Patient and practitioner views on a combined face-to-face and digital intervention to support medication adherence in hypertension: a qualitative study within primary care

Objectives To explore patients’ and healthcare practitioners’ (HCPs) views about non-adherence to hypertension medication and potential content of a combined very brief face-to-face discussion (VBI) and digital intervention (DI). Methods A qualitative study (N=31): interviews with patients with hypertension (n=6) and HCPs (n=11) and four focus groups with patients with hypertension (n=14). Participants were recruited through general practices in Eastern England and London. Topic guides explored reasons for medication non-adherence and attitudes towards a potential intervention to support adherence. Stimuli to facilitate discussion included example SMS messages and smartphone app features, including mobile sensing. Analysis was informed methodologically by the constant comparative approach and theoretically by perceptions and practicalities approach. Results Participants’ overarching explanations for non-adherence were non-intentional (forgetting) and intentional (concerns about side effects, reluctance to medicate). These underpinned their views on intervention components: messages that targeted forgetting medication or obtaining prescriptions were considered more useful than messages providing information on consequences of non-adherence. Tailoring the DI to the individuals’ needs, regarding timing and number of messages, was considered important for user engagement. Patients wanted control over the DI and information about data use associated with any location sensing. While the DI was considered limited in its potential to address intentional non-adherence, HCPs saw the potential for a VBI in addressing this gap, if conducted in a non-judgemental manner. Incorporating a VBI into routine primary care was considered feasible, provided it complemented existing GP practice software and HCPs received sufficient training. Conclusions A combined VBI-DI can potentially address intentional and non-intentional reasons for non-adherence to hypertension medication. For optimal engagement, recommendations from this work include a VBI conducted in a non-judgmental manner and focusing on non-intentional factors, followed by a DI that is easy-to-use, highly tailored and with provision of data privacy details about any sensing technology used.


INTRODUCTION
Medication adherence, defined as the level to which an individual takes medication as intended by their healthcare prescriber, is a worldwide public health concern. 1 Nonadherence to blood pressure lowering medication is estimated at 41%, which is relatively high compared with many other medications. 2 3 This is associated with increased risk of cardiovascular disease related morbidity and mortality. 4 5 Given that high blood pressure is responsible for nearly 20% of deaths worldwide, non-adherence to antihypertensive treatment is a global health concern. 6 Previous research into medication nonadherence has documented its complexity Strengths and limitations of this study ► To our knowledge, this is among the first qualitative studies to gather patient views on the use of sensing technology such as Wi-Fi or GPS within a smartphone app to support medication adherence. ► The study sought the views of a range of healthcare practitioners on incorporating a very brief intervention for medication adherence into a primary care consultation, a topic not previously explored in-depth. ► The use of stimulus materials provided detailed and focused responses for specific intervention components such as feedback on adherence and content of messages. ► While the sample size was small, the depth and focus of insights gained are sufficiently useful in informing the development and refinement of intervention components. Open access and multifaceted nature. 