF and JM and discussed with the rest of the team. Where relevant, sub-themes within each of the themes were then deductively mapped according to perceptions, barriers, facilitators, mixed factors, or strategies, in order to provide further structure and meaning to the data. Finally, all authors were involved in writing up the analysis and findings.

Results

Six focus groups, each with between 4 and 7 participants, were held virtually over a two-week period during July and August 2021. Initially we recruited 36 participants but one participant did not attend the scheduled focus group and another decided to withdraw from the discussion, therefore a total of 34 participants (mean ± SD; aged 45 ± 3.6 years; 64.7% females) were included (Table 1).

Table 1 Characteristics of focus group participants (n = 34)

Thematic analysis

Five main themes were identified through thematic analysis: (i) holistic wellbeing is central to healthy living, (ii) encouraging uptake of a mobile health intervention, (iii) trying out a mobile health intervention is one thing, sticking to it long term is another, (iv) perceptions of chatbots as a tool to support healthy lifestyle behaviour and (v) sharing health-related data is OK, but with conditions.

Theme 1: holistic wellbeing is central to healthy living

Holistic wellbeing was the key overarching theme when discussing the topic of healthy lifestyle. Participants felt strongly that healthy living requires a balance between both physical and mental health, as well as activities of daily life. They viewed exercise, nutrition, and sleep as core pillars, or the foundations, of a healthy lifestyle but they also identified several other aspects of life that are important. For example, having social connections and a good support network, paying attention to emotional and mental wellbeing, being surrounded by positivity, spending time on spiritual practice, or having security and certainty in life were discussed.

Quote: “Healthy lifestyle has two aspects, one is a physical health, which is, of course, your diet and your exercise and the other one is your emotional health, which means to be spending time with your loved one, having some ‘me time’. So, to me that is healthy lifestyle, so everything has to be in balance.” (P32, Chinese Female aged 37 years).

Quote: “I would think that it is more than the physical and the food and the sleeping, for example, to me, also making meaningful connections to people around me.” (P29, Chinese Male aged 41 years).

Participants identified three support mechanisms that they felt would help them to live a healthy lifestyle. These were trained professionals, peers or family members, and the internet. Although professionals were perceived to be the best support for mental health, there was the view that mental health stigma prevents people from seeking their support due to concerns that a diagnosis could lead to negative consequences, for example on future employment. Instead, for mental health support, people may rely more on the internet, family and friends, or a support group.

Quote: “for the mental health challenges right, if they are not open for professional help, I’ll encourage them to join mental health support groups, you know there’s those available where diagnosis is not required, because of the stigma.” (P5, Chinese Female aged 39 years).

Barriers, facilitators, and strategies

The identified barriers, facilitators, and strategies to healthy living are summarized in Table 2. In general, participants felt people know how to lead a healthy lifestyle, but what prevents them from implementing healthy activities is a lack of motivation, discipline, or willpower. They described competing priorities, such as work commitments and taking care of family, as taking up most of their time, with a lack of personal time to engage in healthy activities. Therefore, participants felt a better work-life balance would facilitate healthy living and perhaps allow them the time to commit to a healthy lifestyle.

Table 2 Themes, sub-themes, and example quotes in relation to the participants’ perceived barriers, facilitators, mixed factors, or strategies

Environmental factors were viewed as both barriers and facilitators to healthy lifestyle in Singapore. For example, residents generally have good access to parks and outdoor activity spaces, such as outdoor gyms, and fitness centres are generally low cost. However, the weather conditions limit outdoor activity to the early morning or evening, when temperatures are cooler. Participants also explained that the food environment in Singapore does not promote healthy eating, as unhealthy food is cheaper and highly available, compared to healthy food.

Social influences were also perceived as creating both barriers and facilitators to healthy eating and participation in exercise and physical activity. Participants talked about how food choices are influenced by what others are eating or the food preferences of others, both when dining out and preparing food at home. They also talked about how involving family and friends in physical activity can motivate them to be more active.

Participants also described the actions they were already taking or could take, to lead a healthy lifestyle. Dietary strategies included making conscious diet choices, for example eating more balanced meals, ordering less ‘take away’ food, and cooking at home. Exercise strategies focused on integrating movement into daily activities, for example walking after meals and taking activity breaks during the workday. Different psychological strategies included setting goals, following a routine, journaling, and spiritual rituals like meditation and praying. Finally, time management strategies, including a focus on work-life balance and planning ahead for healthy eating and physical activity, were named by the participants.

