Abstract
Background
Medication adherence is essential for effective atrial fibrillation (AF) management; however, many patients with AF are non-adherent to these medications, necessitating an understanding of adherence barriers. This mixed-methods study aimed to identify factors influencing oral anticoagulant (OAC) adherence.
Methods
A stratified random sample of 2000 patients with non-valvular AF and OAC prescriptions was surveyed from a large, seven state community health system, capturing barriers and facilitators to medication adherence using an adapted version of the Adherence Starts with Knowledge-12 scale. Twenty-eight semi-structured interviews were conducted with a sample of survey respondents to further explore drivers of adherence. Eligible clinicians were surveyed on patient adherence. Survey data were analyzed descriptively and using multivariable regression models. Interview data were analyzed thematically.
Results
Of 213 patients with an adherence score available, 31.5% reported non-adherence. Nonadherent patients were more likely to report poor healthcare experiences. Receiving all necessary information from clinicians was positively associated with medication adherence. Interview analyses showed clinician communication about medications and believing that medication can help manage AF were important facilitators of medication adherence. While clinician response rate was low (3.0%), 87% of responding clinicians perceived high non-adherence among patients with AF, citing forgetfulness, cost, and side effects as the main reasons.
Conclusion
Most patients reported adherence to their medications, whereas clinicians perceived relatively low adherence among their patients. However, clinicians aligned with patients on the need for improved communication and shared decision making around AF treatment. Interventions that promote clinician and patient communication may be pivotal for optimizing medication adherence, which may lead to improved patient care.
Keywords: medication adherence, mixed methods, oral anticoagulants, nonvalvular atrial fibrillation
Background
Atrial fibrillation (AF) affects nearly 3% of the population and increases one’s risk of ischemic stroke five-fold. Oral anticoagulants (OAC) are highly effective at preventing stroke, and therefore are a key component of AF management.1,2 National guidelines recommend the use of OAC in patients with AF at elevated stroke and thromboembolism risk (“Congestive heart failure, Hypertension, Age ≥75 years (2 points), Diabetes mellitus, Stroke or transient ischemic attack history (2 points), Vascular disease, Age 65-74 years, and Sex category (female)” (CHA2DS2-VASc) Score ≥2 in males or ≥3 in females).3–5 However, non-adherence to these medications in patients with AF is common, with an estimated 33% to 50% of high-risk patients with AF being non-adherent to their OACs.6–8 This is alarming, as research has shown even small reductions in OAC adherence in patients with AF are associated with significant increases in stroke risk and poor health outcomes.9
Medication adherence is a complex behavior influenced by patient, provider, and health system factors.10 Given the importance of adherence to prescribed OACs in maximizing treatment effectiveness and improving patient outcomes, it is important to gain a thorough understanding of facilitators and barriers to OAC adherence.1,6 By considering both patient and clinician perspectives to medication adherence, patient centered approaches to managing OACs can be developed to improve adherence among patients with AF. This study used a mixed-methods, exploratory observational design to examine patterns of OAC adherence in patients with AF, including patient and clinician perspectives of drivers of these patterns and mechanisms for improvement.
Methods
Patient Survey
A retrospective cohort analysis of electronic health record (EHR) data from patients in the Providence Saint Joseph Health (PSJH) system was conducted. PSJH comprises 52 acute care hospitals and over 800 ambulatory care clinics distributed across a 7-state health system in the western US and Alaska. Patients ≥18 years of age diagnosed with non-valvular AF and a prescribed OAC medication noted in their EHR were sampled for the survey. Additionally to be included, patients had to have an outpatient encounter in their EHR between January 1, 2016 and June 1, 2023 with at least two A-Fib and/or atrial flutter diagnoses at seven or greater days apart (first instance could be from inpatient, outpatient, or emergency department visit; second instance must have been outpatient visit), have a virtual or face-to-face visit with a Providence clinician in the past two years, have a prescription for an OAC in the past year, speak English as their primary or secondary preferred language, and have a valid mailing address in their EHR. This study complies with the Declaration of Helsinki and was approved by the PSJH institutional review board (Study Number STUDY2023000907).
Using stratified random sampling, surveys were sent to 2000 patients from the overall eligible population. Strata were created using race/ethnicity (non-White/White), insurance type (Medicaid/other), and residence (based on the USDA, Economic Service’s Rural-Urban Commuting Area codes (RUCA) assigned to patient zip code (urban: RUCA <4, rural: RUCA ≥4),11 resulting in eight mutually exclusive groups. While simple random sampling would create groups constituting 12.5% of the survey sample, some groups were underrepresented. In these instances, all members of the specific underrepresented group were surveyed with additional patients recruited from demographics that remained underrepresented. For example, there were few non-White, rural patients in the eligible population. Therefore, the entire group of non-White, rural patients was sampled, and additional non-White urban patients and White, rural patients were sampled to complete the strata.
Participants received the survey, informed consent document, and an introductory letter in the mail which provided an option to send the survey back via mail or complete the survey online via REDCap if preferred. The primary outcome of interest, self-reported medication adherence, was captured using an adapted version of the Adherence Starts with Knowledge-12 (ASK-12), a validated 12-item scale assessing barriers and facilitators of adherence that consists of three subscales: behavior, treatment beliefs, and inconvenience/forgetfulness.12,13 Each subscale score was computed as the sum of all items within that scale, and the total score was the sum of all 12 items. All items used 5-point Likert scale response options. The total score range was 12–60, with higher scores indicating greater barriers to medication adherence. Scores were also dichotomized using a score of ≥ 23 to indicate non-adherence, and scores under 23 to indicate adherence, as described previously.12,14 Prior approval was obtained for use of the ASK-12 (ASK-12 contact information and permission to use: Mapi Research Trust, Lyon, France, https://eprovide.mapi-trust.org).