7 8 Two broad categories within this are (1) non-intentional non-adherence, a passive process due to factors not directly within an individual's control, such as memory or access difficulties, 8 9 and (2) intentional non-adherence, a more deliberate action whereby an individual makes a conscious decision not to take their medication due to their perceptions about or experiences with their medication or condition. 9 The multifaceted nature of non-adherence presents a challenge to those developing interventions to support adherence; for example, determining which factors to target, while balancing feasibility of delivery with likely effectiveness. Digital interventions (DIs) such as text messaging or smartphone applications (apps) offer interactive, low cost and scalable methods of providing support to individuals for whom medication adherence is a challenge. DIs are particularly suitable given the increasing use of these by people across the age groups for dayto-day tasks, such as apps for alarm clocks, calendars and shopping lists. 10 In addition, DIs can potentially lower costs compared with traditional face-to-face approaches, through reducing consultation time required with healthcare practitioners (HCPs), which may be particularly valuable at times when there is a high demand for consultations, for example, during the COVID-19 pandemic. 11 12 Evidence for the effectiveness of DIs in improving medication adherence is promising (eg, see Thakkar et al 13 ). In a recent systematic review of app-based interventions, patients using a smartphone app to support medication adherence for various health conditions were twice as likely to report taking their medications than those receiving usual care. 14 Specific to hypertension, DIs such as short message service (SMS) messages, smartphone apps, email and Bluetooth blood pressure monitors have been shown to improve medication adherence and lower both diastolic and systolic blood pressure. 15 16 Incorporating sensing technology into smartphone apps potentially expands the scope of DIs further. Passive smartphone sensors can collect user location data via GPS or Wi-Fi to enable the delivery of real-time support, 17 which is of particular relevance given that non-intentional non-adherence is strongly influenced by a person's physical environment. 18 Smartphone sensing technology has shown success in DIs across the domain of health and well-being (eg, see Cornet and Holden for a review 19 ) but user acceptability of such technology in a smartphone app to support medication adherence is largely unknown.
While user acceptability is key to use of a DI, potential users first need to install and engage with the DI for it to provide benefit. Primary care professionals, such as practice nurses or community pharmacists are ideally placed for encouraging uptake of DIs for medication adherence, for example, during a medication review or at the point of prescription collection. A DI used as an adjunct to a face-to-face consultation might therefore be a promising approach to support medication adherence. There is some evidence that DIs combined with tailored tele-based or web-based feedback from HCPs, improves adherence to long-term medication 20 and antihypertensive medication. 21 However, evidence is limited on how healthcare professionals can best promote the uptake of DIs for medication adherence. The acceptability of combining a DI with a very brief face-to-face intervention (VBI) delivered by a healthcare professional to support medication adherence has also not been widely explored.
This study aimed to explore patients' and HCPs' views on (1) non-adherence to hypertension medication and (2) a complex intervention designed to support medication adherence. Initial ideas for the intervention consisted of a very brief face-to-face discussion with a primary care provider, followed by ongoing support via a DI (SMS messages or smartphone app). Feedback from participants included preferred content of the intervention and factors likely to influence engagement.

