Barriers and facilitators of habit building for long-term adherence to antihypertensive therapy among people with hypertensive disorders in Los Angeles, California: a qualitative study

Abstract

Objectives

The aim of this study was to a) explore barriers and facilitators associated with medication-taking habit formation, and b) elicit feedback on the components of an intervention designed to help form strong habits for long-term medication adherence.

Design

The study design was qualitative; we conducted semistructured interviews between September 2021 and February 2022.

Setting

The interviews were conducted online, with 27 participants recruited at the Cedars-Sinai Medical Center in Los Angeles, California.

Participants

A purposive sample of 20 patients who were over 18 years of age, had been diagnosed with hypertensive disorder (or reported high blood pressure; >140/90 mm Hg) and who were prescribed antihypertensive therapy at the time of recruitment, along with seven providers were interviewed.

Results

Contextual factors included frequent changes to prescription for regimen adjustment, and polypharmacy. Forgetfulness, perceived need for medication, and routine disruptions were identified as possible barriers to habit formation. Facilitators of habit formation included identification of stable routines for anchoring, planning, use of external reminders (including visual reminders) and pillboxes for prescription management, and extrinsic motivation for forming habits. Interestingly, experiencing medication side effects was identified as a possible barrier and a possible facilitator of habit formation. Feedback on study components included increasing text size, and visual appeal of the habit leaflet; and imparting variation in text message content and adjusting their frequency to once a day. Patients generally favoured the use of conditional financial incentives to support habit formation.

Conclusion

The study sheds light on some key considerations concerning the contextual factors for habit formation among people with hypertension. As such, future studies may evaluate the generalisability of our findings, consider the role of visual reminders in habit formation and sustenance, and explore possible disruptions to habits.

Trial registration number

NCT04029883.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• In-depth interviews helped to contextualise and enhance our behavioural economics-based intervention focused on supporting habit building for long-term adherence to antihypertensive therapy.

• The respondents were recruited purposively from a high-volume specialised medical centre in Los Angeles, California.

• Interviews were semistructured and elicited patient and physician perspectives on the barriers and facilitators of medication-taking habits, along with feedback on critical components of the intervention.

• The study insights may be influenced by participant self-selection.

Background

Hypertension is one of the most common modifiable risk factors for cardiovascular disease, affecting nearly one-third of the population worldwide.¹ Efficacious pharmacotherapies are readily available, but adherence is essential for reaping long-term benefits.² Unfortunately, adherence to antihypertensive therapy (AHTs) remains low; during the first year of AHTs, patients possess medications on only 50% of days, and only 20% take their medications frequently enough to obtain cardiovascular benefit.^{3 4}

Several structural (eg, cost⁵), psychological (eg, perceived need; efficacy beliefs; fear of side effects⁶; cognitive decline among older patients⁷), and social factors (eg, lack of social support⁸) contribute to the problem. However, forgetfulness has been cited as one of the main barriers to AHT adherence.^{9 10} Meta-analyses have shown inconsistent and disappointing results of existing interventions aimed at improving adherence to AHTs.¹¹,¹³ For example, the effects of incentive-based adherence interventions only typically last over intervention period, wearing off once the intervention ends.^{14 15}

Habits are commonly cited by those with sustained high adherence.^{16 17} A habit is an automatic, reflexive behaviour in response to a contextual cue.^{18 19} Forming a habit requires repetition of a behaviour over roughly 3 months in response to the same cue,²⁰,²³ requiring strong intrinsic motivation during the habit formation phase. Once the habit is established, there is little to no need for cognitive effort, that is, one does not require motivation, self-regulation or deliberation to execute the behaviour.^{24 25} Habitual pill-taking could be especially beneficial for those experiencing cognitive decline, as habits do not require reliance on memory for high AHT adherence.

A common habit formation intervention ‘anchors’ or ties the target behaviour to a pre-existing routine acting as the external cue.^{18 26 27} However, typically only about half of the participants receiving this intervention repeated the behaviour long enough to form a habit.²⁸,³² To remedy this drawback, we designed ‘Behavioral Economics for anti-hypertensive Therapy Adherence’ (BETA), an intervention that uses text messages and small conditional incentives informed by behavioural economics (BE) to help participants stay motivated during the habit formation phase. The text messages keep the behaviour salient, and the conditional rewards help counter present bias (where one excessively discounts future health benefits relative to current costs³³) during the 2–3 months it takes to successfully form a habit. Tying incentives explicitly to measures of habit formation (by conditioning them on adherence in line with the participant’s anchoring plan) is conceptually novel and different from existing incentive interventions that tie the reward to overall adherence, rather than specifically to support habit formation.

This qualitative study marks the formative phase of the pilot randomised controlled trial (RCT) of BETA, where we elicited, through semistructured interviews, barriers to and facilitators of medication-taking habit formation among people with hypertension in Los Angeles, California, for intervention contextualisation.

Methods

This study adheres to the Standards for Reporting Qualitative Research guidelines.³⁴

Study setting

The study reported in this paper aimed to provide crucial insights to help support the contextualisation of the BETA intervention, that was set to be tested through a parallel, three-armed pilot RCT (R21HL 156132-01) in a high-volume clinical practice (Cedars-Sinai Medical Center (CSMC)) in Los Angeles, California. The BETA intervention provided habit formation support through three mechanisms: (1) linking medication-taking to an already existing daily routine; (2) increasing information salience through daily text messages, and (3) providing intermittent rewards for medication-taking according to the anchoring plan, that is, coinciding with the time of an existing routine behaviour such as eating breakfast, to sustain motivation.³⁵ All patients received anchoring education through a study leaflet and selected an anchoring behaviour, along with its time of day, place and visual reminder. Patients were randomly assigned to one of three groups: control, receiving usual care (receiving no daily text messages or conditional incentives); messages, receiving daily text messages; and incentives, receiving messages as well as cash incentives conditional on timely adherence.³⁵ The mixed-methods RCT used both in-depth interviews and quantitative survey data to establish feasibility, acceptability, and preliminary efficacy.

