Of the 53 apps, the most common persuasive design techniques were surface credibility (n=48, 90.6%) and liking (n=48), followed by trustworthiness (n=43, 81.1%), reminders (n=38, 71.7%), and suggestion (n=30, 56.6%). Social support persuasive design techniques were the least used across the different apps analyzed for health workforce capability, whereas primary task support techniques were the most common, with the exception of simulation. The apparent lower inclusion of persuasive design techniques around social support is a contrast, given that previous qualitative analyses have shown that rural health professionals perceive social and professional isolation to play a major role in building and maintaining their health workforce capability (Ramsden et al, unpublished data, 2021), and also provide multiple suggestions on how this could be achieved in this study. Examples given were related to online communities of practices, interprofessional learning (social comparison and learning), and telehealth improving trust between different disciplines, but also around having an online career coach (normative influence) and demonstrating online effectiveness of team performance and peer recognition, for example, by showcasing good news stories in health workforce capability. This study was based on the analyses by Oyebode et al [9], who reviewed 80 popular mHealth apps using the 32 techniques as described above. Briefly, the most common persuasive categories identified in Oyebode et al’s [9] study were personalization (n=77, 96.3%), surface credibility (n=69, 86.3%), trustworthiness (n=66, 82.5%), self-monitoring (n=64, 80%), real-world feel (n=59, 73.8%), reminders (n=57, 71.3%), suggestion (n=56, 70%), liking (n=52, 65%), expertise (n= 52), commitment/consistency (n=47, 58.8%), reduction (n=45, 56.3%), and tunneling (n=40, 50%). The authors provided persuasive strategy suggestions that can be used by app developers to increase the likelihood of behavior change. Similar to our study, Oyebode et al [9] pointed out that social interaction can motivate people to reach their behavioral targets, yet social support is rarely used in mHealth apps (17 of 80 apps, 21%). They thus recommend for future apps to include social support strategies.

Across health workforce capability domains, we found that health apps include the most persuasive techniques, as do the apps related to values and attitudes. Interestingly, the apps related to work organization have the least persuasive techniques, even though work organization plays a large role in maintaining or building health workforce capability [13]. For example, the ability to link to local allied health professionals through an online booking system would enhance a general practitioner’s (GP) workforce capability (Ramsden et al, unpublished data, 2021) by improving interdisciplinary, holistic care. The online system would also facilitate sending and sharing reports between the GP and an allied health professional. The online booking system would also reduce the likelihood of the patient not following up with an allied health professional as they get an appointment on the spot (social facilitation persuasive design technique between the GP, allied health professional, and patient). When designing digital tools that focus on work organization to increase capability, designers could focus on including persuasive design techniques, such as real-world feel, tunneling, and tailoring, which are currently underrepresented in the work organization category.

It appears that health apps include a wide array of persuasive techniques, and thus, building a health workforce capability app could learn from the design of capability-related health aps. It should be kept in mind, however, that Oyebode et al [9] did not find a direct relationship between the number of persuasive strategies used and app effectiveness (measured by user app ratings). The authors, therefore, argued that adopting fewer persuasive design techniques could potentially be as effective, as it potentially reduces the complexity of the app and thus prevents user cognitive overload. The number of techniques used is still under debate, and more research is needed in this area.

