Noncommunicable diseases (eg, diabetes, heart disease and cancer as well as poor mental health) are considered key threats to global health [1] and are driven by factors such as physical inactivity, poor diet, tobacco smoking, and excessive alcohol consumption. A key global public health policy priority is to enact policies to ensure that the best possible health care is available for all [2]. In the United Kingdom, aims of the National Health Service (NHS) long-term plan [3] and priorities of UK government executive agencies such as Public Health England (PHE) are to provide a smoke-free society, to encourage healthier diets, and to improve mental health [4]. Encouraging the use of digital health interventions, such as smartphone apps, may be a cost-effective way of contributing.

Health and well-being smartphone apps can be cost-effective solutions for changing health behaviors [5,6]. Such tools can act as ideal platforms to deliver behavior change interventions [7] because of their availability, portability, and easy access [8]. Research has demonstrated early evidence of effectiveness of smartphone apps for smoking cessation [9], healthy dietary and physical activity promotion [5,10-12], weight loss [5,13,14], alcohol reduction among nondependent drinkers [15], and mental health promotion [16]. In addition, health apps can reach those resistant to seeking help in person (because of stigma) by improving access to behavior change interventions [17]. However, low uptake and poor engagement over time compromise the potential of health and well-being apps.

Uptake refers to the decision to select and install a health app [18]. The search for and selection of health apps tend to take place in commercial app stores such as Google Play for Android operating systems and the Apple App Store for iOS [10,19]. Thousands of health and well-being smartphone apps are available in the major app stores, a number that continues to grow [7], and the uptake of apps from commercial app stores tends to be influenced by indicators of popularity such as the app’s rank order, ratings and reviews, and the total number of downloads [19]. However, such popularity indicators are not necessarily positively associated with the effectiveness of an app [20] and may even be negatively related [21]. An associated problem with app uptake is that the vast majority of apps listed in commercial stores lack evidence about their efficacy [22] or effectiveness [23]. The need for quality marks in commercial app stores [24] and regulation of health apps and evidence for their effectiveness has been raised [16]. Better transparency in an app’s description to help people make an informed choice, including how the user’s data are handled, how the app was developed, benefits explained in lay terms, and descriptions of the app content, has been recommended [25-27].

A barrier to the uptake of evidence-informed apps is that not all apps are available to the public, or prominently displayed, via commercial app stores [22,24]. Therefore, fewer people may benefit from the high-quality tools available. Evidence-informed apps tend to be promoted within community or health care settings (often targeting a specific geographic region or country) or on curated health app portals. These portals are websites that present a list of selected health apps [28]. Health app portals can be government funded, such as the UK NHS’s Apps Library or PHE’s One You Apps portal, or curated by private organizations, such as App Script by IQVIA in the United States, the United Kingdom, and the United Arab Emirates; the MyHealthApps by PatientView’s in Europe and the United Kingdom; or ORCHA Health in the United Kingdom. These organizations can lend credibility to and have the potential to promote the uptake of selected health apps [29] by providing a list of safe; evidence-informed; tested; and, where possible, clinically effective health apps for the general public to choose from.

Research has focused on the identification of factors that influence the uptake of health apps in commercial app stores. There is an urgent need to explore whether the general public would be willing to use curated health app portals, which could improve the uptake of evidence-informed health and well-being apps [18]. Despite this need, little is known about the views on curated health app portals. This study aims to explore potential users’ views on factors influencing the uptake of health apps in general and on curated health app portals in particular using think-aloud and interview methodology.

The Capability, Opportunity, Motivation-Behavior (COM-B) model [30] offers a comprehensive framework for understanding behaviors. In the context of this study, the behavior of interest is the uptake of health and well-being apps. The model proposes that behavior arises because of the interaction of three components: capability (physical and psychological), opportunity (physical and social), and motivation (automatic and reflective). The Theoretical Domains Framework (TDF) [31], which contains 14 domains that can be mapped onto the components of the COM-B model, was also used. Together, the COM-B model and the TDF allow for a detailed analysis of data and identification of key factors influencing uptake in general and on curated health app portals in particular (Figure 1) [18].

This qualitative study applied a theoretical framework informed by the COM-B model and TDF to explore (1) factors influencing potential users’ uptake of health and well-being smartphone apps through searching on the internet and (2) their views on available curated health app portals.