In an attempt to better understand the drivers of medication status and medication adherence, additional survey questions were added that inquired about the patient’s health status, health-related self-efficacy (self-efficacy), healthcare experience,15,16 healthcare discrimination, health care insurance and access to care, and other social determinants that impact their health.17,18 The full patient survey can be found in appendix A. Briefly, health status was assessed by asking patients to rate their physical and mental health. Health-related self-efficacy was assessed based on responses to questions inquiring about a patient’s feeling of control over decisions that affect their health, understanding of what to do when they have a health problem, understanding of where to go when they have a health problem, and comfort advocating for their health. Healthcare experience was assessed by inquiring about the care received from the patient’s healthcare provider. Healthcare discrimination was assessed based on responses related to common ways patients may feel discriminated against by a member of their healthcare team, because of their race/ethnicity, language, gender or gender identity, sexual orientation, income, disability status or health needs. Added questions related to healthcare insurance and access to care assessed patients’ access to healthcare services. Finally, a patient’s social determinants of health were assessed by inquiring about other experiences that may impact their health such as having social/economic needs. Participants who completed the survey during the study window received $10 cash by mail. Data collection occurred between January 5, 2024 to March 8, 2024.
Patient Interview
Following survey completion, a member of the research team contacted survey participants who completed the patient survey and opted-in to being contacted for a follow-up interview, self-reported an AF diagnosis, and self-reported being prescribed a medication for AF. Recruitment also included interviewees classified as adherent or non‑adherent according to their ASK‑12 scores. Interviews were conducted and analyzed by two authors (MG and RG). Interviewees were recruited until saturation was reached. Data saturation was defined as the point at which additional data did not lead to new information or themes related to drivers of medication adherence and was determined by two interviewers through post-interview assessment review (ie, brief, written reflections completed by interviewers).
The interviews were conducted by phone using a semi-structured interview guide (see appendix B) that focused on understanding patients’ experiences with healthcare, the ways in which they communicate with their clinicians, and their experience accessing and taking their prescribed medications. Interviews lasted approximately 30–60 minutes and were audio recorded and then transcribed by a HIPAA-compliant third-party professional transcription service. The participants provided verbal informed consent for the information shared during their interview to be used for research.
Clinician Survey
A cross-sectional survey was sent to 2690 healthcare providers within PSJH system with experience treating patients with AF between 2021 to 2023 to assess clinician perspectives on barriers and facilitators to medication adherence. Clinicians without a valid email address, and clinicians with specialties that would not actually require them to manage a patient’s AF diagnosis (eg, psychiatry) were excluded from the study. Clinician respondents were presumed independent from the patient survey sample.
The Clinician survey was sent out via email and responses were captured in REDCap. The survey focused on drivers of effective diagnosis and treatment of AF and medication adherence. Clinician perceptions of medication adherence were measured using a modified version of the ASK-12 scale,12,13 where clinicians were asked to think about a typical patient with AF when answering the adherence questions (ASK-12 contact information and permission to use: Mapi Research Trust, Lyon, France, https://eprovide.mapi-trust.org). The survey also asked clinicians about their communication with patients, decision tools used, and managing patients with AF. Appendix C provides a list of all questions asked in the clinician survey. Survey completion denoted consent to participate. Data collection occurred between November 7, 2023 and February 22, 2024.
Data Analysis
Patient Survey
Descriptive statistics were used to summarize the patient survey data. Demographics were summarized using mean ± standard deviation (SD), or median (interquartile range (IQR)). Overall discrimination score was calculated as the mean (SD) of all six healthcare discrimination survey questions. All other survey questions were collapsed into binary metrics (eg, agree vs disagree, always or sometimes vs never). Categorical variables were summarized using frequencies and percentages. A univariate analysis of variance and chi-square goodness of fit test were used when appropriate to test for significant differences between comparison groups (responders vs non-responders and adherent vs non-adherent). To assess the extent to which patient-level characteristics (eg, healthcare experiences, healthcare discrimination, clinician relationship) were associated with ASK-12 medication adherence, Poisson regression with a log-link and robust standard errors were used for binary outcomes (where an ASK-12 score ≥ 23 indicated non-adherence). The risk ratios (RRs) are modeled such that the reference is Adherent (Adherent = 0, Non-Adherent = 1). Multivariable models were adjusted for potentially confounding demographic characteristics including age, sex, self-reported race, insurance type, geography type (rural or urban), and chronic conditions, to identify the most important associations between covariates and adherence. Covariates were included based on significant univariate associations within each section of the survey. The RR and its corresponding 95% confidence interval (CI) were reported as the effect measures. A p-value of less than 0.05 was considered statistically significant.
Subgroup analyses were performed using the same multivariable Poisson regression models identified for the overall sample. These subgroups included age (70 and under, 71 and over), insurance type (Medicaid, non-Medicaid [Medicare, Commercial, or other]), geography type (rural, urban), and number of non-AF chronic conditions (2 or fewer conditions, 3 or more conditions). Results are presented in Supplemental File A (Tables S1–S8).
Patient Interview
The interview coding and analysis was completed by two authors (MG and RG). Interview data were analyzed thematically to identify common topics and patterns that came up repeatedly in the data. An initial codebook was developed from the interview guide, checked for modifications as post-interview assessments were completed (though no modifications were made), and then revised iteratively when emergent codes were identified during coding of transcripts. To maintain methodological rigor, two interview transcripts were double coded. After each double-coded interview, the coding team met to discuss codes until consensus was reached Once coding was completed, codes were summarized and reviewed. Themes were derived from the summaries by identifying repeated similar ideas and commonalities or contradictions. The themes were refined to ensure that patient stories were represented and spoke to the original objective of understanding factors that influence medication adherence. All analyses were completed in Atlas.ti version 24.
Integration of the Quantitative and the Qualitative Patient Data
Integration of the patient interview and patient survey results was implemented in the research design, methods, and reporting/interpretation of study findings.19,20 At the design level, an exploratory sequential design was used.21 The quantitative and qualitative approaches were linked in the participant sampling, as the interview participants were identified from those that completed the survey.20 Finally, an integrated approach was used to report themes that emerged from the qualitative data analysis corresponding to the quantitative survey results.22
Clinician Survey
Clinician survey data were analyzed descriptively. The modified version of the ASK-12 was scored in the same way as the patient survey, with an ASK-12 score ≥ 23 indicating non-adherence and <23 indicating adherence. Survey items related to patient communication, AF management, and decision tools were analyzed as binary metrics with responses being categorized as either positive or negative.
