We conducted twenty interviews with healthcare providers (n = 10) and patients (n = 10). Recruitment stopped once data saturation had been reached. Healthcare providers had a mean age was 32.9 (± 5.3) years, and 60% were female; they included doctors (40%) and pharmacists or pharmacy assistants (60%). Most healthcare providers (60%) had six or more years of experience in practice; the remainder had five or less years’ of experience. The patient demographics are reported in Table 2.

Table 2 Patient interviewee characteristicsNASSS domain 2: The technology

The participants identified many desired features for the proposed intervention (Fig. 1). Healthcare provider discussions often focussed on the flexibility of the system and the ability to identify specific user attributes. For instance, healthcare providers felt that elderly patients preferred the personal touch of a call, while working-age adults preferred the convenience of an app.

IDE “Middle-aged generations are more tech-savvy and also prefer more convenience.” [app or phone call based]

IDH “An app is plus and minus because a human is personal touch after all, an app can never replace that.”

Fig. 1

A summary of suggested intervention features

Patients mirrored these comments, noting issues with technology literacy and a desire for personal interactions, which would require the proposed intervention to take a multi-modal approach.

ID02 “A phone call may be considered old fashioned to some people. But it is the people’s touch.”

Flexibility was also considered important in relation to intervention frequency. Varying the intensity of the intervention could be one way to prevent users from becoming overwhelmed or losing interest (i.e., habituation to behavioral norms or attention fatigue).

ID02 “You know messages. We get a dozen messages everyday and we just delete it or ignore it.”

Healthcare providers suggested stratification based on the complexity of the case and the degree of adherence. Stratifying would inform the system on the intervention type and intensity, making it personalised and more effective.

Both groups of participants reflected on the challenge of medication routines, understanding the purpose of each medication, and following instructions. Patients and pharmacists suggested simplified prescription instructions, which could include visuals of the medications, and explanations of each medication’s purpose.

IDI “It would be even more helpful that these prescriptions are actually displayed to the patient in layman language right. Cause sometimes, I think right now the prescriptions don’t really show in layman language as well.”

Integration with the existing systems was another point raised by interviewees. For instance, using familiar applications like WhatsApp (a common messaging app) was suggested to avoid introducing a new system. Patients also mentioned the burden of installing new healthcare apps. Hence, avoiding new app or software installation was preferred.

ID05 “There are many apps that were launched by the government, many apps. For us who don’t have much education, we can’t understand…I can’t use.”

At the system level, integration with existing medical databases was perceived as essential for an effective intervention. Participants emphasised the need for real-time accurate data for adherence monitoring. Limited access to complete prescription records was acknowledged, but interviewees were optimistic this would resolve with the new electronic medical record system.

ID A “Hard for the AI. But if you are moving towards EPIC [a harmonized electronic medical record system] then it’s fine for checking for adherence.”

Other considerations for the intervention include data entry, data accuracy and complexity.

ID A “It would be challenging to ascertain reliability.” [patient inputting data] If it’s a caregiver I think they answer truthfully [re data entry].

ID C “So some patients maybe do not want to expose too much of their confidential information to the system.”

NASSS domain 3: Value proposition

Most expressed an interest in the solution and felt that it would enhance existing processes, particularly screening for non-adherers.

IDE “How do I keep track, or how do I monitor patients that do not collect their medicine? We do not have that now, we don’t monitor each patient that we have.”

Clinicians reported the need for such a solution due to existing support gaps but emphasised the importance of context. For example, older age was often reported as a barrier due to technological literacy or low interest, yet some older adults are curious about trying new things if supported. Others noted that caregivers would find the intervention appealing even if the patient did not.

IDG “50-60% of them are quite tech savvy [patients attending now]….If not, they would have family members who are very interested to make use of this.”

However, patient’s opinions were divided on this point. Some reported that technological literacy was prohibitive to this type of intervention as well as phone ownership.

ID02 “If you don’t own a phone how are you going to get the messages? Old folks do not have smart phones”

ID05 I really don’t know how to use, I don’t know anything related to digitalization. I don’t even really know how to use a phone”

During the interviews, healthcare providers rated example nudge interventions in terms of their perceived effectiveness, feasibility to implement, and acceptability to users (Figs. 2, 3, and 4).

Fig. 2

Perceived healthcare provider agreement on nudge interventions effectiveness

Fig. 3

Perceived healthcare provider agreement on nudge interventions acceptability

Fig. 4

Perceived healthcare provider agreement on nudge interventions feasibility to implement

Patients were interviewed on their perspectives of four nudge interventions (Table 3). Patients felt a need for reminders and remote contact from healthcare providers. While healthcare providers rated home delivery highly, patients felt it was not for them. Participants reported that their preference for self-collection was rooted in a desire to see the doctor in person, exercise, and check that the medication dispensed was correct. Home delivery was deemed more suitable for those with mobility issues. Healthcare providers reported that incentives would effectively encourage medication adherence, while patients were particularly critical of this idea. Patients believed they were responsible for managing their medication, and incentives were the wrong motivation. Participants were also concerned that this might be mistaken for a scam. Finally, participants emphasised the importance of keeping caregivers involved, regardless of the intervention.

Table 3 Participants reflections on example nudge interventionsNASSS domain 4: The adopter system

Role changes were anticipated by clinicians, patients, and caregivers if this system was implemented. For example, patients would register on the system while healthcare providers would monitor results. However, concerns over ownership of the system were noted. It was agreed that the pharmacy department should handle medication management matters, but limitations like data access, time, and resources were reported. Clinicians must also be prepared to use traditional methods, like phone calls, to engage less tech-savvy individuals.

For patients, a willingness to share data and interact with the system is important. Using the intervention must be supported by appropriate training and the availability of messages in different languages. The intervention must also avoid creating an additional burden, which may lead to poor engagement. Finally, access to technology should not be assumed, and engagement of the caregiver rather than the patient may be more appropriate in certain cases.