METHODS
This study is reported in line with the Consolidated criteria for Reporting Qualitative research studies checklist (COREQ), 22 see online supplemental file 1).

Design
We undertook a qualitative study using semi-structured interviews followed by focus groups.

Recruitment and sampling
Patients were recruited for interviews from primary care practices based in the East of England (n=3) and East London (n=1). Practices were identified with the help of the Clinical Research Network, an organisation which supports the delivery of research within primary care in England. Patients were eligible to participate if they were: (1) prescribed at least one antihypertensive medication for at least the previous 3 months; (2) deemed non-adherent according to general practitioner (GP) practice records, with a blood pressure reading of over 140/90 mm Hg and/or gaps in filling repeat prescriptions in the previous 3 months and (3) used either SMS or smartphone apps. The practice administrator at each site generated a list of prospective participants that met criteria 1 and 2, which was screened by a GP or practice nurse. Eligible patients received a study pack from their GP practice in the post consisting of an invitation letter and participant information sheet. Posters highlighting the study were also displayed in the GP practices. Patients interested in taking part were invited to contact the researcher (MVE) via telephone or email, at which point the researcher checked that all three eligibility criteria were met before scheduling an interview.
A convenience sample of HCPs were recruited from the four GP practices taking part in this study. Healthcare practitioners were eligible to be interviewed if they were involved in the care of patients with hypertension, for example through medication reviews (conducted by a GP, practice nurse or practice pharmacist) or blood pressure checks and/or health assessments (conducted Open access by a healthcare assistant). The researcher invited HCPs to participate during the face-to-face study set-up meeting where they were given a study information pack. The researcher contacted the HCPs one week later to check willingness to participate and to schedule interviews for those who were interested.
Recruitment for focus groups followed that of the patient interviews. To address the low response from eligible patients, the eligibility criteria was widened to include patients prescribed medication for type 2 diabetes, as research indicates similar rates of medication non-adherence and barriers to adherence as for hypertension. 23 24 The eligibility criteria was also narrowed to ensure that participants were familiar with using smartphone apps (ie, SMS alone was not sufficient). The decision to cease individual interviews and switch to focus groups with patients was due to preliminary analysis from the interviews adding little new information to findings from previous research, and our experience of the usefulness of focus groups to gain feedback on the format, content and structure of DIs. [25][26][27] Data collection Patient interviews were conducted by one researcher (MVE) at patients' home, workplace or local library. HCP interviews were conducted at their place of work by the same researcher (MVE). Focus groups were conducted at community centres local to the patients' general practice and moderated by two researchers (MVE and JJ).
Interviews and focus groups were guided by flexible topic guides 28 developed by the research team, drawing on the perceptions and practicalities approach (PAPA) framework 18 and previous research experience in both the topic area and intervention development.
Topic guides were reviewed by patient and public involvement (PPI) representatives to ensure the questions were easy to understand and appropriate for the study objectives. Broadly, interview topics included: reasons for medication non-adherence, current practice of HCPs during medication-related consultations, and views on a potential SMS text message or smartphone app intervention that could support adherence. Example intervention content included medication reminders, advice and support messages, and feedback on adherence. See online supplemental file 2 for the topic guides and example DI content. HCPs were shown an example protocol for a very brief face-to-face discussion or 'VBI' to generate discussion (see online supplemental file 3). Components of the VBI included: introducing the digital support to the patient and inputting basic patient information via an online questionnaire to generate the tailored digital support.
Focus groups followed similar topic guides to the interviews, focusing on attitudes towards smartphone apps in particular, including the acceptability of sensing technology such as location sensing. To prompt discussion and gain feedback, both interview and focus group participants viewed stimulus materials of example intervention messages, including medication reminders, and smartphone app features, including graphs and images (see figure 1 for examples).
Written informed consent was taken in person by the researcher immediately prior to the interviews and focus groups commencing. All patients received a £20 voucher for taking part. Interviews and focus groups were audiorecorded and professionally transcribed  Open access verbatim. Interviews lasted on average 47 minutes and focus groups 1 hour and 28 minutes.

Data analysis
Analysis was informed methodologically by the constant comparative approach 29 and theoretically by the PAPA, which incorporates the blurring of and distinction between intentional and non-intentional non-adherence. 18 Interview transcripts were read and reread to aid familiarisation and identify preliminary themes; these broad descriptive themes were formed into an initial coding framework related to barriers and facilitators to medication adherence and a potential intervention. Each transcript was then coded systematically (MVE) using NVivo qualitative data-indexing software (V.12; QSR International) and the coding framework was refined throughout the process. The process was repeated for focus group transcripts; the coding framework was further expanded and refined, given the additional topics explored in the focus groups. A sample of interview and focus group transcripts were independently coded by a second researcher (JJ) to confirm and strengthen the validity of findings. Meetings between the research team (MVE, JJ and HE) facilitated data analysis including discussion of themes, subthemes and the interrelationships.

Patient and public involvement
All study materials (participant information sheet, invitation letter, study poster, consent form, topic guides and stimulus materials) were reviewed by representatives from the Cambridge University Hospitals PPI panel. We made a number of changes to the study materials as a result of PPI input. We adjusted the language to make the documents more accessible and ensured interview questions were sensitively worded and easy to understand from a patient perspective. PPI representative Jennifer Bostock provided input throughout the study and reviewed and commented on this manuscript.