Patient Survey Results
There were 266 patients that completed the survey (13.3% response rate). Table 1 provides descriptive characteristics of the overall sample that were sent surveys (n=2000), along with characteristics of responders (n=266) and non-responders (n=1734). Compared to non-responders, survey responders were slightly older (70.2 vs 68.2 years) and more likely to be white (68.4% vs 47.6%), live in a rural area (45.5 vs 35.5%) and be insured by Medicare (51.1 vs 38.9%). Of the respondents, 231 self-reported being on AF medication (86.8%) and 213 completed the ASK-12 medication adherence scale (80.1%).
Table 1.
Descriptive Characteristics of Survey Sample
| Overall (n=2000) | Responders (n=266) | Non-Responders (n=1734) | |
|---|---|---|---|
| Characteristic | n (%) or Mean ± SD | n (%) or Mean ± SD | n (%) or Mean ± SD |
| Age (mean)* | 68.5± 12.7 | 70.2±11.0 | 68.2±13.0 |
| Race* | |||
| Asian | 302 (15.1%) | 24 (9.0%) | 278 (16.0%) |
| American Indian Alaska Native/Native Hawaiian Pacific Islander | 112 (5.6%) | 10 (3.8%) | 102 (5.9%) |
| Black | 181 (9.1%) | 18 (6.8%) | 163 (9.4%) |
| White | 1007 (50.4%) | 182 (68.4%) | 825 (47.6%) |
| Other racial groups/Missing race data | 398 (19.9%) | 32 (12.0%) | 366 (21.1%) |
| Geography type* | |||
| Urban | 1263 (63.2%) | 145 (54.5%) | 1118 (64.5%) |
| Rural | 737 (36.9%) | 121 (45.5%) | 616 (35.5%) |
| Insurance at most recent visit (prior to sample draw)* | |||
| Medicare* | 810 (40.5%) | 136 (51.1%) | 674 (38.9%) |
| Medicaid | 450 (22.5%) | 50 (18.8%) | 400 (23.1%) |
| Commercial, other, unknown insurance | 740 (37.0%) | 80 (30.1%) | 660 (38.1%) |
Note: *Indicates a significant difference between responder and non-responder at p≤0.05.
A complete case analysis approach was used on the 213 patients that completed all ASK-12 scale questions. Overall, 68.5% (n=146) of patients were considered adherent based on their ASK-12 scale score. Table 2 provides descriptive characteristics by adherence status. The mean ASK-12 score for the entire sample was 20.5. The mean ASK-12 scores in the adherent and non-adherent groups were 17.1 ± 3.0 and 28.1 ± 4.8, respectively. Of the three ASK-12 subscales, the behavior subscale score was found to be relatively low (mean score=6.6 (range of 5–25)), suggesting this domain may have contributed more to patients’ self-reported medication adherence.
Table 2.
Descriptive Characteristics by Medication Adherence Status
| All Patients (Complete Case Sample; n=213) | Adherent (n=146) | Non-Adherent (n=67) | |
|---|---|---|---|
| Characteristic | n (%) or Mean ± SD | n (%) or Mean ± SD | n (%) or Mean ± SD |
| Age (mean) | 70.6±10.6 | 72.0±10.0 | 67.5± 11.0 |
| Sex (% Female) | 94 (44.3%) | 63 (43.5%) | 31 (46.3%) |
| Race | |||
| Asian | 17 (8.0%) | 15 (10.3%) | 2 (3.0%) |
| American Indian Alaska Native/Native Hawaiian Pacific Islander | 8 (3.8%) | 6 (4.1%) | 2 (3.0%) |
| Black | 12 (5.6%) | 6 (4.1%) | 6 (9.0%) |
| White | 152 (71.4%) | 105 (71.9%) | 47 (70.2%) |
| Other racial groups/Missing race data | 24 (11.3%) | 14 (9.6%) | 10 (14.9%) |
| Education level | |||
| 4-year college degree or more | 96 (46.2%) | 74 (51.8%) | 22 (33.9%) |
| Vocational training or 2-year degree | 46 (22.1%) | 27 (18.9%) | 19 (29.2%) |
| High school diploma or high school equivalent | 62 (29.8%) | 39 (27.3%) | 23 (35.4%) |
| Less than high school | 4 (1.9%) | 3 (2.1%) | 1 (1.5%) |
| Geography type | |||
| Urban | 112 (52.6%) | 74 (50.7%) | 38 (56.7%) |
| Rural | 101 (47.4%) | 72 (49.3%) | 29 (43.3%) |
| Insurance at most recent visit (prior to sample draw) | |||
| Medicare | 114 (53.5%) | 85 (58.2%) | 29 (43.3%) |
| Medicaid | 34 (16.0%) | 19 (13.0%) | 15 (22.4%) |
| Commercial, other, or unknown insurance | 65 (30.5%) | 42 (28.8%) | 23 (34.3%) |
| ASK-12 Scores | |||
| Overall ASK-12 score | 20.5 ± 6.3 | 17.1 ± 3.0 | 28.1 ± 4.8 |
| Inconvenience/Forgetfulness subscale | 5.4 ± 2.5 | 4.6 ± 1.9 | 7.2 ± 2.8 |
| Treatment beliefs subscale | 8.5 ± 3.8 | 7.0 ± 2.2 | 11.9 ± 4.4 |
| Behavior subscale | 6.6 ± 2.9 | 5.5 ± 1.1 | 9.0 ± 3.9 |
Notes: Adherence was defined based on overall ASK-12 Score. Patients with an ASK-12 score ≥ 23 were identified as non-adherent and patients with scores <23 were considered adherent.
Health Status, Health Care Experiences, and Social Determinants of Health
Overall, 62% and 83% of patients reported good to excellent physical health and mental/emotional health, respectively. However, results from the univariate regression models showed non-adherent patients to be at significantly higher risk of reporting poorer mental/emotional health (RR=1.75, 95% CI: 1.16–2.63). Non-adherent patients were also more likely to report poor healthcare experiences and higher healthcare discrimination (+2.06 points) than adherent patients (p<0.0001) on the patient survey. Finally, non-adherent patients had more than twice the risk of reporting one or more social needs, such as housing, food, transportation, utilities, or help with day-to-day activities (RR=2.22, 95% CI: 1.53–3.22) (Supplemental File B; Table S9). Figure 1 illustrates the percentage of patients that agreed with each survey statement by adherence status.
