Exploring patient and pharmacist perspectives on complex interventions for cardiovascular prevention: A qualitative descriptive process evaluation

Abstract Background The Assessing outcomes of enhanced Chronic disease Care through patient Education and a value‐baSed formulary Study (ACCESS) is a randomized controlled trial evaluating two interventions targeting barriers to care among those at high risk of cardiovascular disease: copayment elimination for cardioprotective medications, and a tailored self‐management support programme. We designed a process evaluation to better understand participant perspectives on the interventions. Design We used a qualitative descriptive study design, collecting patient and pharmacist feedback via individual semi‐structured telephone interviews and in‐person focus groups. Data were analysed inductively using thematic analysis. Results Fifty‐three patients (39 interviews and 14 in two focus groups) and 20 pharmacists participated. Copayment elimination provided quality of life benefits: minimizing the need to 'cut‐back', allowing 'peace of mind' and providing emotional support. Health‐related benefits included: improving adherence to covered medications, and helping to afford non‐covered goods. The only criticism was that not all medications and testing supplies were covered. Patients reported that the educational materials provided helpful information, acted as a reminder, improved confidence, improved adherence to medication, and helped initiate conversations with providers about indicated medication. Some participants felt that the educational materials were repetitive, overly medication‐focused and not tailored enough. Pharmacists felt that their patients benefitted from both interventions, which improved patient adherence and communication with their patients. Conclusion The success of interventions intended to change behaviour is largely dependent upon participant's feelings that the intervention is helpful. This process evaluation provided insights into participants' perceptions on these interventions. Reception of both was largely positive with a few criticisms noted.


| INTRODUC TI ON
Many patients at high cardiovascular (CV) risk have chronic conditions requiring them to engage in self-management to achieve optimal outcomes. 1 Self-management often includes taking medications regularly, following specified diets, self-monitoring and being physically active. 2 Unfortunately, patients often face substantial barriers that impede their ability to self-manage, pre-disposing them to development or progression of cardiovascular disease (CVD). 3 Two particularly common barriers include lack of knowledge regarding chronic conditions/CV risk, and how to manage those appropriately 4 ; and financial barriers, which impede access to the prerequisites needed for optimal self-management. [5][6][7] Recent evidence suggests that addressing these barriers has the potential to improve patient outcomes. Specifically, in health-care systems that require patients to pay for medications, reducing patients' financial barriers through policies such as copayment elimination has been demonstrated to improve adherence and reduce CV risk. [8][9][10] In diverse settings, improving chronic condition knowledge through self-management education has demonstrated promise in helping patients reduce CV risk, [11][12][13] and is recommended by major international guidelines. [14][15][16][17] Conducting qualitative research along side interventional studies is recognized as being important to elucidate feedback on trial interventions and implementation processes. [18][19][20] While there are high-quality trials supporting the role of interventions to target financial and knowledge-related barriers, qualitative studies about the implementation and acceptability of the interventions are not frequently reported in the published literature.
The Assessing outcomes of enhanced Chronic disease Care through patient Education and a value-baSed formulary Study, or ACCESS trial, is an on-going factorial 2 × 2 pragmatic randomized controlled trial, which started in November 2015. The trial aims to evaluate the impact of two interventions targeting financial barriers and lack of knowledge, among low-income seniors with high CV risk in Alberta, Canada. 21,22 The interventions are as follows: (1) a comprehensive tailored self-management education and support (SMES) programme including facilitated relay of clinical information to participants' health-care providers; and (2) elimination of copayments for select high-value cardioprotective medications (comparing with standard care: copayments of 30% of drug costs, to a maximum of $25/medication/dispensation). We hypothesized that these interventions would result in improved medication adherence and health behaviour changes, ultimately resulting in fewer hospitalizations, CV events and deaths.
While our interventions were grounded in theory, 21 their successful implementation was essential for changing participant behaviour.
Therefore, we designed ACCESS as a hybrid effectiveness-implementation trial (type 1) 23 featuring a qualitative descriptive implementation study to help us assess participants' perspectives on the interventions. The objective was to assess the implementation of our interventions and learn how to better support behaviour change for CV risk reduction in order to make further modifications to the interventions for the rest of the trial. We explored participants' and recruiting pharmacists' perspectives on the strengths and limitations of the trial interventions.