Theme 2: encouraging uptake of a mobile health intervention

Theme 2 discusses the factors influencing the initial uptake of mobile health interventions. Participants’ comments were largely in relation to digital health promotion programmes offered by the Singaporean Government’s Health Promotion Board (HPB), although opinions on other commercially available apps were also discussed. In general, participants were open to using digital health interventions and apps. They acknowledged that the Singapore Government has made significant effort to promote health through digital programmes and that the incentives offered with these programmes (e.g., offering free wearable trackers) are effective in getting people to sign up initially. Some also commented that programmes or apps that are popular, and that many people are talking about, encourage wider uptake. Apps with essential features linked to lifestyle behaviours, for example, booking systems for fitness classes, or tracking body weight with a linked digital scale, also have higher uptake.

Quote: “I honestly feel that HPB’s model (free tracker coupled with incentives) is pretty effective and has seen a very good adoption rate.” (P24, Indian Male aged 38 years)

Quote: “For me, I feel probably you need to make the app more popular first, maybe some incentive to initially just jump start, like everyone now notices the [online retailer] app. You need to make the app popular so that word of mouth or that you get more and more people use it then. (P34, Chinese Male aged 42 years)

Quote: “I was introduced to it [Healthy 365 app] because there’s a dance class that goes on right underneath my window, and that was like so attractive, so I just appeared then they told me, oh yeah you can dance here, and you can download this app.” (P9, Indian Female aged 42 years)

Barriers, facilitators, and strategies

The identified barriers, facilitators, and strategies for the uptake of mobile health interventions are summarized in Table 2. In terms of barriers, participants found that the type of user could prevent uptake, for example, people who do not like to wear watches or activity trackers or those who are less technically savvy. They also highlighted decreased trust in health programmes with corporate links or sponsorships, as they found the backing of certain companies stood in direct opposition to the proclaimed health objectives of the programmes (for example in the case of fast-food companies). One key facilitator to the uptake of mobile health interventions was accessibility. Programmes that offer a digital platform that can be accessed by anyone, regardless of demographic or background, were seen as highly inclusive and more likely to be used. As an example, access to free apps and activity trackers, as in the case of the National Steps Challenge in Singapore, was seen as an enabler to join the programme.

The marketing strategies and outreach efforts used to encourage uptake of mobile health interventions were seen as both a barrier and a facilitator. On the one hand, roadshows and the use of radio jingles were mentioned as memorable tools that convinced people to join. On the other hand, one participant said that it was often difficult to find further information about ongoing programmes and it was too effortful to sign-up to participate. Similarly, there were mixed views regarding the cost of commercial apps. While participants felt cost was a limiting factor in the uptake and long-term use of commercial apps which ultimately led the abandonment of the app, they were also dissatisfied with the limited features available with free apps. Still, if apps were able to offer the same service as comparable offline services, such as receiving support from a health professional, participants saw cost benefits for the app.

Strategies for the uptake of mobile health interventions were centred around ways to reach people. Participants viewed places visited frequently by people, such as supermarkets, schools, or workplaces, to be the best places to reach and engage people with mobile health initiatives. In addition, participants felt certain individuals, such as doctors, friends, or community support groups, would be best placed to convince people to use mobile health interventions. Consistent and persistent messaging about a mobile health intervention over a long period of time was also mentioned as a way to encourage uptake.

Theme 3: trying out a mobile health intervention is one thing, sticking to it long term is another

Theme 3 explores user experience and factors influencing long-term engagement with mobile health interventions. Overall, participant’s perceived mobile health interventions as useful and effective, highlighting the benefits of certain features like providing free apps and activity trackers, being able to collect rewards, and using the technology to connect with group activities and workshops.

Quote: ““Lose to Win” is also organized by the Health Promotion Board…I took part in a few seasons, and I think it was really, really helpful. The first time when I take part in it they organize you into groups and then you meet regularly to exercise together, and also on nutrition workshops and yeah talks and workshops, which is very helpful. I got to know a few friends from that programme.” (P2, Chinese Female aged 46 years)

Participants mentioned that the novelty and excitement of a new app encouraged them to use it more in the beginning, but after some time, the novelty effect would wear off and they would either stop using the app as frequently or give up using it altogether.