Figure 1.
Percent of patients that agreed with each survey statement by adherence status.
Poisson Regression models were run (Table 3) to identify factors associated with respondents’ AF medication non-adherence. Overall, receiving information needed to manage one’s care was associated with a 42% lower risk of non-adherence (RR=0.58, 95% CI: 0.36–0.94), and low self-efficacy was associated with 73% higher risk of non-adherence (RR=1.73, 95% CI: 1.03–2.90). Sub-group analyses were also performed to assess the associations between different factors and medication non-adherence in different age groups, insurance types, geographic locations, and other chronic health conditions (Supplemental File A).
Table 3.
Multivariable Poisson Regression of ASK-12 Medication Adherence Status for Overall Survey Sample
| Risk Ratio (RR) | 95% CI | p-value | |
|---|---|---|---|
| Healthcare Experiences | |||
| You ALWAYS got all the information you needed to help manage your care | 0.58 | (0.36–0.94) | 0.026 |
| Discrimination in Medical Settings | |||
| Overall discrimination score | 1.05 | (0.99–1.1) | 0.084 |
| Healthcare Insurance Coverage and Access | |||
| Obtained all the medicines you or a doctor believed necessary? | 0.56 | (0.31–1.01) | 0.055 |
| Social Determinants of Health | |||
| See or talk to people that you care about and feel close to less than 5 times a week | 0.87 | (0.55–1.36) | 0.5334 |
| Reported at least one need | 1.24 | (0.7–2.19) | 0.464 |
| Self-Efficacy | |||
| Disagree to any of the 4 questions | 1.73 | (1.03–2.9) | 0.040 |
| Health Conditions and Status | |||
| Fair/Poor physical health | 0.88 | (0.56–1.38) | 0.577 |
| Fair/Poor mental emotional health | 1.14 | (0.65–2.01) | 0.647 |
Notes: Adjusted for age (continuous), gender, race, geography type, Medicaid at most recent visit, and number of chronic health conditions (other than AF). The reference category is Adherent. Higher RR (>1.00) indicate higher likelihood of non-adherence, and lower RR (<1.00) indicates less likelihood of non-adherence.
Age
For patients ≤ 70 years old (Table S1), receiving information needed to manage one’s care (RR=0.41, 95% CI: 0.25–0.67), and having coverage and access to obtain all necessary medications (RR=0.46, 95% CI: 0.26–0.81) were associated with lower risk of non-adherence, while higher healthcare discrimination scores were associated with a higher risk of non-adherence (RR=1.07, 95% CI: 1.03–1.12). Among patients ≥ 71 years of age (Table S2), self-efficacy was the only statistically significant predictor of adherence; low self-efficacy was associated with nearly four times the risk of non-adherence for this group of patients (RR=3.73, 95% CI: 1.55–8.95).
Insurance
Among patients with Medicaid insurance (Table S3), greater discrimination in the healthcare setting was associated with an increased likelihood of non-adherence (RR=1.29, 95% CI: 1.02–1.63), whereas having coverage and access to obtain all necessary medications was associated with adherence (RR=0.13, 95% CI: 0.02–0.88). For patients with Medicare, commercial, or other types of health care coverage (Table S4), always receiving the information needed to manage their health was the only significant factor associated with adherence (RR=0.55, 95% CI: 0.31–0.96).
Geography
Differences in predictors of medication non-adherence were found based on geographic locations. For patients living in rural areas (Table S5), low self-efficacy was associated with over two times the risk of non-adherence (RR=2.31, 95% CI: 1.08–4.96). Whereas greater discrimination in medical settings was the only statistically significant predictor of adherence for patients living in urban areas (Table S6), where those experiencing higher levels of discrimination in medical settings had a greater risk of non-adherence (RR=1.06, 95% CI: 1.01–1.11).
Chronic Health Conditions
Patients with AF and ≤ 2 chronic conditions (Table S7) that reported getting all necessary information to help manage their health were 63% less likely to be non-adherent (RR=0.37, 95% CI: 0.16–0.84). Whereas for patients with AF and ≥3 other chronic conditions (Table S8), having access and coverage to obtain all medicines needed was significantly associated with a 65% decreased risk of non-adherence (RR=0.35, 95% CI: 0.16–0.76).
Patient Interview Results
Sixty-one patients were contacted to schedule an interview. Data saturation was determined following 28 patient interviews. Table 4 provides descriptive characteristics of the patient interview sample. Overall, interviewees represented both urban and rural locations (64% and 36%, respectively), were primarily White (82%) and largely using Medicaid insurance (64%). Slightly over half of interviewees (53.6%) were considered medication adherent based on their ASK-12 survey scores.
Table 4.
Patient Interview Characteristics (n=28)
| n (%) or Mean ± SD | |
|---|---|
| Age | 62 ± 11 |
| Sex | |
| Female | 14 (50%) |
| Race | |
| White | 23 (82.1%) |
| Other Race | 5 (17.9%) |
| Geography type | |
| Urban | 18 (64.3%) |
| Rural | 10 (35.7%) |
| Insurance at most recent visit | |
| Medicaid | 18 (64.3%) |
| Non-Medicaid | 10 (35.7%) |
| Education | |
| 4-year college degree or more | 10 (35.7%) |
| Vocational training or 2-year degree | 8 (28.6%) |
| High school diploma, equivalent, or less | 10 (35.7%) |
| ASK-12 Score Category | 23 ± 7 |
| Adherent | 15 (53.6%) |
| Non-Adherent | 11 (39.2%) |
Note: Other race and non-Medicaid insurance categories are collapsed for small number suppression two patients missing adherence classification.
During interviews most patients self-described as being very adherent to their AF medications; however, survey responses classified 46.4% of the interviewees as non-adherent to their medications. Patients attributed this adherence commonly to perceptions of AF as a serious health condition, positive attitudes towards medications, adequate communication and support from healthcare clinician, and medication reminders and organization systems.