Khakurel et al [14] conducted a systematic review of 34 wearable device studies and concluded that wearable technology can potentially increase work efficiency for staff, improve workers’ physical well-being, and reduce work-related injuries. However, they also reported that technological, social, policy, data, and economic issues related to wearable technology remain. The authors developed a categorization of wearable devices, including monitoring, assisting, augmenting, tracking, and delivering content. Alhasani et al [15] applied a similar systematic review to 60 stress apps. The authors found that most apps use manual tracking, and pointed out that this can become boring and users might forget to log data. A combination of manual entry and automated tracking sensors (hybrid data capture) is recommended to improve the likelihood of reaching a set behavioral target change. For example, accelerometer data in a mobile phone can be used to predict the stress level at work [16] through the use of a data collection app stored on the phone that measures the continuous process of recording active apps, starting and ending time stamps, and duration of app use. By linking these data with the participant’s self-reported stress levels, researchers could accurately predict future stress levels. Ferdous et al [16] demonstrated an average accuracy of 75% and a precision of 85.7% as indicators of overall stress levels in work environments. This information can be used to inform stress reduction, organizational policies, and the interrelation between stress and productivity of workers. Reactions from health professionals varied according to their willingness to trial wearables, but at the same time, qualitative analyses identified that health professionals are receptive to certain elements of persuasive design techniques that relate to wearables, such as monitoring, feedback, reminders, reduction, and tunneling.

Content filtering and personalized interventions that match users’ needs have been demonstrated to improve persuasion power [15], such as the persuasive design techniques analyzed in our study: tailoring, tunneling, and reduction. The authors also recommend that users can customize the user interface and app features, such as color background, or their user profile as this gives them increased self-agency, a sense of control, and identity. Persuasive design techniques such as these could be implemented within a capability-building digital solution.

We propose to use a strengths-based approach to improving rural health professionals’ capability and include persuasive design techniques in digital solutions. For example, future applications can measure the health professional’s vitality or health workforce capability scores at different times throughout the day by asking them to rate their capability at random times or a set time throughout the week. Alternatively, some health professionals may be willing to use a hybrid model by augmenting self-ratings with wearable technology, such as vital signs or step counts. In either case, the health professional can view graphs or trends and analyses showing self-rated health workforce capability levels over time and time of day and display simple statistics, such as minimum, maximum, and average scores. This would reflect the persuasive design technique monitoring. Moving forward, the data can potentially also be used to predict future health workforce capability and recommend specific behavioral change interventions for the health professional to allow for tunneling and reduction. Furthermore, if the health workforce capability score falls, a reminder could pop up for the health professional to take appropriate action. This could include contacting a trusted organization or coach (system credibility persuasive techniques, such as trustworthiness and real-world feel, as linkage occurs with real people) to discuss the health workforce capabilities that they feel are hindering their current capability and identify solutions to increase their capability. It could also involve the health professional doing some exercise or other activity that helps them with their well-being and feelings of capability. Other simple persuasive design techniques, such as limited ads, will increase system credibility support through increased surface credibility. Given that health workforce capability is determined by the interplay of a complex array of factors [2] (Ramsden et al, unpublished data, 2021), this may well be the first step toward building a futuristic, completely digitalized tool to improve the health workforce capability support tool.

There are many ethical aspects that needs to be considered in developing a digital solution supported by real-time support by real people [17].

First, further research is needed on the ethical aspects of this type of data capture, as this will have major security implications. Indeed, Alhasani et al [15] note that every interaction with the app creates behavioral data, via audit trails or sensors, that can be analyzed in real time to predict users’ needs. This means that sensitive information about a user is being generated and stored.

Second, there is likely to be a perceived conflict of interest if an organization would own a health workforce capability digital tool and also provide the support to help health professionals with their capability.

Third, the active participation of a health profession can be strongly affected by factors such as the health professionals’ perception of how the data are owned, stored, and used. There is thus a need to balance the individual privacy of the data generated with the benefits that an amalgamated data set could provide to the workforce as a whole. As discussed, the ethics of using such granular data need to be further evaluated, and an acceptability study regarding the scope of use for the collected data would be timely. Nonetheless, as alluded to before, it is key to ensure that the data are kept secure and only used for their intended purpose, whatever that is determined to be. Users must also be reassured of this and their confidence maintained.

Lastly, relying on self-rated measures may reinforce unconscious incompetence [18] or result in social desirable answers where a health professional knowingly scores themselves optimistically due to the perception of negative effects on their professional registration or professional standing. This can be mitigated by building a culture of trust and support between the organization and the health professional and a shared understanding that these analyses are being done for the benefit of the health professional.