RESULTS
Of the 126 eligible patients prescribed medication for hypertension who were sent an invitation, 6 were interviewed. All 11 HCPs approached by the researcher were deemed eligible and agreed to take part. Of the 218 patients prescribed medication for hypertension and/or type 2 diabetes who were then sent an invitation to a focus group, 14 participated (four focus groups with 3-5 patients per group). Recruitment of participants to focus groups continued until no new themes were emerging in relation to the specific topics covered. Patient participant characteristics are reported in table 1. Their mean age was 62.7 years (range 47-79 years), 60% identified as male and 85% as White British. Eighty per cent of patients reported using both SMS and smartphone apps, with the remaining 20% using SMS text messages only. All patients selfreported having occasionally missed or skipped their medication in the previous 3 months. HCP participant characteristics are reported in table 2; six practice nurses, two healthcare assistants, two practice pharmacists, one GP. Participants were recruited from four GP practices based in urban (n=3) and rural (n=1) locations. GP practice Index of Multiple Deprivation (IMD) scores, a measure of relative socioeconomic status in England based on postcode, ranged from 'least deprived' (n=2), to 'less deprived' (n=1) and 'more deprived' (n=1), see table 2.
To present the findings, we broadly categorise the key themes identified into the following categories: reasons for non-adherence, recommendations for message content, tailoring the DI, acceptability of sensing technology, and attitudes towards a VBI. We provide illustrative quotes below. See online supplemental file 4 for additional quotes from participants. For reference, DI refers to both SMS text messages and smartphone app, as the same intervention messages can be delivered using both formats.
Reasons for non-adherence Participants provided two key explanations of nonadherence to antihypertensive medication. First, for Open access non-intentional non-adherence, forgetting was the most common reported reason and was mentioned in three ways: forgetting to take medication, forgetting whether or not medication had been taken and forgetting to reorder the prescription in time. These overarching explanations were apparent when participants discussed the merits of a DI to improve adherence, as presented in the following sections. We begin with participants' views about a DI's messaging content, followed by tailoring and then sensing technology; the final section considers the role of the VBI component.

Recommendations for message content
Simple reminder messages were perceived as useful for both taking medication and re-ordering prescriptions.
Even if I'm in a hurry, [ The idea of receiving feedback on one's adherence, generated from self-report via SMS message or app, in a message of encouragement (eg, 'Well done!') was viewed as unnecessary. Participants were more receptive to schematic feedback in the form of a graph, score or percentage.  A 'snooze' function (similar to an alarm snooze) was well-received by participants, provided users could set their own parameters, for example, length of snooze duration and maximum number of snoozes.

It would be good for me 'cos I'm often not home when I'm supposed to take them, so if you hit the 'snooze' for an hour or whatever you choose it to be, […]and it'll keep reminding me again and I'll take the tablets [when I'm home]. [FG1, male]
A suggestion for tailoring by adding images of medications into the app raised more problems than benefits; participants pointed out that 'every time you get the medication, the box changes' [FG1] and it was felt this would create confusion, rather than help.

Acceptability of sensing technology
Participants were initially wary about the incorporation of sensing technology, such as GPS or Wi-Fi to determine location, into an app. They raised concerns about surveillance, typically referred to as 'Big Brother' [FG1, FG3] watching them. Participants were more likely to accept sensing technology if the perceived benefits (such as tailoring medication reminders to their specific schedule and locations) outweighed concerns about data privacy.

It would make it impossible to forget 'cos I'd just walk through the door and take 'em. That would be brilliant. [FG1, male]
Participants requested information to address these concerns, including: who has access to their data beyond the university (in particular, less trusted organisations such as insurance or marketing companies), where data are stored, and what happens in the event of hacking.

A: Who are you gonna share this with? That's all I'm worried about […] B: It could be pretty valuable information for insurance companies to put their premiums up. [FG4, male (A), female (B)] I think it'd be more reassuring to know it was a medical body behind it or a university body behind it; it gives it some substance and credibility. [FG1, male]
Participants wanted to retain personal control over the sensing function, with the ability to choose when the app tracks and records their location data as well as the ability to opt in/out at any point. Open access I think it's a good idea but, you see, some people wouldn't like it. [FG4,female] Discussions about sensing technology prompted participants to suggest further ideas for functions of an app. Participants in two focus groups suggested linking the sensing technology with the smartphone calendar, to proactively detect periods when away from home, triggering reminder messages to pack medication or reorder prescriptions. Above all, participants emphasised the importance of the DI being user-friendly for the target group, many of whom may be less familiar with smartphones.
The caveat I suppose might be that those that tend to have the chronic diseases tend to be the older age group so they may not be so tech savvy. We've got some patients who don't use mobile phones even now. [