Four major topics related to medication adherence arose from the interviews, including 1) diagnosis, medications, and co-morbidities, 2) experiences with clinicians and access to care, 3) system and clinician-level factors, and 4) patient level factors. Quotes related to each topic are provided in Supplemental File C.
Diagnosis, Medications, and Co-Morbidities
Most interviewees reported being diagnosed with AF in the emergency department while there for heart-related complications, with few participants reporting being diagnosed during a routine check-up. Interviewees also commonly noted being on multiple cardiovascular medications, with OACs being the most common. Participants also reported multiple co-morbidities along with AF, such as hypertension, hypercholesterolemia, Type II diabetes, congestive heart failure, and hypothyroidism. A common barrier noted by interviewees was the uncertainty around how all their medications interact, as well as prior experience with side effects from medications. The most frequently mentioned challenge for medication adherence was being on several medications to manage their conditions and the possibility of needing to be on all medications for the rest of their lives.
Experience with Clinicians and Access to Care
Interviewees reported wide variation in their experiences with clinicians and access to care, which influenced how they perceived and adhered to their medication regimen. Most interviewees reported receiving AF care from a cardiologist, though some were managed by nurse practitioners and primary care clinicians. Many described frequent visits early after their AF diagnosis that later tapered to annual or biannual appointments. Overall, participants felt they had trusting relationships with their healthcare clinicians, noting that their concerns and preferences were generally respected during treatment planning. However, a few interviewees expressed not feeling heard and excluded from decisions about their AF management. These patients expressed frustration with their clinicians and felt they had to defer to clinicians’ recommendations without being able to voice concerns.
Communication with clinicians was also noted as a major factor influencing medication adherence. Many interviewees described how support and positive communication with their clinicians led to enhanced medication adherence. Some patients described working with their clinician to try different medications and dosages until landing on their current prescription. On the other hand, some interviewees shared concerns about side effects of medications and/or interactions between different medications they were taking for other health conditions. These concerns seemed to largely stem from insufficient support and communication from their care team to help explain or address any concerns.
System-Level Factors That Affect Medication Adherence
Interviewees identified cost as a primary external factor affecting medication adherence. Generally, medications were viewed as expensive, and interviewees felt they would likely be unable to afford their medications without health insurance. Patients described situations where their insurance coverage would change, causing medications to become unaffordable, often resulting in having to change medications.
Patient-Level Factors That Affect Medication Adherence
A major driver of AF medication adherence was patient’s perceiving AF as a serious health condition that affects their life, and the related perception that their medication is helping the condition. Interviewees described feeling tired, lightheaded, and easily out of breath during episodes of AF and shared that if they did not treat their AF, they would be constantly worried that something would trigger an episode, ultimately decreasing their quality of life. Interviewees went on to describe feeling less anxious when their AF was managed through a combination of medical procedures, medication, and lifestyle modifications. Yet, while the majority of interviewees felt their AF medications improved their health and kept their condition under control, many did not like feeling as though they were dependent on medications.
In addition to the perceived benefit of their AF medications, interviewees also reported that making their medication part of their daily routine helped with adherence. Many patients reported the use of various reminders and organization systems to stay on schedule and adherent.
Mixed Methods Findings – Integration of Patient Survey and Patient Interview Findings
Findings from the quantitative analyses using data from the patient survey showed low self-efficacy and healthcare experience, primarily clinician support and communication, as influencing factors of medication adherence. Those that felt they received all information necessary to manage their care and believed they could manage their own care, were more likely to be adherent to their AF medications. Further, coverage and access to obtain all medications arose as a significant predictor of medication adherence in several of the group sub-analyses. Themes from the qualitative data obtained from the patient interviews that correspond to these quantitative results related to predictors of medication adherence are presented using an integrated approach in Table 5.
Table 5.
Integrated Mixed Methods Results of Medication Adherence
| Quantitative Data Summary | Patient Quote | Qualitative Data Summary | Integrating Results |
|---|---|---|---|
| Healthcare experiences and access to care | |||
| Patients that always received the information needed to manage their care were 42% more likely (RR =0.58 [95% CI 0.36–0.94]) to be adherent to their medications Non-adherent patients were also more likely to report poor healthcare experiences and had a discrimination score that was 2.06 points higher than adherent patients (p<0.0001) |
“They explained what it’s doing to the heart, how it’s not pumping like it used to, how blood will pool up in there. They pretty thoroughly explained it. Of course, I was a little shook up in the hospital, so went on to research it myself. Yes, I was well-informed and educated, had questions answered while I was in the hospital”. “…We’ll just add that to your regimen”. I was like, “Tell me what it’s going to do for me. Is it going to help me?” She said, “Just take it. I’m the doctor here”. That bothered me…. “I lost confidence in the cardiologist”. |
Most participants felt that their clinicians were responsive to their care and sensitive to their concerns and preferences when determining their treatment plan. Some participants felt that their clinicians explained their medication and treatment plan well, explaining side effects alternative options, etc. On the other hand, some participants felt that they were not heard and instead they had few choices in their AF management plan. When this was the case, participants expressed frustration and reported poor relationships with their clinician and healthcare experiences. | Adequate support and communication from clinicians consistently emerged as important factors in AF management and satisfaction with care. It is also important to note that given some subgroup findings (eg, discrimination as a barrier to adherence among patients living in urban areas and those with Medicaid insurance), non-adherence may be related to systemic issues that result in health inequities. |
| Social determinants of health | |||