Quote: “that’s quite exciting for the first two seasons of this ‘National Steps Challenge’, then after that, even though I sign up on the latest ‘National Steps Challenge’ but because I’m not clocking much [steps], I kind of also semi give up, yeah [laughs]. So yeah, the initial excitement has gone.” (P2, Chinese Female aged 46 years)

Quote: “I think the initial stage, because of the novelty people will get a tracker try to clock 10,000 steps, every day, but after a number of days, it wears off, so once the novelty wears off, we are back to our usual: people who are active remain active, people who are not active will still be not active because it’s like there’s no more incentive anymore really.” (P32, Chinese Female aged 47 years)

Barriers, facilitators, and strategies

The identified barriers, facilitators, and strategies linked to long-term engagement with mobile health interventions are summarized in Table 2. Participants felt technical issues, for example, problems synching apps with trackers or measurement inaccuracies, and high levels of user burden, such as when manually tracking dietary intake, were key barriers to continued engagement with apps. Tracking health and behaviours, particularly via passive sensing, and being able to compare data with others, were seen as facilitators to engagement. Additionally, social influences were identified as facilitating engagement in two ways, first in the sense that people want to use an app that everyone else is using, and second, it can be motivating when other family members and friends are using the same app and provide support to engage in healthy activities.

Rewards and incentives via mobile health interventions were seen as both facilitators and barriers to engagement with healthy lifestyle behaviours. On the one side, participants said incentives, such as receiving vouchers for performing healthy activities, can be effective in engaging people with a mobile health intervention and help to motivate them to change their behaviours. On the other side, they acknowledged that this form of extrinsic motivation is unlikely to work long-term and can lead to users abusing the system. For example, when rewards and incentives are no longer available users become disengaged, suggesting they will only perform the desired behaviours when an incentive is involved.

There were also mixed views on tailoring and personalisation of app content and features. Although tailoring was perceived as a very desirable feature, participants commented that existing apps often take a ‘one-size-fits-all’ approach and, while that approach might work for some people, it is unlikely to work for the majority. In general, if a mobile health intervention is not tailored, personalized, or bespoke to the individual, people will eventually abandon it.

Participants identified several strategies that they felt would improve engagement with mobile health interventions. Social strategies included collaborating as a group with family and friends to achieve prizes or using peer pressure to help drive behaviour change. In relation to incentives, while they were identified as creating both barriers and facilitators to longer-term engagement, participants still felt that they could be used to further facilitate long-term engagement. For example, they suggested different models of incentivisation such as delayed rewards or only allowing users to redeem points against healthier choice products. Regarding desirable features, participants highlighted the importance of personalization, whereby an app can cater directly to a user’s specific needs and tailor support or content as needed. They also discussed a desire to be able to track and visualize their progress and achievements easily. In addition, apps should be extremely easy to use, offer multiple health and wellbeing services, and integrate different data sources in one place to reduce burden on participants.

Theme 4: perceptions of chatbots as a tool to support healthy lifestyle behaviour

Participants’ views on chatbots were largely based on their experiences using customer service-related support. They had a negative perception of chatbots, perceiving them as useless, and described the process of interacting with them as frustrating. Specifically, chatbot interactions are time-consuming and the chatbot itself is often unable to understand questions or provide helpful answers. These issues were perceived to have a demotivating effect on users and made them sceptical about how useful a chatbot would be in the health context.

Quote: “I think it’s just based on past or present experiences that we had on chatbots, they’re usually not answering your question directly, I think a lifestyle, digital coach, whether health exercises or even mental health, I guess, I wouldn’t want to use it.” (P3, Chinese Female aged 48 years)

Quote: “Well, if it is still not a real person, then after all it is programming, so I wouldn’t say that I have much confidence in it, because, after all, for everyone the problem is unique.” (P32, Chinese Female aged 37 years)

Participants were also concerned that chatbots, or digital interventions more generally, might be used to replace face-to-face services and could lead to job losses for certain professionals, like psychologists.

Quote: “…might people lose their job or get replaced? If you can understand yourself, why would we need a doctor, why would we need a psychologist, physically?” (P26, Malay Female aged 38 years)

While participants were open to the idea of a chatbot supporting them with certain health issues or behaviours, there was concern that, for mental health support, a chatbot might not only be demotivating but also problematic and that AI might never be able to replace a human being.