HCP 02, GP, male]
Attitudes towards a VBI Patient-participants' discussions about the DI functions largely focused on addressing non-intentional nonadherence-mainly forgetting. On the whole, they were sceptical about a DI's success in addressing intentional non-adherence:

If they're not taking the tablets and they don't wanna take the tablets, why would they sign up for the app? [FG1, male].
HCP-participants suggested that the DI encouraged users to contact their healthcare provider if experiencing problems with their medication.
That would be really useful in that if they're stopping it for any reason it needs to come up with a message to say, "Please make an appointment with your GP. There may be alternative medications available which would suit you and you need to make an appointment to discuss that". [ All HCPs perceived the VBI element as feasible to deliver within primary care and recommended incorporating it alongside a medication review or blood pressure check. HCPs had two key provisos: training to help them deliver the VBI within the tight timing of 5 minutes, and a 'user-friendly' template that could be incorporated in existing computer systems for inputting patient data to inform the subsequent DI. HCPs also noted the need for sufficient training in using the DI itself, given their role in encouraging its use in their patients following the VBI.
I think that will be important, that whoever is talking about the app needs to know how it works and how you use it… Because if somebody who is recommending it doesn't know how to use it then you're not gonna buy into it. [HCP 09, Practice Nurse, female]

Summary of main findings
Patients prescribed antihypertensive medication and the HCPs that care for them, highlighted non-intentional (forgetting) and intentional (side-effects, reluctance to medicate) reasons for their non-adherence. Participants found a mobile DI that provided simple medication reminders and feedback messages acceptable. To facilitate engagement with the DI, participants recommended it was tailored to the needs of the individual and their medication regime as well as providing user control over the tailoring and other optional functions. The use of sensing technology within a smartphone app was acceptable to participants provided they received Open access comprehensive information about the associated use and confidentiality of their data.
While the DI was considered limited in its potential to address intentional non-adherence, HCPs saw the potential for a brief face-to-face discussion (or 'VBI') with patients in addressing this gap, when delivered alongside a DI. Incorporating a VBI into routine primary care was considered feasible, if it could be integrated into existing practice software systems and if training were provided.
Strengths and limitations of the study Drawing on relevant theory, 8 9 18 this study was conducted as development work with a target patient group to inform aspects of an intervention as part of a larger research programme. 30 While previous research has investigated the use of sensing technology and smartphone apps for health, 19 this study is among the first to gather qualitative data on the acceptability of such technology (eg, Wi-Fi or GPS) in a smartphone app designed to support medication adherence (see also Kassavou et al 31 ). While advances in technology can provide additional features to smartphone apps, it is important to assess the intended user group's views of such technology before its implementation. 32 We gained insights from a range of HCPs on the acceptability and feasibility of incorporating a VBI for medication adherence into a primary care consultation, a topic that has not been previously explored in-depth. The recommendations arising from our findings can inform the development and implementation of a medication adherence VBI in primary care. Developers should consider the following: the importance of the practitioner-patient relationship when discussing medications, exploration of patient-specific barriers to adherence, templates embedded within existing GP practice systems and sufficient training for HCPs.
The use of stimulus materials generated discussion in the interviews and focus groups, and provided focused responses for specific hypothetical intervention components.
We acknowledge that this is a small-scale qualitative study, where 85% of the patient sample were White British and 91% of the HCP sample were female. As such, the findings may be limited in their application to a patient and healthcare professional population. However, the depth and focus of insights gained provided rich data that were sufficiently useful in informing the development and refinement of intervention components for the wider programme, and to similar interventions.
We experienced challenges with recruiting patients through GP practices, particularly those who were nonadherent to their medication, a group who may be less likely to participate in a study of this nature. For future studies we would recommend widening recruitment methods to include patients not tied to a specific sample of GP practices, for example, via social media channels or community groups. We acknowledge the possibility that patients who are intentionally nonadherent to their medication may be unwilling to download an adherence app or receive SMS support messages. In these instances, alternative, more intensive intervention methods involving multiple behaviour change technique components may be considered appropriate, such as motivational interviewing delivered face-to-face and/or over the telephone. [33][34][35] Comparisons with existing literature The findings echo previous research that has identified the main reasons for non-adherence to cardiovascularrelated medication as forgetting and side effects, 24 36 as well as the broad categorisation of reasons into intentional and non-intentional. 37 38 In our study, this distinction was particularly helpful when considering which elements of an intervention were appropriate for targeting these two broad categories.