| ACCESS trial
This process evaluation was a substudy nested within the ACCESS trial. Eligible participants were adults over 65 years with an annual household income <$50 000 who were at high risk of CV events (diagnosis of any one of: chronic kidney disease, coronary artery disease, heart failure or stroke; or at least two of: diabetes, high cholesterol, high blood pressure or smoking). Enrolment was conducted through a variety of sources, the largest of which was community pharmacists, starting in November 2015 and ending in September 2018.22 4674 patient participants were randomized into one of the four groups (self-management education, copayment elimination, both interventions or control). Follow-up surveys and administrative health data are being used to collect data over a 3-year period. The primary outcome of the ACCESS trial is to determine the effect of these novel interventions on relevant clinical endpoints (mortality, myocardial infarction, cerebrovascular events, need for revascularization and chronic disease-related hospitalizations). The trial is on-going, with continued follow-up of patients happening through 2021. The qualitative work presented here was designed in the protocol to provide feedback on the study interventions and to allow for changes to the interventions to improve their acceptability to study participants. 21 There were approximately 500 participants enrolled and randomized in the ACCESS trial at the time this qualitative work was launched, with roughly equal numbers between the four arms of the trial.
The SMES intervention, branded as MOXIE, was codeveloped by the ACCESS team and a marketing firm (EMERGENCE Creative).
Both the branding and the fact that the messaging was provided by a fictional 'peer' (named Moxie) were designed to engender Health Research Foundation Grant (grant number 148441). evaluation provided insights into participants' perceptions on these interventions.
Reception of both was largely positive with a few criticisms noted.

K E Y W O R D S
participant focus group, pharmacist, qualitative, randomized controlled trials, vulnerable population personalized patient engagement. Those enrolled in this arm of the trial received weekly 'postcards from a friend', colourful trifold mailers with graphics focused on various aspects of CV risk reduction. This strategy was used as it was hoped that by giving people physical objects to place around their homes, they may serve as subtle reminders for behaviour change. Participants who stated a preference for electronic communication during eligibility screening received regular emails and access to a personalized electronic platform in addition to the mailers. The details of the programme are reported elsewhere, 21 but tailoring of health information was on the basis of a few specific variables collected at the baseline assessment: patient's individual CV risk factors, smoking status, current medication use, and patient's self-reported barriers to medication adherence. MOXIE provided additional 'health tools' for patients including pedometers and health tracking books.
Finally, MOXIE included a facilitated relay intervention whereby patients were sent letters regarding indicated medications, which were tailored based on the information provided at baseline.
Patients were instructed to take the letters to their prescribing physicians and pharmacists to start a discussion about cardioprotective medications the patient should be taking (specifically statins and ACE inhibitors/angiotensin receptor blockers).
Standard publicly funded health benefits for seniors in Alberta provide them with premium-free medication insurance coverage, but patients are required to pay a copayment of 30% of the list price of each medication, to a maximum of $25 CAD per prescription. 24 The copayment elimination intervention was designed so that those randomized to this would have their copayments reduced to 0 for select high-value CV preventive medications including the following: statins, beta blockers, ACE inhibitors, angiotensin receptor blockers, calcium channel blockers, diuretics, antiplatelet agents, anticoagulants, oral antidiabetes agents, insulin and smoking cessation aids.
Other types of medications were not covered by this programme, neither were diabetes self-monitoring supplies and other medical devices.
Participants who were randomized to the control arm of the trial received neither MOXIE nor copayment elimination. They continued to receive standard care through their regular health-care providers and retained their usual copayment for medications.

| Conceptual framework and study design
Two conceptual frameworks influenced the design of this study. We used a qualitative descriptive study design to achieve the objectives of our process evaluation. 27 Ethical approval was received from the University of Calgary Conjoint Health Research Ethics Board (REB13-1241).