Quote: “One of the bad things about digital health coaching is that it can come across as slightly impersonal and the other thing is that it may try to be motivational but it might end up having a really opposite effect” (P11, Chinese Female aged 37 years)

Quote: “For me, I think, I would like to know whether it is a chatbot or real person. I mean if it’s for food, okay lah, just a chatbot will do. But if it’s a mental health coach, I would prefer a human being. I mean a real-life human behind the chat. I mean, that’s why I feel that robotic or AI can never replace a human being, I’ll need the touch, yeah, so I think for mental health, I prefer a human being behind it.” (P1, Chinese Female aged 46 years)

Quote: You can have the Doctor Google or Alexa, you know, answering your queries anytime of the day, but if you’re talking about psychological well-being it’s pretty hard to trust just a robot, you know, answers which are like generic anyway.” (P23, Chinese Female aged 46 years)

However, they also acknowledged that the technology is still in its infancy and apps using digital health coaches could have potential due to their availability, accessibility, and cost-effectiveness when compared to offline options.

Quote: “So I mean of course there are like some established chatbots where people use it so much that the AI is good enough to be able to give good answers, but it always takes – yeah I mean it’s a good time to start – but it always takes a while before the data collected is good enough for it to give reasonably good answers.” (P6, Chinese Female aged 39 years)

Quote: “compared to like a physical coach, right, I think I would be more receptive to having a digital coach. Simply because of the availability, the accessibility and then I think working with a digital coach would also be more cost effective. In a sense, I think, probably for a digital coach, at most, the cost involved probably would be from your monthly subscription if it’s chargeable in that sense. But then, whereas you know if you have a physical coach, I think in the terms of costs here will be higher. So in that sense, I think I would be more receptive towards the app.” (P24, Indian Male aged 38 years)

Potential strategies to improve chatbots for the purpose of digital health coaching were largely focused on the qualities of the chatbot and the type of support that the chatbot could provide. For example, the chatbot should be empathetic, interactive, encouraging, and helpful, like a virtual friend who can provide motivational support and makes personalized health suggestions based on current progress. Participants also suggested giving users different ways of interacting with the chatbot, for example through speech or text, offering free trial periods to test out the chatbot, and providing an option to link users to real-life coaches.

Quote: “…the interactive element there, so if a digital health assistant wants to maybe be a little bit more effective, it should be your friend. It shouldn’t be something that is very much like your doctor sitting there and then looking at vitals. It should be the friend who says ‘hey, wanna come with me today and let’s go for just a short walk?’ and then after that before you know it, that app has brought you on a slightly longer walk, than the short walk that you originally wanted. And then the app says ‘hey I had lots of fun with you’ and all that kind of thing, so it has to have a very human side to it in order to be effective.”

Theme 5: sharing health-related data is OK, but with conditions

Participants were largely unconcerned about health-related data sharing, especially if they could see a benefit for them or the wider community. However, this sentiment was conditional on three aspects: (i) who will have access to the data, (ii) how it will be stored, and (iii) for what purpose it will be used. Participants needed to be able to trust the entity accessing and using their data. In this regard, government agencies were viewed as more trustworthy than private companies, as the latter were perceived to benefit from the data themselves by monetizing it. Participants were generally happy to share their data if it was aggregated, anonymized, and securely stored. Participants highlighted the importance of being informed about the proposed use of their data and being explicitly asked to provide their consent for this use.

Quote: “I know I’m sharing my personal data for the better of the community.” (P21, Chinese Male aged 47 years).

Quote: “I think what’s important is to generate more trust or like you know, for us to want to be more willing to share data, we’d like to know, what actually is being done with the data, like, why do you need this information and what’re you going to do with it and what do I get back in return for sharing that bit of data and how it can help me? (P6, Chinese Female aged 39 years).

Discussion

The aim of this qualitative study was to explore perceptions, barriers, and facilitators to mobile health interventions targeting healthy lifestyle behaviours in Singapore. We identified one broad theme relating to healthy living and holistic wellbeing, two themes relating to the uptake and continued use of mobile health intervention, and two narrower themes discussing the potential of chatbots in supporting health behaviour change and in what circumstances health-related data might be shared. These themes and relevant sub-themes are discussed below, with a particular focus on the practical implications for developing and implementing mobile health interventions in Singapore and, potentially, other Southeast Asian countries.