Study design

This study involved semistructured interviews conducted with patients and providers to refine BETA intervention parameters prior to RCT implementation. The interviews sought to understand anchoring strategies, attitudes to text messages and incentives, as well as perceived barriers to and facilitators of context-dependent medication-taking. They also elicited information on patient context, preferences and values. The provider interviews elicited perceptions on the intervention’s ability to promote medication-taking habits based on their experiences with this patient population.

Sample selection and recruitment

Patients were eligible to participate if they were 18 years or older, had received a hypertension diagnosis or reported high blood pressure (>140/90 mm Hg) during their last visit, and had been prescribed at least one AHT medication at the time of enrolment. The study coordinator screened for potential patients using CSMC’s electronic health record system. On the day of the scheduled clinic visit, the patient’s provider introduced the study and screened out ineligible patients. If eligible, the study coordinator met with the patient at the end of their clinic visit with study information, consented the patients, and scheduled a Zoom appointment with the second author, AP. Patients were compensated US$30 for their time. A couple of these patients participated in the subsequent BETA pilot RCT.

Providers were recruited based on their experience working with our target population and were not offered any monetary compensation, as is customary. Figure 1 presents the recruitment procedures and final patient numbers.

Figure 1. Patient recruitment procedures and final numbers.

Data collection

All interviews were conducted remotely using Zoom between September 2021 and February 2022. The team developed two interview protocols (one for patients and one for providers; onlinesupplemental files 1 2) that explored contextual barriers and facilitators of AHT-taking habit formation and feedback on intervention components. For contextual factors, the questions covered topics including, existing medication-taking behaviours, facilitators of habit formation for adherence, barriers to habit formation for adherence, and adherence during the COVID-19 pandemic. The intervention development questions sought input on the habit leaflet (online supplemental file 3), three text messages (text 1: ‘don’t forget to take your pills with your meal.’; text 2: ‘this is BETA. Don’t forget me after your meal’; text 3: ‘use your visual reminder to help take your pills with your meal.’), attitudes towards and perceptions of rewards, and integration of the intervention into the clinical workflow (providers only). Each of the protocols was designed to last 30–45 min. The interviews were conducted in English only by the second author, a trained qualitative researcher with experience in studying habit formation. All interviews were audio recorded and transcribed verbatim.

Patient and public involvement

This study did not include any patient or public stakeholder advisory board. Given the formative nature of this study, we instead included patient and provider voices through in-depth interviews.

Data analysis

We used both inductive and deductive approaches to analyse the data.³⁶,³⁸ The deductive code categories were derived from domains in the interview protocol, whereas the inductive codes captured emerging subthemes grounded in the data, for instance, positive or negative sentiments, or types of anchoring routines. Two qualitative researchers from the team (AP and IG) developed the codebook iteratively by reading several transcripts. Once the codebook had been sufficiently developed across the main categories of interest, they coded content independently and engaged in three rounds of discussion and reconciliation of codes, which resulted in merging of some codes and refining of others. Kappa metrics ranged from 0.71 in the early stages of reconciliation, to 0.84 when measured across four co-coded transcripts. All analyses were conducted in NVivo.³⁹

Both qualitative researchers (AP and IG) are trained in qualitative methods in behavioural science and health services research. One of them (AP) has extensive experience interviewing patients and analysing qualitative data in the context of patient-centredness and health disparities. The other (IG) has experience with mhealth interventions in other patient populations and countries.

Results

Table 1 shows the summary statistics of study patients (n=20) and providers (n=7). Sixty-percent of our patients were female, and just under half were over 65 years of age. Most patients (75%) identified as white (non-Hispanic) and 35% disclosed a range of diagnoses linked to hypertension, such as cardiomyopathy and tachycardia.

Table 1. Summary of patient characteristics.

	Patients		Providers
	N	Mean (SD) or %	N	Mean (SD) or %
Total	20		7
Gender
Female	12	60	4	57.14
Male	8	40	3	42.86
Age (years)
18–44	5	25	5	71.43
45–64	6	30	1	14.29
≥65	9	45	1	14.29
Race/ethnicity*				Not collected
White (non-Hispanic)	15	75
White (Hispanic)	2	10
Black/African American	2	10
Asian	1	5
Treatment Timeline*				Not applicable
Less than a year	2	11.76
1–4 years 11 months	6	35.29
5–9 years 11 months	4	23.53
10 years or greater	5	29.41
Prescribed medications (#)*	20	7.4 (3.73)		Not applicable
Antihypertensive therapy (#)	20	2.8 (0.95)
Regimen for anti-hypertensive therapy*				Not applicable
1×per day	5	25
2 ×per day	1	5
3×per day	1	5
Mixed regimen†	13	65
Half-a-tablet regimens*	7	35		Not applicable
Professional role‡		Not applicable
Cardiologists			5	71.43
Registered nurse			1	14.29
Clinical pharmacist			1	14.29
Years in practice‡		Not applicable
<10 years			1	14.29
10–20 years			4	57.14
20–30 years			1	14.29
>30 years			1	14.29

‘#’ indicates that the variable is a count variable (e.g., number of prescribed medications)

^*tThese variables were only collected for participating patients.

^†tThese were patients taking multiple medications with varying dose amounts and dose frequencies.

^‡tThese variables were only collected for participating providers.