| Sampling
Our target population for this qualitative process evaluation were trial participants (patients) and pharmacists who were involved in the recruitment process for the ACCESS study. We used separate purposive sampling strategies for patients and pharmacists. For each group, we defined several important characteristics that we wished to have represented in our final sample and contacted potential participants by phone. For patients, we anticipated needing to interview 30 individuals, though we planned to continue sampling until thematic saturation was achieved. Given that our objective was to receive feedback on the study interventions, we only sampled from those who received at least one intervention (i.e. not controls). We recognized that a variety of patient characteristics might affect their experience and perspectives on the interventions. The sampling considerations for participation in interviews included the following: • Gender (men/women): gender differences have been noted in self-management behaviour and perspectives. 28 • Income ($≤30 000 vs $30-50 000): those with lower incomes may stand to receive greater value from copayment elimination.
• Those who had perceived financial barriers on intake questionnaire: as above.
• Those who were not on indicated medications (i.e. ACE inhibitors/ ARBs or statins) during enrolment: as those not on these medications were thought to be able to provide an important perspective on how the intervention impacted their medication use.
• Those who indicated less-than-perfect medication adherence during enrolment: as above.
• Type of intervention received (MOXIE, copayment elimination, or both): in order to ensure that we had perspectives of each intervention adequately represented.
For focus groups, we used a similar, but more limited, set of sampling criteria, including the following: gender, income, type of intervention received and method of recruitment (word of mouth/referral vs seeing recruitment material in a public space).
For pharmacists, we anticipated the need to interview 20 individuals and we used the following sampling strata: • Urban/rural location: Pharmacists in rural locations often have more involvement in patient care than in busy urban pharmacies.
• Gender (men/women): gender-based differences in pharmacy practice have been noted. 29,30 • Those who had more significant involvement in recruiting for the study (i.e. recruiting 3 or more participants into the study) and those with less substantial involvement: those with more involvement were felt to be likely to provide more rich data, while those less involved could provide rationale for why they had not been more involved.

| Data collection
Individual interviews were chosen as the principal means of data collection, as we were asking patients about how their personal situations changed as a result of the intervention. Patients were eligible to participate in an interview 6 months post-enrolment.
Given that our participants lived all over the province of Alberta, in-person interviews were deemed to not be feasible. Therefore, we used semi-structured telephone interviews 31 to collect data from patients. Individual interviews were approximately 30-60 minutes in duration. Interviewers were aware of some of the participants' responses from their baseline questionnaire (ie whether they were taking indicated medications), which helped tailor the questions that would be asked of participants in the interview.
Interviews were supplemented with 2 in-person focus groups, held 1.5 years post-ACCESS trial launch, which were conducted at the University of Calgary and lasted 2 hours. These focus groups were primarily focused on patients' recruitment experiences, and using this forum helped provide a richer discussion and to inform a relaunch of trial recruitment activities.
Additionally, we conducted telephone interviews with pharmacists, which ranged from 20 to 50 minutes in duration. The domains included in the interview and focus group guides for patients were (see Appendices A and B) as follows: • Barriers to adherence/self-management, and changes to these.
• General feelings towards the specific intervention they received (copayment elimination, self-management education, or both).
• Whether the intervention had addressed financial and/or knowledge issues.
• Which aspects of their intervention had been helpful, and which had not.
• Why the intervention was successful at changing behaviour or not.
• What could be changed in the intervention to make it more helpful.
Pharmacists who helped recruit for ACCESS 22 were also interviewed. Topics included in the interview guide for pharmacists were (see Appendix C) as follows: • Familiarity with the ACCESS trial.
• Experiences with patients receiving the copayment elimination intervention.
• Experiences with patients receiving the MOXIE intervention, including the embedded facilitated relay intervention.
Patient and pharmacist interviews were conducted individually by two female research assistants who were trained in qualitative interviewing and had no prior relationships with participants.
Interview guides were piloted with other members of the research team to ensure the questions flowed and the interviewers were comfortable with the sequence of questions. Focus groups were conducted by DJTC (male, Co-Principal Investigator) and TSS (female, Research Coordinator) with other research staff present to help take notes. Participants were informed of the purpose of the interviews and focus groups. Notes were taken during all interviews and focus groups, and the proceedings were digitally recorded and transcribed verbatim by a professional transcriptionist.