| Non-adherent patients had more than twice the risk of reporting one or more social needs, such as housing, food, transportation, utilities, or help with day-to-day activities (RR=2.22 [95% CI: 1.53–3.22]) Adherent patients were 69% more likely to have health insurance over the 6 months prior to the interview (RR=0.31 [0.25–0.38]). Adherent patients were 53% more likely to obtain all medicines (RR=0.47 [95% CI: 0.25–0.89]) and 57% more likely to receive all medical care, tests, or treatments (RR=0.43 [0.30–0.62]) that their doctor believed necessary. |
“Until you get costs under control, you’re not going to have a lot of compliance for some of these…” “Because I can see a lot of people on fixed income who just have to decide to either pay their bills or take their meds…That’s a major issue”. “For me personally, I think my insurance is going to end very soon, but as of its current state, I am going to have to juggle a lot of medications, including potentially changing a couple of them that are proprietary as well as seeing how much the total cost is going to be and seeing what else I have to eliminate from my life in order to keep taking the medications. It might just be not going out to a meal twice a month, fine, no big deal”. For other people, it’s, “Can I pay my gas bill?” |
Cost of medications was commonly discussed as a barrier to adherence. In general, medications were largely viewed as expensive, and interviewees felt they would likely not be able to afford their medications without health insurance. Patients reported that having medications at least partially covered by insurance helped ensure that they were able to consistently access their medication, an important first step towards adherence. | Quantitative and qualitative study findings highlighted social and economic needs as drivers of adherence. Specifically, quantitative results showed that non-adherent patients were more likely to report social determinants of health needs and qualitative results highlighted cost and lack of insurance as main drivers of non-adherence. |
| Self-efficacy and perception of medication effectiveness | |||
| In the overall sample, low self-efficacy was associated with 73% higher likelihood of non-adherence (RR=1.73 [1.03–2.90]). | “…We’ll just add that to your regimen”. I was like, “Tell me what it’s going to do for me. Is it going to help me?” She said, “Just take it. I’m the doctor here”. That bothered me…. “I lost confidence in the cardiologist”. “I feel like the doctor-patient relationship between me and him is that I do not really understand. He does and he knows what’s best for me and that I should just be okay with that. I should just do what he says and not really question it over and over again. Like, ‘I told you, you need to take [your blood thinner]. I told you, you need to take blood pressure medication. I told you to change your diet. I told you to exercise. Are you doing all of those things?’ ‘Yes.’ ‘Okay. I’ll see you next year.’” |
Patients expressed feeling as if they did not have control over their treatment decisions. Some patients also described not understanding why clinicians prescribed certain treatments. An important facilitator to adherence was feeling a sense of responsibility for one’s own health and having self-discipline to stick with a medication routine. |
Patients that felt less control over their health and treatment plan, showed lower self-efficacy based on survey results. Low self-efficacy was associated with medication non-adherence. Further, interview findings suggested patients that did not feel control over treatment decisions or feel comfortable advocating for their health shared more barriers to medication adherence. |
Notes: Themes from the qualitative data analysis corresponding to these quantitative results related to predictors of medication adherence are presented using an integrated approach.
Clinician Survey Results
Of the 2690 clinicians who were distributed a survey, a total of 78 (3.0%) completed the survey. Most respondents were credentialed as MDs or DOs (62.8%), with physician assistants and nurse practitioners making up the additional third of responding clinicians. Specialties of responding clinicians varied, with 38.5% of respondents from family medicine, 28.2% from internal medicine, and 21.8% from cardiology (general cardiology and electrophysiology) departments. Over half of respondents had been in their current role for over 10 years (57.7%), with only 3.9% being in their current role for less than two years. Table 6 provides the descriptive characteristics for the clinician sample.
Table 6.
Descriptive Characteristics of the Clinician Sample (N=78)
| Characteristic | N (%) |
|---|---|
| Credential | |
| MD | 43 (55.1%) |
| DO | 6 (7.7%) |
| PA/NP | 24 (30.8%) |
| Other | 5 (6.4%) |
| Specialty (mark all that apply) | |
| Family medicine | 30 (38.5%) |
| Internal medicine | 22 (28.2%) |
| Cardiology/General cardiology | 11 (14.1%) |
| Electrophysiology | 6 (7.7%) |
| Other | 12 (15.4%) |
| Time in current role | |
| 1–2 years | 3 (3.9%) |
| 2–5 years | 14 (17.9%) |
| 5–10 years | 13 (16.7%) |
| 10–15 years | 16 (20.5%) |
| 15–20 years | 4 (5.1%) |
| 20+ years | 25 (32.1%) |
Overall, clinicians reported low levels of perceptions of medication adherence for their patients with AF. Specifically, 87% of clinician responders perceived their typical AF patient as non-adherent. The mean clinician ASK-12 score for a typical patient with AF, which ranged from 12 to 60, was 33.2. Mean ASK-12 scores were similar between specialty (cardiology) and primary care (family medicine and internal medicine) clinicians. The mean clinician ASK-12 score was 12.3 points higher than the mean score reported by patients. Figure 2 provides the mean overall and subscale clinician ASK-12 scores, as well as the percentage of clinicians that agree with each statement by subscale. Within the inconvenience and forgetfulness subscale, clinicians most frequently reported that their patients “just forget to take their medications some of the time (62.8%)”. When asked about treatment beliefs, clinicians largely felt that patients were able to work with their care team to make decisions and ask questions about medications (83.3%), but less frequently agreed that patients felt confident that their medications helped them (52.6%). Finally, clinicians cited cost and side effects as primary reasons why their patients would skip or stop taking medication (51.3%) when answering questions about patient behavior.
Figure 2.
ASK-12 subscale scores and percent of clinicians that agreed with each statement. Footnote: ASK-12 is a 12-item scale that assesses barriers and facilitators of adherence that consists of three subscales: inconvenience/forgetfulness (3 items), health beliefs (4 items), and behaviors (5 items).11,12 Each subscale score was computed as the sum of all items within that scale.
When clinicians were asked about tools and decision aids they use when assessing and addressing medication adherence, 83.3% clinicians reported asking patients if they changed their medication regimen. However, only 21.8% reported using a tool to assess patient adherence levels. Primary care clinicians were much more likely to use a medication adherence tool than specialty care clinicians (27.5% vs 6.7%, respectively). Overall clinicians reported frequent communication with patients, offering information about their condition and treatment options (90%). Although, much smaller percentages reported frequently asking patients about their desired level of involvement in their AF management decision making (58.0%) and/or asking patients about the support they have or need for taking their medications as prescribed (51.0%).