Participants showed a generally positive attitude towards digital health interventions in general and highlighted the importance of targeting holistic wellbeing, including body, mind, connectedness, and spirituality. Current health-related interventions, however, rarely address these elements together but rather focus on a single domain such as doing more exercise [48]. Even with health promotion programmes that target different domains, there is generally a greater emphasis on combining physical activity, sleep, and diet, while other less tangible elements related to mental health and wellbeing (e.g., emotions, social relationships, life values) are hardly integrated (e.g., [49]. While further research is needed, our findings suggest that digital health interventions targeting Asian populations in Singapore might benefit from adopting a ‘body and mind’ approach, highlighting the significance of the whole human entity and the interdependence of its parts (physical, emotional, and spiritual). This coincides with contemporary views on mental health, which emphasise the care of both mind and body [50]. Face-to-face programmes targeting holistic wellbeing exist and could serve as an inspiration for future digital health interventions (e.g., The Integrative Body–Mind–Spirit (I-BMS) psychosocial programme [51].

Most of the identified barrier and facilitators to leading a healthy lifestyle were in line with previous literature in Singapore and beyond. For example, several studies have highlighted lack of time (e.g., competing priorities) and cost (e.g., affordability of healthy food or exercise equipment) as key barriers to healthy living [51,52,53,54,56]. Similarly, social support (e.g., pairing with someone with similar goals) has been identified as a facilitator for health behaviour change [56, 57]. There were, however, some environment-specific barriers that are unique to Singapore and other Asian countries which should be considered by future intervention development teams. One of such is tropical weather, which features hot and humid temperatures all year-round and might discourage individuals from exercising or taking part in sports activities. Mobile health interventions including a physical activity component should consider such environmental factors and tailor their recommendations and activity suggestions accordingly (e.g., prompting individuals to exercise indoor or early / late in the day).

The food environment in Singapore was also widely discussed during the focus groups, with participants agreeing it is not conducive to healthy eating as energy-dense, nutrient-poor foods are widely available, affordable, and heavily promoted. This is in line with recent studies analysing the health of the Singapore food environment [58], and it might hamper individuals’ intentions and goals related to healthy eating. In addition, diets traditionally considered as healthy and widely recommended by public health organisations such as the Mediterranean diet might not be as accessible in Singapore as they might be in Western countries, considering availability and costs (e.g., Singapore imports 90% of its food). This should be considered by mobile health intervention developers and other public health initiatives targeting healthy eating in Asian populations, with dietary suggestions tailored to the local context and variety of ethnic backgrounds (e.g., Chinese, Indian, and Malay).

A key challenge for any mobile health intervention lies in how to motivate users to start using the app and/or wearable device and adhere to it long enough [27, 30]. Participants reflected on a wide range of variables influencing uptake and engagement with mobile health interventions. Consistent with previous research, factors such as user-friendliness, personalization, gamification and social influences were suggested as facilitators to engagement [58,59,60,62], while lack of digital literacy, poor clinical workflow integration, and technical issues were highlighted as barriers [30, 33, 60, 63]. A noteworthy difference to previous studies (mostly conducted with Western populations) is the influence that government endorsement has on a users’ decision whether to start using a given digital health intervention or not. Participants were also more willing to share health-related data with government agencies compared to private companies. This is consistent with previous research showing that Singaporeans have a higher level of confidence in the government, and are more supportive of government surveillance, compared to Western countries [64]. In addition, a substantial proportion of Singaporean adults have experience participating in government-supported, nationwide mobile health interventions such as the National Steps Challenge [14], which reached approximately 26% of the adult population in Singapore between 2017 and by 2019 [65]. Taken together, these findings suggest that future mobile health interventions in Singapore might benefit from partnering with government and/or local (non-profit) institutions, as this is expected to facilitate trust (e.g., data-sharing) and maximise uptake.

The use of rewards and incentives was generally viewed by participants as a useful strategy for both initial uptake and sustained engagement with a mobile health intervention. Financial incentives in the form of redeemable points or coins are common in the Singaporean app landscape, both for health- and non-health-related apps. However, the effectiveness of financial incentives is debated in the behaviour change field [66, 67], with some arguing that they undermine intrinsic motivation, which is thought to be a key process for sustained engagement [68]. In contrast, others defend their use as a useful strategy to encourage initial uptake, leading to more self-determined forms of motivation in the long-term [69]. Participants of digital health programmes implementing a reward system have also shown mixed views on the use of incentives [70]. Regardless of the intervention developers’ position, they should consider that some sort of external reward system might be expected in Singapore and act upon this, either by incorporating incentives or focusing on expectation management. Other engagement strategies mentioned by participants that could substitute the use of incentives are gamification, use of social support, personalisation, and app integration (‘one app for everything’, in line with the demanded holistic health approach).