Patients differed in how long they had been on AHT: about 12% started less than a year ago, about 35% had been on the AHT medications for <5 years, and a quarter had been on AHTs for more than a decade. Patients were taking about 3 AHTs (SD=0.95), as is typical,^{40 41} and about 7 (SD=3.73) different medications in total, on average. Sixty-five-percent of patients had a mixed regimen, that is, the medications they were prescribed had a variety of intake frequencies. Thirty-five-percent patients were on half-a-dose regimens, that is, were expected to break their pill in half to be taken at different times of day or on different days.

Participating providers were generally balanced across gender, and most were younger than 45 years of age. Most (over 71%) were cardiologists at Cedars-Sinai and had over a decade of experience.

Below, we present findings in detail for each of the emerging themes. Tables2 3 provide supporting quotes from patients and table 4 provides the same from providers.

Table 2. Selected quotes from patient interviews on the barriers and facilitators of medication-taking habits.

Theme/code	Quotes (patient)
Patientprofiles
History of changes to anti-hypertension therapy regimen (n=18; 90%)
Dose regulation	‘Well, [the doctors] changed [prescriptions] because something recently has just started happening to me is, well, used to be my basic problem was my blood pressure was continually rising. We have a new [problem] now because my blood pressure just started dropping and falling to the floor.’ (Pat16).
Concurrent treatment for other chronic conditions (n=8; 40%)
Treatment prior to AHTs	‘I think my cholesterol was a little high and I started taking cholesterol-lowering meds before I started to take blood pressure meds.’ (Pat06)
Polypharmacy	‘I had cancer twice, and I also have osteopenia. So they wanted me to take something for my bones, and then after I had cancer there’s a medicine, Evista, that actually is supposed to help your bones but also is somewhat of an anticancer drug, so I’ve been taking that for probably about ten years. And then allergy meds.’ (Pat21)
Attitudes to medication (n=11; 55%)
Necessity	‘This is a daily part of my life. This will be a daily ritual for the rest of my life. I mean, this is the reality of it. So, I don’t try to hesitate. I don’t try to sneak around the corner. I don’t try to say I don’t want to do it today. I must take my meds until further notice where [the cardiologist] can tweak it.’ (Pat02)
Reluctance	‘[I]deally, I’d like to be off medicine all together, if I can. But I don’t want to if this is helping. And there’s a reason why my blood pressure’s perfect, obviously, because I’m not doing anything else differently, … it’s obviously the medicine.’(Pat25)
Nervousness	‘I’ve only had an allergic reaction to one [medication] in my lifetime. But I’m very sensitive. So the minute I get even a normal dose of something I can feel it, or just my body, it makes me feel like a little on edge, I’m very nervous about new medications, especially heart medications are very serious medications that could potentially have a lot worse of a side effect. So I’m just probably more nervous and cautious.’ (Pat05)
Overwhelm	‘I forget to take my blood pressure pills sometimes because I’m on so many frickin medications that I get tired of it and then that creates like a no-no sign because I’m playing with my health but I get so tired of taking all these medicines.’ (Pat26)
Barriers to habit formation
Psychological and cognitive barriers (n=8; 40%)
Forgetfulness due to unstable anchoring	‘[(Y)ou know that I do have an exact routine when I get up, and [that] I do every morning. But I don’t know why the pill taking don’t snap in there… I don’t usually eat breakfast. But when I do eat breakfast, I automatically remember to take my pills, but if I don’t eat breakfast, I don’t think about my pills until, you know, it’ll pop up in my head sometime or other. And sometime[s], I’ll say, okay, I’m going to get them in a minute and then I forget.’ (Pat22)
Forgetfulness due to routine disruptions	‘The only time I really miss [my medication is] let’s say, [when] I go out for the night and I come back late, and I might forget to take them because I just go right to bed.’ (Pat35)
Forgetfulness due to novelty	‘When I first started taking the medications I would forget. I would, you know, instead of taking three metoprolol a day, I would only take one, or sometimes two, and I would forget the evening pill. Or sometimes I would forget the lunch pill.’ (Pat42)
Forgetfulness due to perceived need	‘I think it’s because I’ve been taking [the AHT medications for] so long I feel like, oh, if I skip a day, whatever. But say if it is my birth control, I know I’m like, oh, no, I have to take that. … So, for some reason I don’t feel the need for my hypertension meds.’ (Pat14)
Hypertension medication side effects (n=6; 30%)
Self-dosing to accommodate side effects	‘The prescription initially was to take [the medication] once in the morning and once in the evening. I was a very nervous patient when I was taking it twice a day, especially with what I’d experienced with the metoprolol, so when I took it was during the tachycardia event.’ (Pat05)
Routine changes to accommodate side effects	‘I tried that one time when I went back to my desk job at my high-rise building as concierge. We decided to wait until I get home to take [the medications] because it just, it makes you tired, it makes you sleepy, a little drowsy, and driving the car from where I’m coming from to get home, I can’t take a chance like that because you lose your balance, and driving at night, I do have my eyeglasses on, but they get you tired. They get you very, very, very tired.’ (Pat02).
Side effects due to missed dose	‘So at the beginning when I started these, you know, a new habit for me, I would forget. So, then I said to myself I’m not taking them the right way so sometimes I would get a headache, or I would feel a little dizzy or, I would [have an] upset stomach,[and] I would think, okay, I forgot my medication.’ (Pat42)
Routine disruptions (n=11; 55%)
Travel	‘(W)hen we’re on vacation, I’m not in a routine. Sometimes, when we go camping, that’s a little harder because things are everywhere. When we’re on vacation or if we’re at my in-laws, that’s easy to put things off [especially] if you’re in a bit of a mess.’ (Pat25).
Unstable routines due to rush	‘(S)ometimes in the morning I get bombarded even before I wake up. I have clients and stuff like that, work, and everything. And so, I will forget to [take the] vitamins that I wanna [take] in the morning. And sometimes I’ll never [take] them.’ (Pat35)
Work	‘Well I am semi-retired now, so I don’t have work issues that get in the way and are distractions. In the past, it was much more helter-skelter and more of a hit and miss, although I understood the importance of blood pressure medication. I was pretty conscientious about taking it, but I did start missing some [doses].’ (Pat24).
Facilitators of habit formation
Stable routines for anchoring medication-taking activity (n=20; 100%)
Routinisation	‘When I wake up in the morning, the first thing I do is set off my alarm, and I go straight [to the kitchen] and I take my medication. I leave [them] on my counter at night. And that’s the first thing I take in the morning. And when I go to work then before I have my lunch which is at 12:00, I go and take my medication. And then when I get back home in the evening, I leave my medications there how I left them in the morning, on the counter, so I remember to take them again at 6 p.m.’ (Pat42).
Anchoring to pre-existing medication routine	‘Well, I take my vitamins every day so I take the [AHT] pills at the same time and I’ve been doing that for about 55 years or so as long as I’ve been taking a vitamin or a pill.’ (Pat32).
Visual reminders	‘I put [the medications] in a little [pillbox] which has the day on it, and I have that on the sink in the bathroom. So they’re staring me in the face’ (Pat06)
Reminders from pets	‘(I)just go brush my teeth, wash my face, get my pills, get a snack for the dogs, and come to bed. That’s, basically, it. And I’m gonna tell you, if I forget the snack, they let me know, so.’ (Pat35)
Motivations for habit building (n=17; 85%)
Personal experiences	‘Having to be rushed to the ER is really a scary experience. So I’ve always exercised but I wanted to do everything I can to prevent that from happening again.’ (Pat10)
Other’s experiences	‘I will tell you that I have a friend of mine lose her partner because her partner did not take her meds. And I kinda always remember that she would miss them all the time, like she would always have to be [reminded], and it caused her to have a heart attack. And she was older than me. So, that always stuck in my head with the blood pressure stuff.’ (Pat35).
Planning (n=20; 100%)
Organising; tracking	‘I load the [pill] box every Sunday night; it’s that routine, that ritual, of getting it done because I put all of my vitamins in there. I put everything else in there.(F)or my arthritis I also put in one Tylenol in case I need it for the day. It’s just right there and ready for me. I know if I’ve taken it or not, then. Because if next week comes by and it’s still sitting in there then I know I didn’t need it that day.’ (Pat39).
Managing changes to visual characteristics of the medication	‘So I had to make the pill thing as simple as possible. That’s why I put them in the boxes. And I put them all at the times that I had to take them because what I found out is, each month the pharmacy changes to the pill maker that is the cheapest for them. So the way they look changes. …(Y)ou can’t depend on the pills always to look different [from each other] in appearance because sometimes small pills look exactly the same as another med. So a senior citizen might be taking two prednisone or two blood pressure meds [instead of two different medications].’ (Pat16).
Failure of single strategy	‘[I store my medications] in a little pill thing on the nightstand next to my bed… And that still don’t work sometime(s).’ (Pat22).
Reminders; combination strategy	‘I have alarms set on my cellphone for 8:00 a.m. and 8:00 p.m., to remind me to take meds. But, basically, [when at home], I keep the pillbox on the counter as you enter the kitchen. So, I wake up in the morning, and I go to the kitchen to make coffee or have breakfast, it’s right there, and that helps to remind me to take it. And then each week I put not just my prescriptions, but other things some vitamins and supplements, low-dose aspirin, magnesium, and so on, so this is ready to go each day. And between the alarms on my cellphone and pillbox, I never forget to take my blood pressure meds or the other things.’ (Pat27)