Participants with lived experience of hypertension saw little value in information-style messages (eg, about the consequences of non-adherence) in addressing intentional non-adherence. Rather, they suggested that such messages may be most helpful for newly diagnosed patients. This follows previous qualitative research in which mHealth interventions were deemed especially appropriate for 'newbies', 39 that is, patients with less experience in managing a health condition compared with those with established medication routines, for atrial fibrillation 40 and type 2 diabetes. 39 Similar to previous studies, 26 participants expressed concerns about receiving too many messages, suggesting this would influence (dis) engagement with the DI. Participants also emphasised the need for a DI to be as simple and easy-to-use as possible, another common theme in usability studies for medication adherence DIs, whereby difficulties with navigating a website, SMS or smartphone app have presented barriers to usage. 41 42 A related concern is the potential burden that self-monitoring DIs place on the user, for example, asking patients to self-report their medication taking behaviour within a set timeframe. 43 44 Our findings support the need for usability testing with the DI target users, which could include assessing any associated burdens or extra responsibilities placed on the user.
Participants in this study saw the benefit that sensing technology could provide but raised data privacy concerns about its use within a medication reminder smartphone app, requesting comprehensive information and user control. Similar concerns have been identified in previous research into location-sensing apps. For individuals living with HIV, the acceptability of location-based self-monitoring reminders was dependent on the purpose of the app and who would have access to their data. 45 Similarly, young adults in Dennison 46 worried about the storage of personal location data collected by health apps and wanted control over personalising the app settings.
Despite the privacy concerns, participants in this study viewed a location-sensing smartphone app more favourably if it was created by a university or charity rather than a commercial company. This follows user feedback of Open access other location-based apps for smoking cessation, 47 medication adherence 31 and mental health, 48 in which apps designed by universities or for research purposes were deemed more trustworthy by participants. This reflects the discourse around the ethics of mHealth, whereby third parties and insurance companies pose potential threats to the safety of patients' health data collected by sensors or smartphone apps. 43 These ethical considerations are of particular importance given the rise of mHealth in the healthcare sphere.
Recommendations for an intervention to support medication adherence The findings from this study have several implications for the development of a DI to support medication adherence. To encourage engagement with an intervention, it needs to be highly tailored to each individual. This includes: the timing and content of reminder messages (to address non-intentional non-adherence) and the content of support messages (for intentional non-adherence), where knowledge and duration of health condition varies between individuals. Furthermore, a key tailoring variable as recommended by HCP-participants was the individual patient's specific barrier(s) to adherence. Tailoring data can be collected using various methods, ideally before the start of the intervention for optimal impact. This could include a short questionnaire, in person or by phone with a practitioner, within a smartphone app, or via a set of SMS messages requiring responses. 26 It was common for participants in this study to be taking multiple medications per day, and most wanted this to be reflected in the medication reminders. This requires a balance between providing appropriate adherence support without over-complicating the DI or overburdening the user, resulting in reduced intervention engagement. 49 This study obtained novel insight from patients on the use of passive sensing technology within a medication adherence smartphone app. To increase the acceptability of sensing technology, future apps should explain the benefits that it can provide to the user, such as tailored medication reminders based on real-time location, or prompts to pack medication for upcoming holidays detected via calendar syncing. The app must provide a flexible opt in/out option for the collection of sensing data as well as information on how personal data will be used and stored within the app. Lastly, users may be more accepting of a location-sensing app created by a university or charity rather than a commercial company.
Primary care was viewed as an appropriate setting for HCPs to introduce patients to a DI and pair it with a brief behavioural face-to-face discussion, or 'VBI'. More specifically, this could address intentional non-adherence by exploring the specific barriers to medication adherence with patients. Using a non-judgemental approach for this, would encourage patients' openness, which in turn would provide more useful information for tailoring the DI and possibly making adjustments to the patient's regime as part of the usual care. This supports a body of literature on shared decision making, 50 which has demonstrated an association between an improved patient-professional partnership and medication adherence, for a variety of conditions 51 and for hypertension specifically. 50 52 Delivering a VBI requires skill, in order to incorporate all elements and within the short time frame. 53 54 Our findings indicate the importance of comprehensive training for healthcare professionals which incorporate the principles of shared decision making and the skills to deliver the intervention in under 5 minutes, as well as proficiency in using a DI. Lastly, the template for HCPs to complete the VBI and/or enter patients' details into the DI should be user-friendly and embedded into existing GP practice software systems.