This study elicited views and preferences of a sample of members of the public. The Consolidated Criteria for Reporting Qualitative Research checklist guided the design of the study [32] (checklist given in Multimedia Appendix 1). The think-aloud methodology [33] was applied to collect real-time data about health app selection on the internet and involved asking participants to verbalize their thoughts and impressions throughout the selection process. The researcher intervened only when a prompt was considered necessary (eg, during silent moments, asking questions such as “What are you thinking now?”). Following the think-aloud tasks, follow-up questions were asked to better understand the statements or utterances made during the tasks. Finally, semistructured interviews were conducted. The think-aloud tasks and the topic guide were informed by stakeholder consultation, which included views and opinions of lay persons (patient and public involvement representatives) and expert opinions of policy makers of this study. The study protocol was preregistered on the Open Science Framework [34]. The Faculty of Medicine and Health Sciences Ethics Committee at the University of East Anglia approved this study (reference number: 201819-089). The collected data are stored following the European Union General Data Protection Regulation and the University of East Anglia Research Data Management Policy. The data were anonymized, and all personal identifiers were removed. All participants read the participant information sheet and provided consent before participating in the study.

Participants were recruited through paid advertisements on Facebook. Adults in the general population were eligible if they were 18 years or older; were able to provide consent; owned a smartphone; would consider using a smartphone app to change their behavior in the future; and were able to attend an interview in Norwich, England, where the work took place. As a standard practice in qualitative research, the aim of this study is to gain a better understanding of the phenomenon of interest and to increase the coverage of perspectives rather than to recruit a population-representative sample [35]. Therefore, purposive sampling was used to promote the diversity of the sample (ie, age, gender, ethnicity, educational level, and employment) [36]. This included targeted advertisements on Facebook and the selection of participants to ensure the diversity of the sample. A total of 114 individuals responded to the Facebook advertisements and read a brief participant information sheet and completed the screening questionnaire. Of the 38 participants invited to an interview, 14 did not respond and 24 agreed to participate. Of these 24 participants, 6 were canceled for various reasons.

Before completing the online screening survey, participants were asked to read a brief participant information sheet describing the study. After reading and agreeing to participate, participants were asked to complete an online questionnaire to assess their eligibility and to collect descriptive data (Multimedia Appendix 2). Data were collected on age; gender; ethnicity, measured using the Office for National Statistics’ index; level of education; employment status; whether they had ever used health or well-being apps; whether they currently use a health or well-being app; last time they had downloaded an app; and frequency of app use. Participants who met the inclusion criteria were sent an email with a comprehensive participant information sheet and invited to participate in the interview. On the day of the interview, the interviewees received a printed copy of the participant information sheet, and written consent was obtained.

Face-to-face interviews were conducted between July and August 2019 and took place at the University of East Anglia (n=17) or participants’ homes in Norwich (n=1). The interviews were conducted by a single female researcher (DS), and no one else was present during the sessions. Each session started with a think-aloud exercise, with participants being instructed on how to verbalize their thoughts. First, they were asked to perform a search for an app they would potentially use to change the health behavior of their choice. They had a choice of using either a study laptop or their smartphone. Second, the researcher asked them if they were familiar with curated app portals. If they were not, DS briefly explained the principle and asked them to repeat the search using the NHS Apps Library and the PHE’s One You Apps curated health app portals (Figure 2). During the think-aloud sessions, positive reinforcement using verbal (eg, “You are doing great” and “Right”) and nonverbal (eg, nodding) communication was used to encourage participants to continue to express their views. In quiet moments, prompts were used (eg, “What are you thinking now?” and “Tell me what is on your mind”). Following the think-aloud task, questions regarding their experience with the uptake of and engagement with apps were asked (the topic guide is given in Multimedia Appendix 3). The sessions lasted between 26 and 63 minutes. Participants received a US $27.50 (UK £20) gift voucher as compensation for their time.

The sessions were audio recorded and transcribed verbatim by an external company. The transcriptions were checked for accuracy by the researcher undertaking the interviews. The data were analyzed using framework analysis following the stages of familiarization, identification of thematic framework, indexing, charting, mapping, and interpretation [37]. To ensure rigor, trustworthiness, and consistency, a percentage of randomly selected transcripts (2/18,15%) were independently coded by the second author (OP). The deductive thematic framework based on TDF was refined iteratively through repeated discussions with the second author (OP), and any discrepancies were resolved through discussion with the senior author (FN). Indexing was completed by the first author (DS) using QSR NVivo 12. The data were charted, and the responses were grouped according to the finalized thematic framework. During mapping and interpretation, the grouped data were examined by DS to identify patterns. During mapping, identified factors were classified according to their organic position rather than what they affect (eg, an opportunity factor may indirectly influence the behavior by increasing the motivation for uptake of a health app and influencing it directly). To aid comprehension of the findings for uptake in general and on health app portals in particular, data were analyzed and presented separately for these 2 topics.

To enhance the credibility and trustworthiness of the results [38], 30% (6/18) of participants were randomly selected and requested via email to provide feedback on a document with a summary of the findings and conclusions (member checking). They were asked whether they recognized their opinions and whether they agreed with the interpretation of the findings. A total of 2 participants responded to our request and confirmed that their opinions had been captured. In one case, our email was not delivered.