| Data analysis
Interview transcripts were imported into NVivo 11 software (QSR) to help organize the qualitative data. Analysis was undertaken by two independent reviewers (research assistant and TSS). Thematic analysis techniques were used to code the transcripts. 32 We started with a preliminary coding template based upon the interview guides.
Codes were added by each individual reviewer in an inductive fashion. The coders met to resolve discrepancies, involving a third reviewer when necessary (DJTC). The team met to amalgamate similar, granular codes into broader theme groups.

| RE SULTS
We conducted individual interviews with 39 patients (Table 1); 38 patients declined to participate in an interview. More of those interviewed were men with lower incomes and native English speakers (Table 1). In addition, we hosted 2 focus groups (n = 8 and n = 6); these patients were generally older and had lower incomes (Table 1).
We approached 27 pharmacists for interviews, and 20 agreed to participate. As expected, the pharmacists had a relatively even distribution across the sampling criteria (Table 2). There were, however, slightly more men than women, and more pharmacists from rural settings than from urban centres ( Table 2).
The three main topical areas covered by the interviews/focus groups were as follows: (1) feedback on the copayment elimination intervention, (2) feedback on the SMES intervention, and (3) feedback on the facilitated relay intervention embedded within the SMES. Further supporting quotes for each area and subarea are provided in Table 3. Identifiers are provided for each quote, which link to the participant tables (Appendices D and E).

| Area 1: copayment elimination feedback
Participants who received the copayment elimination intervention were overwhelmingly positive about the intervention. There were two main themes of positive feedback regarding this intervention: benefits to quality of life and direct health-related benefits. The only negative feedback received about the copayment elimination centred on sentiments that the intervention did not cover enough medications or health supplies. Participants explicitly described that they wished the formulary would cover all medications (not just the high-value preventive medications) and their diabetes testing supplies. However, even though these were not covered, some participants voiced that because some medications were free of charge, their adherence to non-covered medications also improved and the cost savings enabled them to purchase health supplies and eat a healthier diet: 'Between the insulin and heart meds and all the rest

Pedometers 'I'd put it back on my belt, but it just wouldn't stay on. So, when I mailed the last thing in, I had got the notice from you about how I could get a new Pedo-thing' [Pt5].
Blister packing 'They were like more comfortable in taking them [ was too medication-focused, repetitive and/or not tailored enough.

| Area 3: facilitated relay intervention feedback
Those randomized to MOXIE also received tailored letters, instructing them to take the letters to their family physician and pharmacist.
The objective was to stimulate a discussion between participants and their health-care providers about CV preventive medications they should ideally be taking. This portion of the intervention received mixed reviews. Some participants showed these letters to their health-care providers (n = 15/24), while others did not (n = 9/24).
In our focus group, only 1 of the 6 patients brought the letter to both their pharmacist and family physician. Of the 20 pharmacists interviewed, only 6 of them recalled receiving or reading this letter.
One participant commented on why they did not bring these let-