Table 3. Selected quotes from patient interviews on the perceived effects of COVID-19 on habits and general feedback on the intervention.

Theme/code	Quotes (patient)
Perceived effects of COVID-19 on habit formation (n=16; 80%)
No impact	‘At the beginning, at the very beginning, it was kind of scary not wanting to go into the pharmacy kind of a thing. But no, there hasn’t been any problems.’ (Pat10)
Negative impact	‘I try and take all the medication same place, same time, and I found that the routine in my life was actually going to an office to work. So I would go into the office. So I’d get ready whatever in the morning, and actually go into the office, sit at my desk, and that was the first thing I did. And then we had COVID. And I’m still at home so I had to change that and it’s messing me up. I think two days ago I just forgot all my meds. You know? It’s not consistent.’ (Pat14).
Positive impact	‘Well, one thing that worked really well, we had this thing called a pandemic, and I was never allowed to leave my apartment. So sticking to a calendar pretty much kept me on a schedule.… Very few, very, very few interruptions. In other words, I didn’t have anything else to do.’ (Pat16)
Intervention feedback
Habit leaflet (n=14; 70%)
Content	‘I think it looks good. It’s clear. It’s well laid out. It seems sensible. My experience is pretty consistent with that. It mentioned time, place, and activities as kind of anchors or parts of the habit. And it’s true. Going into the bathroom and seeing my toothbrush reminds me about, you know, did I brush my teeth yet?’ (Pat27).
Layout	‘(Maybe) some of the message or some of [the content] could be represented in a visual manner because a patient may or may not want to read the whole thing. With the visuals it’s easier to remember, it’s easier to understand, perhaps, it’s quicker. ‘ (Pat10).
Text messages (n=19; 95%)
Function (n=16; 80%)
Helpful reminder	‘I think that that’s a great idea. I think, I, probably, would only need the one reminder. And then, I’d just leave it up on my phone. I think one would be super helpful and I’d appreciate it.’ (Pat25)
Selective impact	‘Most people have smartphones but not everyone. A lot of people who have to take blood pressure meds are older and may be not tech savvy.’ (Pat27).
Cadence (n=11; 55%)
Lower frequency	‘You know, I think that might be helpful. I think the only downside would be people might find it annoying. But, once the habit was established, if they could opt out [of the messages], that might be helpful.’ (Pat21)
Time of anchor	‘My point is if I needed to take this, and I was going to get a text, what I would do would be work out a way that the texts come to me at those times every day. And they wouldn’t have to be very long. It could say, Citrucel time. That’s a great idea. I think the text idea is a great idea.’ (Pat06)
Random time	‘I kind of like the idea of random times, informing people of the principles rather than, oh, it’s 8:00 a.m., take your meds.’ (Pat27)
Content (n=13; 65%)
Informational	‘If you were to send that ‘don’t forget to use your visual reminder [message]’ every day, then over time, it might motivate the person to get a visual reminder that really works for them. You know? They would say, well, that visual reminder I did yesterday was just useless because I totally forgot to take my pills. So I think I’ll try something different today.’ (Pat06).
Tone	‘I think any of them are okay, but a lot of people who take these medicines are older, and so you know, a little friendliness probably might engage them more than some of the other things.’ (Pat21).
Word selection	‘I think for me seeing the word pills doesn’t make it seem important. When you say medication, it’s like, oh, my prescribed medication not my vitamin pills or my fish oil…’ (Pat14).
Incentives (n=20; 100%)
Perceived role of rewards	‘(T)he reward might encourage people. They may not do it for the money, but getting that reward is like a sign of reinforcement’ (Pat27).
Negative impact	‘(w)ell, what happens when the medication is giving you a side effect or it’s not working properly and then you’re still just taking it because you want that incentive? Because it could potentially become ‘I’m going to start taking this because I like that incentive.’ I don’t like that.’ (Pat05)
Perceived need	‘It’s nice to receive a reward. Of course, it’s nice. But I think that you shouldn’t get a reward for you to be reminded to take your medications. I think that as a patient we should remember to take our medications. I know that a lot of people forget and don’t have the capacity to remember to take the medications, but I don’t think that it should be a reward.’ (Pat42)

Table 4. Selected quotes from provider interviews.

Theme/code	Quotes (provider)
Barriers to habit formation
Psychological and cognitive barriers (n=6, 86%)
Lack of understanding	‘the older folks especially, we try to explain to them the natural history of hypertension, that many patients get worse over time.’ (Pro2)
Nature of condition	‘[Hypertension is] an illness that’s not apparent to[the patients]. They don’t feel poorly. You don’t see anything physically that’s wrong with them and they think they’re fine. And so I’ve had to deal with that throughout my career.’ (Pro3).
Hypertension medication side effects (n=5, 71%)
Fear of side effects	‘I think side effects are the primary reason people don’t take medication. And maybe, even if they haven’t experienced a side effect, it’s the fear of a potential side effect. Something that their cousin, or their mother, or somebody has told them that caused them a problem and they’re fearful that it will cause them an issue.’ (Pro4)
Other patient contextual factors (n=6, 86%)