The researchers involved in this study are mixed methods researchers with experience applying the COM-B model and TDF to qualitative data. She disclosed her research interest to participants on the day of the interview, and no previous relationship was established between her and participants. The interviews were conducted by the lead author, a PhD candidate who has undertaken extensive training in the collection and analysis of qualitative data. Participants were encouraged to share their thoughts (both positive and negative) and to be honest. The interviewer felt that good rapport was built with the interviewees, and most participants (n=16) expressed their interest in learning more about the findings of the research. Field notes and a research journal were kept during data collection.

A total of 18 participants completed the interview. The average age of participants was 43 (SD 14) years, 50% (n=9) were female, 78% (n=14) were of White British ethnicity, 72% (n=13) were employed full time, 11% (n=2) had postgraduate qualifications, 94% (n=17) had used health apps before, and 61% (n=11) were using health apps at the time of the interviews, out of which 73% (n=8) reported daily health app use. Most participants were interested in changing more than one behavior (eg, losing weight, getting more active, and managing mood), and only 16% (n=2) of participants expressed a desire to change only one behavior. Participants’ characteristics are presented in Multimedia Appendix 4.

A total of 2 participants were satisfied with the app they were already using and did not wish to take part in the think-aloud exercise to look for a different app. The remaining 16 participants searched for apps targeting physical activity (n=6), weight management (n=4), mood and mental well-being (n=3), smoking cessation (n=1), alcohol reduction (n=1), and sleep (n=1).

The findings pertaining to factors relevant for both the uptake of health apps and views on curated health app portals are presented under the components of the COM-B model. Higher order themes and subthemes informed by the COM-B model and TDF are reported in Table 1.

Half of the participants who agreed to search for a health app (n=8) used Google Search as their first choice to find a suitable app, whereas the other half opened a commercial app store. The latter search among hundreds of available apps was described by most participants as difficult or a “minefield” (P2, P4, and P6). One participant described this task as being “far more complicated than I thought it would be” (P2). By the end of this exercise, only 3 participants found an app that they were willing to download and engage with further to change their behavior.

None of the participants spontaneously used a curated portal. Curated portals were then introduced to the participants, but none were previously aware of them. Curated health app portals were appealing to all participants, and they believed the portals would be likely to engender trust. However, searching for a health app on the NHS Apps Library and the One You App portal was a generally disappointing experience. Only 2 participants chose a health app from a health app portal (One You Apps), whereas the rest of the participants decided to continue the search in commercial app stores.

Online searches for health and well-being apps were found to be difficult. Factors influencing the uptake of health apps were mapped using the COM-B model and TDF. We found that social influences and participants’ beliefs about consequences (the perceived utility of the app) are key factors influencing the uptake of health apps. This conclusion was based on the frequency and salience of the themes that occurred during the interview. Curated health portals were found to be appealing despite the lack of awareness of their existence. However, the way apps are currently presented on these portals did not meet users’ needs because of a lack of certain features, such as lack of tailoring to the user’s requirements.

In line with previous research, the findings revealed the importance of the capability and opportunity factors, such as app literacy skills; health awareness and app awareness; esthetics of an app; low cost of an app; reading reviews and checking ratings; credible sources; and recommendations of apps from others, including health professionals [18,22,39,40]. Interestingly, the perception of the cost of an app appeared to be related to the perceived utility and credibility of the source. Although at the start, some participants were against paying for apps, the more useful an app was perceived, the more inclined participants felt to pay a fee. This phenomenon was observed for apps listed on health app portals, which were considered a credible source. More importantly, unlike apps listed on commercial app stores, there was implied trust in apps listed on curated health app portals by participants. In addition, some health apps are not available for downloading in both commercial app stores. Participants found it disappointing that some apps were only available for iPhone users. This is in line with previous research that found that out of 18 investigated health apps, only one-third were available to download on both major commercial app stores [28].

In terms of motivational factors, we found that perceived utility included aspects related to individuals’ perceptions about the presentation of an app and their previous experiences with health apps. Together, these shaped the way participants judged the usefulness of an app. This characterization underlines the need expressed by others previously for a better way to present health apps through a description that would lead to an informed choice (eg, the content of the app) [25-27] and potentially positively affect other motivational factors, such as the accuracy of an app and data protection [41]. Notably, concern about data protection and the accuracy of a health app was minimal when participants navigated on health app portals as opposed to commercial app stores.

There is a need to understand what design aspects generate positive or negative emotions and for whom. Emotions are powerful drivers of a behavior, which affects decision making (eg, app uptake) [42]. A key emotion identified in this study directly influencing the uptake was curiosity. However, this study emphasized the importance of positive emotions triggered by, for example, the credible source of an app and negative emotions triggered by restriction of information (eg, lack of understanding of the necessity of GP referral to download an app). Taking these factors into consideration may lead to better uptake with such tools.