| D ISCUSS I ON
This qualitative process evaluation of the ACCESS trial implementation explored patient and pharmacist perspectives on the study interventions. The interventions (copayment elimination and/or a tailored SMES programme) were received positively and seemed to address the underlying barriers to chronic disease managment and CV risk reduction (financial difficulties and lack of education).
Participants receiving the copayment elimination believed they  Table 1) contributed to lower uptake of facilitated relay, as patients may have been hesitant to start a discussion for which they felt ill-equipped. Some of those who did as the letters instructed had positive experiences, but the majority felt their providers were dismissive of the letters. Pharmacists' responses largely mirrored those of participants-that both interventions were helpful to their patients, but they were much less aware of the educational intervention than we anticipated.
Reflexivity is an important element of qualitative research. 33 We realized early on that participants' candour in interviews may be limited if they perceived that this feedback was being collected by the study team as they may have felt that negative feedback could jeopardize their on-going receipt of the trial intervention. To mitigate this view, we had an external team conduct the interviews who had no relationships with participants. However, the fact remains that there is a significant power differential between interviewers, as emissaries from the University, and study participants who were low-income seniors with chronic health conditions-this may well have contributed to the predominantly positive feedback received.
Throughout this process, the investigators of the randomized trial were able to reflect upon the feedback we were receiving. In response to this feedback, after approximately 1000 participants had been enrolled into the trial, the ACCESS trial interventions were significantly modified to address the concerns raised (Appendix F).
Our qualitative study found that the copayment elimination had the potential to confer benefits that were broader than those simply related to coverage of medications included on the study formulary. Reduced medication costs allowed participants to afford other health-promoting goods and enhanced communication between patients and pharmacists, since patients were no longer as focused on money. This would suggest that potential differences in outcomes from this intervention may be related to these 'side benefits' in addition to the direct increased adherence to covered medications.
While important, it is clear that financial barriers are not the only reason that patients may be non-adherent to medical therapies, given that adherence remains suboptimal, even in countries with full public pharmaceutical coverage. 34,35 Therefore, other interventions are also likely needed to enhance adherence. Self-management education programmes can contribute to individuals' health behaviours, including medication taking, in a variety of ways. Even though our intervention tried to provide tailored messaging, our results suggest that it would be strengthened by further and more detailed tailoring Our evaluation has limitations that merit discussion. First, the findings are subjective, and not all study participants have the same opinions of the interventions. However, given that we reached saturation in our sampling for both groups, we feel we have captured the breadth of perspectives on these interventions. Second, given that we intentionally chose to use a qualitative methodology, we cannot ascertain how prevalent positive or critical views of the interventions might be in the broader study sample; a quantitative study on this is currently underway. A strength of the study is that we explored perspectives of patients and pharmacists; however, we did not include physicians who may have received the facilitated relay letters, which would have added understanding of this perspective.
Third, we largely explored the acceptability of the study interventions, which does not necessarily translate into altered patient behaviour (ie improved medication taking, diet and physical activity).
These outcomes will be analysed in the on-going clinical trial. Finally, this study was conducted after the intervention had been received, over a relatively short period of time. With this design, we cannot speak to the sustainability of the programme in supporting behaviour change in the longer term, which merits investigation.

| CON CLUS IONS
The findings from this study provide insight into how interventions, like those being tested in the ACCESS trial, might work to help reduce patients' CV risk. This work is important as there are numerous information technology firms working with health-care organizations to provide digital platforms to patients to encourage healthy living. These qualitative findings will be bolstered by the findings of the trial, which will be able to answer whether adherence and clinical outcomes are improved as a result of these interventions. Future qualitative work is required to explore the experiences of people in clinical trials-particularly among individuals whose adherence/ health behaviours are not improved through receipt of the study interventions.

ACK N OWLED G EM ENTS
The authors would like to thank research staff Pam Leblanc, JoAnn Plested and Jane Fletcher for their help conducting the interviews, and Alexa Desjarlais for her help in arranging and leading focus groups.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request. If not on med → sƟll not on med When you started this study 6 months ago you indicated that you were not taking a staƟn/ACEi, and it seems that you are sƟll not taking this medicaƟon. In the past 6 months have you or your physician considered starƟng a staƟn/ACEi? If Moxie: did you take the leƩers you received from Moxie to your family doctor or pharmacist? If no, why did you not bring them to your healthcare provider? If yes, did you have a discussion around starƟng this new medicaƟon? If yes, why did you ulƟmately decide to not start the medicaƟon?

O RCI D
If non-adherent → now adherent When you started this study 6 months ago you indicated that you were taking a staƟn/ACEi, but that you weren't taking it 7 days of the week. Since then it appears you have started taking this medicaƟon more regularly.