Work	‘I had lots of patients where they might be doing a really intensive, like their job is really intensive and they’re not at home for many hours of the day so if it were a twice a day medication they have to bring the bottle with them to work and it’s like it’s easy to forget, it’s kind of embarrassing to take medications.’ (Pro5)
Structural barriers	‘(A) lot of times, and especially now in the last years, when we’ve got new medications and a lot of the meds aren’t on the formularies for patients that have, you know, insurance coverage. And so a doctor will prescribe telmisartan. And it’s not on their, you know, it’s cost prohibitive to them. It’s not on coverage. And they don’t call. They’re not proactive to say, by the way, can you order something different?’ (Pro3)
Language and literacy	‘Things that make a really big difference actually would be the language concordance between the provider and the patient. It’s much easier for me to explain something in English than in, for example, like French or Spanish or something which at my previous clinic in the county hospital a lot of my patients were Spanish speaking and obviously use an interpreter but it’s like 10 times harder to even go through the process of why hypertension treatment is important and you know. I guess baseline medical knowledge is really important.’ (Pro5)
Forgetfulness	‘Mostly it’s just you just can’t remember it.’ (Pro5)
Facilitators of habit formation
Stable routines for anchoring medication-taking activity (n=7, 100%)
Patient counselling to support routinisation	‘I do try to minimize the number of pills the patients are on to help them achieve that and when I discuss with them taking the pills I always tell them that if it’s gonna be a daily or twice a day dose and I ask them if it’s gonna be a problem for them and I usually ask them what are their habits, when do they take their pills, do they take it with their supper or do they take it when they wake or with breakfast and I try to set up a time for them like when would it be the best time to take it especially if they have other pills that they’re taking and I tell them you could take this pill with this. So, I kind of give them a bit of a guideline of how to take it and I find that that works.’ (Pro6).
Visual reminders	‘I usually tell folks who are gonna take it in the evening, to leave it on their nightstands or in their nightstands, so it’s right there, looking them in the eye before they go to bed. And most people bring a glass of water with them anyways to bed. So, you’ve got your pill bottle there, and now water. So, yeah, I think that all of that looks good.’ (Pro4)
Anchoring	‘I think if they incorporate it into one of their daily activities, activities of daily living, that we all do, take a shower, brush our teeth, wash our hair, whatever it is, or waking up in the morning and going to bed at night, those, to me, are three instances, three touchpoints, where you can create that habit. You go to bed every night.’ (Pro3)
Intervention feedback
Habit leaflet (n=7, 100%)
Layout	‘It’s always nice to have images. I mean that was a lot of text and maybe that was why I didn’t read it when I first saw it. It’s a lot of text and so if you kind of broke it up into some like home, or maybe some pictures of like a bathroom cabinet or food or something like that.’ (Pro7)
Text messages (n=7, 100%)
Function (n=4, 57%)
Selective impact	‘I think if it’s the younger population they’re so used to getting text messages and notifications that it would be a good cue. I wonder with elderly people if that would work as well because they generally don’t have their phone or don’t use their phone as often.’ (Pro 06)
Messaging fatigue	‘If it’s somebody who texts all day long, then it may not even be looked at, or looked at late, or not at all.’ (Pro3).
Cadence (n=3, 43%)
Lower frequency	‘It would probably be slightly torturous, I think. And people might get a little ticked off by that. If you tie it to some positive message or little information every day, it might be less annoying. Maybe sometimes you wanna send them a little information about….’Did you know there’s one 1.8 billion hypertensives in this world? Congratulations for taking your medications today,’ so it doesn’t get too monotone. It would drive me crazy, I think, to get the same text, maybe even the same time, but it sounds, like, it’s not gonna be the same time.’ (Pro1).
Content (n=3, 43%)
Informational	‘I think that text one (see Text 1) is more concrete, and more specific, and tells you what to do. I think those people might have better success in remembering to keep your pills near where you eat, but maybe that’s the point.’ (Pro3)
Rotating content	‘It would probably be slightly torturous, I think. And people might get a little ticked off by that. If you tie it to some positive message or little information every day, it might be less annoying. Maybe sometimes you wanna send them a little information about….’Did you know there’s one 1.8 billion hypertensives in this world? Congratulations for taking your medications today,’ so it doesn’t get too monotone. It would drive me crazy, I think, to get the same text, maybe even the same time, but it sounds, like, it’s not gonna be the same time.’ (Pro1).
Incentives (n=7, 100%)
Useful	‘I think rewards always improve habits. That’s been proven in other types of studies as well.’ (Pro7)
Uncertain about usefulness	‘I don’t know how effective the incentives will be, to be perfectly honest. In my past studies, I’ve, we’ve used incentives to try to get people to show up for their appointments. And people like incentives, but I don’t know if it makes them better at, better patients, to be perfectly honest, you know.’ (Pro4)
Acceptability and feasibility
Clinic feasibility parameters (n=7, 100%)
Room use	‘We don’t want our rooms to be taken up by research because there really is a flow to it so mainly that the consent could be done outside of the clinic space.’ (Pro7)