Uptake and engagement are connected. Engagement without uptake is not possible, and uptake without taking into consideration the factors that are important for engagement is impractical. Some factors might influence both uptake and engagement; for example, our research suggests that the perceived utility of an app is one of the main factors for uptake. However, a previous study found that perceived utility was a predictor of engagement with an alcohol reduction app [43]. Therefore, where possible, uptake and engagement should be considered together as 2 linked constructs.

The main strength of this study lies in its methodology. Given that the aim of this study is to explore uptake with health apps and by applying a user-centered approach, the think-aloud methodology was the appropriate technique to use [33,44] as it will minimize recall bias. Involving stakeholders—patient and public engagement representatives and policy makers—in the design of the research enhances scientific rigor. The purposive sampling technique adopted enabled the recruitment of a wide range of participants that included the same number of females and males and having different levels of education and employment status, and the sample overrepresented ethnicity relative to local rates. The use of the COM-B and TDF to guide the data analysis is another strength of this study.

This study had several limitations. First, asking participants to perform the think-aloud task under observation may not be fully analogous to how they would perform a search when on their own. Second, some identified factors were difficult to define and describe because of the lack of specificity of the description provided by participants. These include esthetics of apps, often described vaguely (nice and elegant) and the cognitive load associated with engagement with these (easy to use). Third, for a qualitative research study exploring such a broad topic, we felt that information saturation was reached; however, it is possible that additional participants with more varied characteristics would have allowed us to identify additional concepts. Finally, during external validation, a randomly selected subsample of participants was asked via email to provide feedback on the summary of the findings. A total of 50% (3/6) of participants did not reply, and it is unclear whether these participants ignored our request or did not agree with the interpretation of the results.

This study has important implications for stakeholders in public health and policy makers who target prevention and health promotion using digital technologies and governmental bodies and trusted health organizations that provide curated health app portals. Low awareness, low app literacy skills, lack of availability on all major app stores, and lack of recommendation in primary care were identified as factors limiting the uptake of health apps in general and on curated app portals. These factors are important for improving the uptake of health apps. Selection was described as difficult. Therefore, there is a need for public guidance on how to identify evidence-based tools [18,22] and for health practitioners to promote and advise their patients on how to select appropriate health and well-being apps [40]. Raising awareness of such tools through both online and offline promotion channels might provide better access to effective apps.

Our findings could also help developers to reconsider the ways in which apps are currently presented on commercial app stores and app portals, which might, in turn, increase the uptake of evidence-informed health apps. The idea of selecting an app from a health app portal was appealing to all participants, although individuals’ needs were not met. These findings describe essential barriers and facilitators related to participants’ capability, opportunity, and motivation to take up health and well-being apps. For example, app descriptions and presentations that better align with individuals’ needs may increase the uptake of health apps on health app portals. These findings can also be used to inform the development of interventions that specifically aim to promote the uptake of and engagement with evidence-informed health and well-being apps, a priority within the NHS long-term plan (ie, digital first). By targeting the identified psychological influences on app uptake through further interventional work, organizations that provide app portals (eg, the NHS and PHE) should be able to increase their impact by helping people to better select appropriate apps. A summary of the recommendations for policy makers, providers, and developers is presented in Textbox 1.

Future research is needed to minimize factors limiting uptake, such as low awareness, low app literacy skills, and a lack of recommendations in primary care. Our results suggest that there is a need to better tailor the design and content of health app portals to better meet individuals’ needs. However, the mixed views on specific app designs indicate that more research is needed to investigate whether there are general design principles that are missed and could be followed to accommodate the majority of people or whether better tailoring and/or adaptive interventions should be considered instead. Future research may also want to consider comparing curated health app portals developed by private organizations with those developed by governmental bodies to investigate whether portal design–related features are considered less or more important than credibility and trust in apps listed on them. Experimental research is needed to assess whether there is a trade-off between credibility, social influences, and perceived utility of the apps presented on curated health app portals. Furthermore, with a growing concern around widening inequalities [45], solutions should be focused on reducing the digital divide and health inequalities that may appear as a result of the financial constraint of owning a smartphone and lack of sufficient app literacy skills.

Among the factors mapped under capability, opportunity, and motivation components of the COM-B model, social influences and the perceived utility of an app appear to be the core factors influencing uptake in general and on curated health app portals. Curated app portals are considered trustworthy and serve as a credible source for apps; however, there is disappointment with their current implementation of these portals. Uptake of health and well-being apps on health app portals, as opposed to uptake in general, appears to help address people’s concerns regarding data protection and the accuracy of apps. Health organizations that develop app portals may consider targeting the factors identified across the COM-B and TDF as part of additional experimental work, as this could help to increase impact through better selection of appropriate health apps.