Chatbots are rapidly becoming important gateways to delivering digital services in multiple areas, including health, customer service, or work assistance [71]. However, participants had mostly a negative opinion towards chatbots and their potential usefulness. This might be because their experiences were limited and mostly based on artificial intelligence (AI) chatbots used in the customer service domain, which are powered by natural language processing and use machine learning to understand the context and intent of a given question before formulating a response [22, 71]. AI chatbots are complex and highly dependent on the amount and quality of the data used to train the system. Interactions with AI chatbots often fail to meet user expectations, as questions are sometimes misinterpreted or cannot be answered, and this can be a source of frustration [72]. We argue that the use of rule-based chatbots might be a better alternative for digital health interventions until AI technology matures. Rule-based chatbots are based on decision trees instead of AI, using pre-defined questions and answers throughout the interaction (e.g., [21, 73]. With a rule-based chatbot the intervention developer has a high degree of control over the conversation flow and can ensure a smooth and effective interaction. This is especially relevant in the health context where information needs to be delivered in the most accurate and comprehensive way [74].

Lastly, it is worth mentioning that participants were particularly concerned about the use of chatbots for mental health support, compared to physical activity and/or healthy diet promotion, and felt strongly that chatbots should not replace humans in this regard. These views have also been reported elsewhere [75, 76] and may stem from the perceived lack of emotional connection with a chatbot, which is a key component of mental health interventions, as well as the reluctance to disclose potentially sensitive mental health information through a digital device. Participants mentioned different types of interaction styles that the chatbot could employ (e.g., empathetic, pragmatic, interactive), which is a relatively novel research topic in the field [77, 78]. Further research is warranted to explore chatbot interaction styles and their impact on engagement, acceptance, and working alliance in specific populations, as well as identifying potential alternatives to chatbots for delivering scalable and engaging digital health interventions.

Strengths and limitations

Strengths of this study include the use of the COREQ reporting guidelines [35] and the 15-point checklist criteria for ‘good’ thematic analysis throughout the transcription, coding, data analysis, and writing processes [43]. Specific good practices include ensuring the data have been transcribed to an appropriate level of detail, giving equal attention to each data item in the coding process, and clarifying our epistemology stance. The fact that several co-authors acted as ‘critical friends’ throughout the thematic analysis (e.g., offering critique and asking reflective questions, suggesting data to be examined through another lens, etc.) was also a key strength as it allowed the main analysts to think more deeply about the data and consider alternative explanations and groupings of codes. We believe this led to a richer, more comprehensive account of our participants’ perceptions and experiences of healthy lifestyles and mobile health interventions.

Our study also has some limitations that need to be considered. Results are based on a predominantly Chinese sample with a majority of women. Therefore, findings may not be applicable to all middle-aged adults in Singapore. Despite our purposive sampling procedure prioritising individuals with low SES, nearly 80% of our participants had a university degree, which is a strong indicator of middle-to-high SES. There is a steep socioeconomic gradient for both the incidence of NCDs [79] and engagement with digital health interventions and other health promotion initiatives [33]. Therefore, it is important that people with low SES are actively recruited and take part in future formative studies. Additionally, the purposive sampling procedure to recruit those with prior experience using lifestyle apps, and the fact that most of the participants felt confident or very confident using apps and the internet (as outlined in Table 1), means the opinions described in this paper are drawn from people who are comfortable using such mobile interventions. Furthermore, the specific recruitment strategy used in our study involved an active role from the participant, who had to contact the research team to participate in the study. Therefore, it is possible that the participants who took part were those already interested in health research or engaged in the topic. Finally, we did not gather data on health status and therefore we do not know whether participants’ opinions were influenced by existing health conditions.

Conclusion

This study adds to the limited body of literature exploring public perceptions of mobile health interventions in Asian populations. Findings highlighted several factors that are relevant for promoting a healthy lifestyle and for the effectiveness of existing and emerging mobile health interventions in Singapore and other southeast Asian countries. Future work in this area should consider: (i) targeting holistic wellbeing, (ii) tailoring content to address environment-specific barriers, (iii) partnering with government and/or local (non-profit) institutions in the development and/or promotion of mobile health interventions, (iv) managing expectations regarding the use of incentives, and (iv) identifying potential alternatives or complementary approaches to the use of chatbots, particularly for mental health.