CONCLUSION
Overall, patients and HCPs saw the benefit of receiving medication reminders via SMS text message or smartphone app. Intervention developers should consider an intervention that is highly tailored to the user, straightforward to use, and addresses data privacy concerns. The use of sensing technology was acceptable to patients, therefore, future research could investigate the feasibility of incorporating such technology into a smartphone app for adherence. A routine primary care consultation was viewed as an appropriate setting to introduce the DI to patients and discuss medication-taking behaviour with patients, but the feasibility of delivering it as 'very brief' that is, under 5 minutes, should be explored further.

Open access
Disclaimer The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

Competing interests None declared.
Patient consent for publication Not applicable.

Ethics approval Ethical approval for this study was obtained from the West Midlands -Solihull Research Ethics Committee, as part of NHS Health Research Authority approvals (Reference: 18/WM/0050).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available on reasonable request. Data are available on reasonable request from the corresponding author.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/ licenses/by/4.0/.

Topic guide: patient interviews
In the each interview, the interviewer will remind the participant of the project aims and check that they fully understand the participant information sheet, and will then take them through the informed consent process. Audio-recording of the interview will then begin.

Part 1: Questions to assess participant's experience with taking their medication
First I want to ask you about the medications that you're taking at the moment.

▪ How many medications have you been prescribed and what medications do you take?
o How many doses do you need to take and how often? ▪ What would you say are the benefits of taking your medication? ▪ What about the downside of /problems with taking your medications?
o How does this affect whether you take your medication or not? ▪ You said you're meant to take [XX] medications a day. How easy it is to remember to take them?

E.g. prompts o Do you have a particular routine? If so, what is it?
▪ What about altering the dose, is this something that you ever do? [can you tell me more about that?
Which medication? How do you alter it? Can you tell me why you alter it in that way? Can you think of any other reason why you don't take your tablets regularly? E.g. side effects] ▪ Is there anything that helps you to take you medication? Anything that makes it easier?

Part 2: Views on an intervention to support medication adherence
▪ Interviewer will outline the basic structure of the proposed intervention, using visual aid/schematic to help explain.

Topic guide: healthcare practitioner interviews
In the each interview, the interviewer will remind the participant of the project aims and check that they fully understand the participant information sheet, and will then take them through the informed consent process. Audio-recording of the interview will then begin.

Part 1: Questions to ascertain participant's experience with addressing non-adherence issues with their patients
▪ Can you describe your role in the GP Practice? ▪ How involved are you with prescribing and monitoring patients' medication? ▪ How do you usually ascertain whether a patient is taking their medication as prescribed? ▪ What do you tend to do if/when you think a patient is not taking their medication as prescribed?