Patient profiles

Patient profiles varied greatly based on their condition and timeline of diagnosis.

History of changes to AHT regimen

Most patients (n=18; 90%) reported adjustments to their regimen over time, including changes in dosage and medication type, typically due to changing health condition of the patient (n=17; 85%). Medication intolerance and sensitivity (n=4; 20%), and side effects (n=3; 15%) were also cited by some. Only one patient reported regimen changes due to unemployment and consequent loss of health insurance.

Concurrent treatment for other chronic conditions

Almost half of patients indicated taking medications for other conditions concurrently with their AHTs (n=8; 40%). These ranged from multivitamins to medications for hyperthyroidism, cholesterol, diabetes, asthma, and cancer. Some noted that routines for other medication predated the antihypertension regimen while others indicated otherwise.

Attitudes to medication

Patients generally recognised the importance of taking their AHT medications. More than half (n=11; 55%) labelled their medications as ‘essential’ for their health. Despite that, patients shared a range of feelings about medications. Many (n=8; 40%) showed acceptance of their health status and discussed continuing to take their medication even when sick. Others expressed their dislike for medications but suggested they would take them, nonetheless. Some patients (n=2; 10%) described how they felt burdened because they were prescribed too many medications. One similarly shared how past medication sensitivity-triggered reactions left them in anticipation of possible side effects and made them feel nervous about taking AHTs as prescribed.

Barriers to habit formation

Psychological and cognitive barriers

About one-third of our patients (n=6; 30%) cited forgetfulness as a major barrier to habit formation. Supporting quotes in table 2 show that many statements about forgetfulness were not necessarily related to cognitive decline, but rather disruptions to routines (such as travel) or work, or unstable anchoring. Two providers spoke about this issue among their ageing patients. Most patients (n=4; 20%) described forgetfulness in the context of other barriers, including feeling a sense of ‘medication fatigue’, and distractions such as online work calls or night outs. In contrast, some patients (n=2; 10%) described how forgetfulness was a result of tying medication-taking to a cue that does not happen every day, or due to the novelty of their prescription.

Some patients (n=3; 15%) questioned the need for continuing adherence, such as when they experienced a change in their lifestyle, or when they weighed the side effects more strongly than the perceived benefits.

Most providers (n=6, 86%) noted that while some patients had a general aversion to medication, in many cases the issue is a lack of understanding of hypertension and the role of medication for this condition. Moreover, they underscored that patient attitudes to antihypertensive medication are often driven by the asymptomatic nature of the condition.

Hypertension medication side effects

Several patients (n=6; 30%) noted side effects with their hypertension medication but indicated that they found ways to cope, such as changing routines to accommodate the side effects or changing dosage to a more comfortable dose (self-dosing), as illustrated in the quotes in table 3. Some patients (n=2; 10%) described experiencing side effects upon missing doses, and others mentioned awareness of potential side effects but did not consider it a barrier to habit formation (n=2; 10%). A quarter of the patients did not report any side effects.

Most providers (n=5, 71%) spoke about side effects as a dominant factor for adherence as prescribed, both those related to antihypertensive medication (such as fatigue, erectile dysfunction), as well as related to medication for other conditions that patients wrongfully ascribe to their hypertension regimen.