Finally, when clinicians were asked about the types of information regarding AF guideline recommendations that they would find helpful if available electronically, the three most selected types of information were cost of medications/formulary information (85.9%), interaction with other drugs or side effects (73.1%), and CHA2DS2-VASc score (71.8%). Table 7 provides a comparison of the type of information regarding AF guidelines recommendations that primary versus specialty care clinicians would find helpful if available electronically.
Table 7.
Types of Information Regarding AF Guideline Recommendations That Select Clinicians Ranked as Being Helpful When Seeing Patients
| ALL Clinicians (n=78) % (95% C.I.) |
Primary Care (n=51) % (95% C.I.) |
Specialty (n=15) % (95% C.I.) |
||||
|---|---|---|---|---|---|---|
| An alert that blood pressure is out of normal range | 32.05% | (22.67–43.15) | 41.18% | (28.61–55.01) | 13.33% | (3.36–40.54) |
| Reminders of steps of diagnostic workup (eg screening for other CVD risk factors) | 26.92% | (18.26–37.8) | 33.33% | (21.83–47.23) | 0.00% | - |
| An alert when medication is recommended | 50.00% | (39.08–60.92) | 50.98% | (37.52–64.3) | 40.00% | (19.18–65.19) |
| A bottom line statement of what medication is recommended | 52.56% | (41.54–63.35) | 54.90% | (41.22–67.88) | 40.00% | (19.18–65.19) |
| Alternative medication recommendations | 42.31% | (31.88–53.47) | 47.06% | (33.9–60.64) | 26.67% | (10.38–53.31) |
| Interaction with other drugs or side effects | 73.08% | (62.2–81.74) | 70.59% | (56.79–81.42) | 86.67% | (59.46–96.64) |
| Cost of medications/formulary information | 85.90% | (76.3–92.02) | 88.24% | (76.19–94.62) | 86.67% | (59.46–96.64) |
| Resources for lifestyle modification counseling | 32.05% | (22.67–43.15) | 35.29% | (23.5–49.2) | 26.67% | (10.38–53.31) |
| Links to patient educational materials | 55.13% | (44.02–65.75) | 60.78% | (46.91–73.11) | 46.67% | (24.09–70.7) |
| An abstract of a relevant journal article | 7.69% | (3.5–16.08) | 7.84% | (2.98–19.11) | 6.67% | (0.93–35.2) |
| Full copy of an important clinical trial | 5.13% | (1.94–12.88) | 5.88% | (1.91–16.71) | 6.67% | (0.93–35.2) |
| Summary text of the guideline | 38.46% | (28.37–49.65) | 39.22% | (26.89–53.09) | 33.33% | (14.6–59.4) |
| CHA2DS2VASC score | 71.79% | (60.85–80.65) | 80.39% | (67.25–89.11) | 60.00% | (34.81–80.82) |
Discussion
Results from this mixed methods study reveal that medication adherence is complex in patients with AF, with several factors contributing to adherence. Overall, 68.5% (n=146) of patients were considered adherent based on their ASK-12 scale score, a rate that is similar to past research reporting an OAC adherence rate (Proportion of days covered/medication possession ratio ≥80%) of 66% in 594,784 patients with AF.23 The survey highlighted poor healthcare experiences, social and economic needs, and self-efficacy, as key factors for supporting patients to become or remain adherent to their medications.24 The patient interviews supported these findings and offered more in-depth insight between these factors and medication adherence, including the poor healthcare experiences being described as not having a say in their AF management, insufficient support and communication from their care team to help explain or address concerns, and discrimination in the healthcare setting.
Medication adherence is multidimensional; therefore, interventions must address multiple factors to improve and sustain adherence.6,9,19,25 The results from this study highlight several key opportunities that patients with AF and clinicians can take to improve medication adherence. There were observed differences between clinician and patient perceptions of adherence. Clinicians generally perceived lower medication adherence among their patients and patients, whereas patients reported higher levels of adherence. This discrepancy is consistent with prior literature and may reflect factors such as self-report bias, clinician assumptions, and the limited use of objective adherence measures.26–28 Although over 80% of clinicians reported that they regularly communicate information and address patient beliefs about medication, only a little over 20% reported using structured tool to assess medication adherence. Evidence suggests that tools and training designed to support and improve clinician-patient communication, can enhance education, shared decision-making and trust-building, which can substantially reduce this discrepancy in perceived medication adherence and improve outcomes.19,25 This supports the results from the survey showing that when patients feel they have received all information needed to manage their care they are more likely to be adherent to medications.
Social and economic needs were also highlighted in the quantitative and qualitative findings as related to adherence. Coverage and access to needed care and medications arose in both the patient survey and interviews as facilitators of medication adherence, with interviewees particularly mentioning cost as a potential barrier. Clinicians also noted cost as a main driver of non-adherence among their patients with AF. Social determinants of health also emerged as a strong predictor of adherence. In fact, non-adherent patients were more than twice as likely to report one or more social needs, such as housing, food, transportation, utilities. Based on these findings, as well as past research showing that social determinants of health account for 30–55% of health outcomes, it is important that clinicians consider their patients’ lives holistically, including their social determinants of health to improve chronic disease treatment outcomes.29–31 One step clinicians can take to lessen the cost burden, especially for patients with economic needs, is to prescribe medications that are on the patient’s insurance formulary when clinically appropriate and/or provide information on prescription discount cards, prescription assistance programs, or other coupons/savings programs that can be used.32
Self-efficacy was found to be a strong predictor of adherence in patients with AF, especially those living in rural areas and older adults. These results are similar to previous research showing low self-efficacy to be associated with non-adherence to health recommendations and behaviors among chronic disease patients, rural patients, and older adults.33–36 Increasing a patients’ self-efficacy could improve their engagement in health-promoting behaviors and treatment adherence. In the current study, assessed self-efficacy seems to be correlated with understanding what to do when they have a health problem, understanding of where to go when they have a health problem, and comfort advocating for their health. Therefore, as seen in other studies, the current sample’s medication self-efficacy seems to be related to health and medication literacy.19,35,37 This finding reveals an opportunity for health systems and clinicians to consider their roles in supporting patient self-efficacy, offering patients with AF information on their disease and various treatment plans, as well as ensuring that patients feel control over decisions that affect their health.