E.g. probes:
o Medication-taking can be perceived as a sensitive topic, how do you tend to address this when starting your conversation? o Do you explore patients' reasons for not taking it; how easy/difficult is this? o What challenges do you come across? o What do you think can help support patients in their medication taking?

Part 2: Views on an intervention to support medication adherence
Interviewer will outline the basic structure of the proposed intervention, using visual aid/schematic to help explain.
o A very brief face-to-face session with a practice nurse or healthcare assistant or pharmacist followed by (ii) ongoing support via mobile/internet. o Include a draft outline of the 5 minute face-to-face consultation between HCP and patient.
▪ How comfortable/confident would you feel in delivering this?

Draft outline of a 5-minute face-to-face consultation (VBI)
• Start the consultation with de-stigmatising medication non-adherenceforgetting to take tablets and/or having concerns is common.
• Support is available. Introduce the digital support (SMS texts or mobile phone app) as a (fun?) way to help patients take themperhaps share personal experience with using it, demonstrate a key feature of the app/ text-messaging support.
• Complete short online questionnaire on behalf of patient: o Input basic information (patient name, age, mobile phone number) o Choose from drop down list of medications that patient is prescribed (tablet name, dose, frequency) o Ask patients to give their main reasons for not taking their medication (e.g. choose top 3 reasons from a list, or rank from most important to least important).
• Generate a code for patients to download the app/start the support. E.g.
QR code, link to website for patients to download app.
• Hand out pamphlet with further information.
• Endorse the patients' use of digital support.
• Record in the system for follow-up purposes. Other days I've forgotten completely and then I get home and I think, "Shall I take 'em now or shall I wait 'til tomorrow now?" I've done that once or twice.

(M)
That is a problem, whether you can take it if you've forgotten. (F) I want to know precisely, "Well, will that affect it big or a lot or not a lot or...? And is taking them all together more important than the hour?" […] It's the occasional expert knowledge that motivates people to stick with an app isn't it?
[…] It's the fact that if you could press a button and say, "What should I do now 'cos I've forgotten 'em?" might be just a great help.

Attitudes towards a VBI (HCP only)
Barriers and facilitators to implementation: must be straightforward to incorporate and deliver within primary care; five minute time limit potentially unfeasible; sufficient training in delivery of VBI It would have to be something that's very accessible on your desktop or as part of integrated into the clinical system. [HCP 03, Practice Nurse, female] I think this is quite a straightforward process, as I say, because a lot of this we should be doing already and then it's the add on bit at the end. [HCP 06, Practice Nurse, female] I don't think it's going to end at five minutes. I think it's probably going to go a bit more than that in some cases [HCP 05, Healthcare Assistant, female] I'm quite chatty so they will probably just chat with me, 'cause I make them chat, but it could be easily done in five minutes if you were concise and just got on with it. HCP 08, Practice Nurse, female] It would be really good that we've had the training on how to use the app, how to use the questionnaire… It might be useful even just to show us how you might do a consultation. You know, sometimes visually seeing something can be… Even Quite often what works well is just an almost practical session of: "This is what we're proposing. This is how you do the face to face consultation. This is what the apps look like" and then, you know, it then develops discussion. [HCP 06, Practice Nurse, female] The more you do and the easier would become and you get more slick at it, wouldn't you… What we would need is a little bit more information on the type of app and text that you are going to try and set up. And maybe have a dummy run and see how comfortable we felt with it and if there was anything we could do to make it slightly more slicker and more professional… A video clip or something of somebody doing it so we would know how to approach it and address it and stuff like that, will probably be the best thing. [HCP 09, Practice Nurse, female] BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) It's about sort of exploring why, what their understanding is of the medications that they are taking and trying to get an understanding of whether they recognise the importance of compliance. So it's kind of just exploring in a deeper way as to why they've not been taking. The reasons why patients don't take their medication we should really be exploring in any case… our role in the face to face is actually we can explore some of these things a bit easier. [HCP 06, Practice Nurse, female] Note: HCP = healthcare practitioner BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)