Routine disruptions

Travel was one of the most frequently cited disruptors of routines (n=4; 20%), as was work, mentioned by both patients (n=3; 15%) and providers (n=2; 29%). Patients described how the morning rush to the office often resulted in unstable routines. Other disruptors included participation in social activities, changes in weekend schedules and caregiver responsibilities. Importantly, however, patients reported most of these disruptions to be rare while labelling them a disruptor of timeliness of adherence rather than adherence in general.

Facilitators of habit formation

Stable routines for anchoring medication-taking activity

Patients provided a vivid account of the cues, the times of day, as well as the loci and place of medication-taking to describe their daily medication routines. Several (n=8; 40%) described a sequence of activities to demonstrate how medication-taking seamlessly integrated into their routines.

Many patients (n=10; 50%) considered the use of stable anchors, or cues that were least likely to be disrupted. Commonly described anchors included getting out of bed, morning coffee or breakfast, lunch, dinner, getting ready for bed and pre-existing medication taking routines. Importantly, a few patients (n=3; 15%) described cues that triggered their caregivers or pets to remind them to take their medication.

Home was the most frequent place for taking medications; the locus centred around the kitchen, the bed or the bathroom, where the patient carried out other activities related to the anchored routine. Keeping the medications in these places also served as a visual reminder. Workplace and car were identified as other areas where patients took their medications, especially in instances where regimen included an afternoon dose, or when the patients expected to be away from their primary locus.

Most providers (n=5, 71%) felt that anchoring medication taking to existing routines was the strongest facilitator of habit formation. In their narratives, this went hand in hand with patient counselling based on each patient’s context (n=6, 86%) and use of visual reminders (n=3, 43%) as a strategy to guide them towards a habit of taking their pills as prescribed.

Motivations for habit building

Patients described several sources of motivation to maintain high medication adherence, including trust in the physician, knowledge of condition, support from caregivers, prior experiences with long-term regimen adherence, and strong perceived benefits of medications. Importantly, the motivation to ensure high adherence seemed to have resulted in the formation of medication-taking habits.

Fear, however, was the most prominent source of motivation for more than half of the patients. For many (n=7; 35%), fear was a consequence of prior experiences that increased visibility of their condition. Other patients (n=3; 15%) felt motivated to take their medication after encountering or hearing about others’ experiences (including family or friends). Some patients (n=3; 15%) also reflected on their personality types without explicitly stating specific motivators.

Planning

One of the most frequently cited facilitators of medication-taking habit formation was planning, with the following strategies to support routinisation: identifying convenience factors (such as keeping water near the medications, placing medications in more than one area, requesting regimen changes to help couple medications with non-AHT regimens, and using a pillbox), setting external reminders, and preplanning for possible disruptions. A few patients (n=3; 15%) even discussed caregiver support as a way to plan and manage medication routines.

Pillbox use was the most frequently cited planning activity (n=17; 85%). Patients used pillboxes for managing a portfolio of medications across conditions, ordering refills and storing extra supplies, tracking adherence and missed doses, storing halves of pills for half-a-pill regimens and organising for travel or other possible routine disruptions. Convenience was a commonly mentioned thread, as the pillbox reduced the general burden of managing medications, especially in cases where the visual characteristics of the medications changed unexpectedly.

Some patients, however, described how using the pillbox alone did not help with forming medication-taking habits. To counter this, many patients (n=5; 25%) employed external reminders in combination with other strategies for continued adherence. The reminders included digital alarms, calendar reminders, and in some instances visual aids (such as pill boxes, medication bottles or medications in a tray) kept strategically near their medication-taking loci.

Perceived effects of COVID-19 on habit formation

Patients had mixed experiences with managing their medication routines during COVID-19. More than half (n=11; 55%) indicated that COVID-19 had no impact as they continued to follow their pre-existing routines. Others (n=4; 20%) experienced disruptions to routines that led to lower adherence or found the COVID-19-related disruption to be helpful in forming a good medication-taking routine.

Intervention feedback

Habit leaflet

Half of the patients (n=10; 50%) described the content of the habit leaflet as simple, easy to understand and relatable. A few patients (n=5; 25%) recommended contextualising the layout by increasing the font size and adding visuals to make it appealing and easy to remember. This was echoed in a few provider comments (n=3, 43%), who recommended visual accommodation for elderly patients by replacing text with cartoons or images.

Text messages

Patients provided a range of responses across three different dimensions of text messaging: function, cadence and content.

Function

Most patients (n=15; 75%) agreed that text message reminders could be helpful for facilitating habit formation, especially for those patients who are prone to forgetfulness. Though, a few cautioned that their use may only be limited to those patients who are tech savvy. Phone fatigue and inability to snooze messages were concerns that were also brought up, though they were only referred to by one patient each.

Cadence

Patients’ reactions were split evenly across two stances when considering the cadence of the text messages. While some patients thought that daily messages would be appropriate (n=4; 20%), others shared concerns about the redundance of these messages and how it could become ‘annoying’ over a long period of time (n=4; 20%). Similarly, while some patients additionally suggested tying the messages to the patient’s anchoring time(n=4; 20%), a few thought that messages should come at random times during the day (n=2; 10%). Half of the providers highlighted similar concerns regarding the use of text messages and phones among older demographics.

Content

Patients used words such as ‘simple’, ‘clear’, ‘helpful’ and ‘engaging’ to describe the content of the text message. A few patients (n=3; 15%) thought them to be educational, with potential for helping patients find strategies for routinisation. Some others recommended that the text content be made more friendly to appeal to the older audience (n=4; 20%). A few additionally suggested keeping the sender and content formal to project credibility (n=2; 10%).

Providers were mixed in their reactions. Most simply said they liked the messages. One provider was concerned about patients receiving the same text every day, suggesting instead that messages be embedded within pieces of information that generate interest among recipients.

Perceived role of rewards

Patients provided a range of perspectives on the function and design of the incentives. Most patients (n=16; 80%) considered conditional incentives as useful motivators for medication-taking habit formation. A few patients (n=3; 15%) thought that the incentives idea was ‘genius’ as it ‘is an example of how to get people to do the humane thing’. (Pat26). Some (n=4; 20%) thought this would make the otherwise mundane task of medication-taking fun and enjoyable. One patient even insisted on the rewards being introduced earlier in the study period as ‘[t]hirty days is too long. [It should be t]en days.’ (Pat23).

In contrast, some patients (n=4; 20%) expressed scepticism around the effectiveness of incentives for those who are self-motivated or have a routine established already. A couple of patients (n=2; 10%) described possible heterogeneity in effects, as their behaviour might ‘depend on what I’m winning’ (Pat22). A few, however, disliked the idea of rewards altogether, either because they considered medication-taking to be a self-motivated behaviour, or because they thought incentives could mis-incentivise those who struggle with adverse side effects.

Providers offered mixed perspectives on incentives as well. Some (n=4, 57%) welcomed it as an evidence-based approach to increasing pill-taking compliance. Others (n=3, 43%) were not certain about the effectiveness of the small incentives, citing both prior work in this area and personal experience. One provider emphasised the discrepancy between patients saying they like incentives and actual behavioural change.

Acceptability and feasibility

Providers agreed on the importance of helping clients build AHT adherence habits, and on the need to harness technological advancements to help patients sustain healthy medication-taking habits. There was also agreement on the proposed intervention’s acceptability and feasibility in the context of the CSMC system and the patient population it serves. A key recommendation was to ensure that any efforts to recruit into the study should be coordinated in advance with care teams by a research coordinator dedicated to the study. Most warned against burdening care staff with screening and consenting to the study and flagged physical space as an issue.

Discussion

This study fills an important gap in the literature around habit formation for long-term medication adherence among people with hypertension. The interviews conducted as part of the formative phase of the pilot RCT of the BETA intervention provided important patient and provider perspectives around the barriers and facilitators of habits, along with critical feedback to inform the design of the intervention.

Our findings shed light on some crucial contextual factors at play within our population of interest. First, they illustrate how forgetfulness may not only be a function of the predominantly geriatric population, but also of low perceived need (ie, low salience) for medications arising from the asymptomatic nature of the condition, which could potentially misrepresent the costs associated with non-adherence. This supports the BETA intervention’s focus on daily reminders and conditional rewards to increase salience of and motivation for forming habits. Second, the results illuminate how considerations concerning polypharmacy would be crucial in the intervention design. For example, multiple regimens for comorbidities may increase the anchoring-behaviour choice pool for BETA. They could also induce medication fatigue, however. Frequent regimen changes could additionally increase the chances of selecting unstable anchors, which could effectively undermine our efforts towards habit building.⁴² This, combined with awkward dosing (such as half-pill regimens) could result in additional barriers to habit building. These considerations may require frequent monitoring of regimen changes and anchors to efficiently identify and replace unstable anchors with more stable and appropriate ones.

Insights from our interviews also provide other possible mitigation strategies to overcome such barriers. The first insight considers the importance of motivation in facilitating habit building, which, as with existing habit literature, suggests the need for intrinsic motivation to continue repetitive behaviours until a habit is formed.¹⁹,²² The conditional incentives may help shift patients’ perspectives towards their medication and provide additional motivation to continue repeating the anchored behaviours until a medication-taking habit is formed. Once formed, habits are sustained without the need for intrinsic motivation.^{24 25} The second insight consists of active planning between providers and patients to make medication-taking as convenient as possible, which is supported by prior literature.⁴³ Our findings highlight a range of strategies for planning, including stable anchor and locus selection, pill box use and management to promote convenience, and employment of external reminders (both audio and visual) to overcome forgetfulness. A possible future intervention could also include a checklist for providers to ensure they discuss these strategies with study participants. As such, our work generates a third insight: motivation must be combined with planning to maximise the impact of this intervention. This is further supported by existing work on habit formation^{42 43} and consequently supports the concept underlying the BETA intervention.

Our study, while limited in its focus on the components of the BETA intervention, highlights important considerations for any future mhealth, and behavioural economics-based interventions for long-term adherence, especially in geriatric populations. For example, text message-based interventions should account for cadence, content, and delivery mechanisms that are seen as credible, salient, and useful for continued engagement. Components associated with patient education and motivation should be designed to (a) highlight information that is most pertinent to the audience, (b) include images and graphics with simple language to allow for quick reading, and (c) pay close attention to font size for readability. Additionally, adherence measurement mechanisms must be tweaked to fit what is contextually acceptable and convenient in a polypharmacy setting. In the case of BETA, this could involve replacing the use of the medication event monitoring system caps with adherence measurement devices that mimic a pillbox, which is more commonly used by our population of interest.

Our study is not without its limitations. First, our sample was recruited from one centre in Los Angeles, which primarily caters to a relatively high-income group of hypertensive patients and specialises in complex cases involving medication sensitivity and intolerance. Self-selection may also be a factor, as patients who chose to be interviewed may be different from those who did not. Future studies should, thus, consider comparing barriers and facilitators identified here, with those for other populations and settings. Second, the study was limited in its scope to focus on the BETA intervention, even though it elicited contextual factors associated with medication-taking habit formation more generally. As such, this leaves many areas for future investigation, including the identification of mechanisms involved with the use of visual reminders to support anchoring of medication-taking behaviour to a pre-existing cue, assessment and tracking of disruptions to chosen anchors during the habit formation process, and the possible impact of choice and guidance for anchor selection on habit formation and maintenance. Additionally, future studies on adherence measurements may consider the acceptability of other objective adherence measurement systems in the context of polypharmacy.

Data availability statement

Data are available on reasonable request.