This study used a mixed methods approach to understand medication adherence in patients with AF. The integration of quantitative survey results with qualitative interview findings, provides a more comprehensive understanding of barriers and facilitators to OAC adherence. Further, both patients and clinicians were surveyed. Gaining the perspective on medication adherence from both patients and clinicians may allow for the development of more effective and acceptable interventions. For example, the current results indicate the need for more open communication between patients and clinicians and shared decision making. Shared decision-making may also help clinicians tailor medication regimens to fit patients’ daily routines, schedules, and existing behaviors. Previous research in patients with AF suggests that integrating medication into patients’ established routines can minimize medication burden, which can support better adherence and persistence.38 Further, system level interventions, such as integrating adherence assessments into routine visits or EMR alerts for skipped refills, may open discussion between patients and clinicians related to medication adherence barriers and facilitators. This can help inform care decisions, build trust, and improve patient motivation to take their medications as prescribed.39
The findings from the current study should also be interpreted considering the following limitations. Medication adherence was assessed using a survey that relied on self-report rather than using objective measures, such as pharmacy fill data. Self-reported medication adherence data are prone to systematic biases, including recall biases and social desirability bias.40 Future research should seek to triangulate self-report adherence data with objective measures of adherence, including claims/pharmacy fill data to strengthen the current findings. Further, while stratified random sampling was used when determining who would be sent the survey, the low response rate should be considered when interpreting results and may have introduced participation bias, as these patients may have been healthier, more health-literate, and more willing to engage than the source population, increasing their likelihood of adherence.41 Further, the survey responders were older, more likely to be white, and live in a rural area than non-responders, all of which may limit the generalizability of the results. While the survey sample appeared to be representative of the US population, differences in medication adherence facilitators and barriers may exist by race, ethnicity, geographic location, type of OAC prescribed, etc.23 that cannot be fully understood from these results. Patient survey and interview respondents might have more trust in the health care system or be less likely to have experienced discrimination in a health care setting, compared to non-responders. This may have limited the ability to fully capture the scope and impact of factors such as discrimination and social determinants of health on medication adherence. Additionally, while all patients had an AF diagnosis and a prescription for an OAC in their medical chart, indicating elevated risk, CHA2DS2-VASc scores were not obtained for the survey participants. Future research should assess stroke risk profile using the CHA2DS2-VASc score to ensure accurate evaluation of anticoagulant adherence. There was a low response rate for the clinician survey, with only 3.0% of clinicians within PSJH that treat patients with AF being represented in the sample. Therefore, the inconsistency in clinicians’ and patients’ perceptions of anticoagulant adherence may simply be due to participation bias, especially of clinicians. Future research would benefit from gaining perspective from more clinicians, especially when trying to determine actionable steps that health systems can take in improving medication adherence of patients.
Conclusion
Past studies have shown high rates of medication non-adherence in patients with AF.6–8 This study used a mixed methods approach to provide a comprehensive understanding of facilitators and barriers to medication adherence, a complex behavior. Predictors of adherence differed across patient populations, highlighting the importance of developing personalized strategies to improve OAC adherence. Even still, common factors influencing adherence included obtaining all the information necessary to manage AF and patient self-efficacy. While clinician responders perceived relatively low adherence among their patients overall, they did align with patients on the need for improved communication and shared decision making around AF treatment. Interventions that promote clinician and patient communication and medication adherence education should be explored, as they may be pivotal in improving adherence and patient outcomes.
Acknowledgments
The abstract of this paper was presented at the American Heart Association’s 2024 Scientific Sessions as a poster presentation/conference talk with interim findings. The poster’s abstract was published in “Poster Abstracts” in Circulation: https://www.ahajournals.org/doi/10.1161/circ.150.suppl_1.4118668
Funding Statement
IncThis study was sponsored by Pfizer Inc. and the Bristol Myers Squibb Alliance.
Data Sharing Statement
Aggregate data may be made available with reasonable request to the corresponding author.
Ethics Approval and Consent to Participate
This study was approved by the PSJH institutional review board (Study Number STUDY2023000907). Survey completion denoted consent to participate. The participants that were interviewed provided verbal informed consent for the information shared during their interview to be used for research.
Consent to Publish
All patient quotes are de-identified and interviewees provided verbal informed consent for information shared during their interview to be published.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
DF and LM are employees of Pfizer Inc. KT and MA were employees of Pfizer Inc. at the time of this work. Monique Gill reports Support for the manuscript from Pfizer, during the conduct of the study. Eleni Wilson reports Support for the manuscript from Pfizer, during the conduct of the study; Grants or contracts from Boston Scientific, outside the submitted work. Ty Gluckman reports Consulting fees from OptumRx, outside the submitted work. Mohammad Ateya reports Support for the manuscript from Pfizer Inc., during the conduct of the study; Stock or stock options from Novartis, Pfizer; Other interests from Novartis, Pfizer, outside the submitted work. Feng Dai reports Stock or stock options from Pfizer Inc, outside the submitted work. Laura Manzey reports Support for the manuscript from Pfizer Inc., during the conduct of the study; Stock or stock options from Pfizer Inc. outside the submitted work. Kevin Townsend reports Support for the manuscript from Pfizer Inc., during the conduct of the study. Staci Wendt reports Support for the manuscript from Pfizer, during the conduct of the study. The authors report no other conflicts of interest in this work.
References
- 1.Safari A, Helisaz H, Salmasi S, et al. Association between oral anticoagulant adherence and serious clinical outcomes in patients with atrial fibrillation: a long-term retrospective cohort study. J Am Heart Assoc. 2024;13(18):e035639. doi: 10.1161/JAHA.124.035639 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lin DS, Lo H, Huang K, Lin T, Lee J. Efficacy and safety of direct oral anticoagulants for stroke prevention in older patients with atrial fibrillation: a network meta-analysis of randomized controlled trials. J Am Heart Assoc. 2023;12(23):e030380. doi: 10.1161/JAHA.123.030380 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Joglar JA, Chung MK, Armbruster AL, et al. 2023 ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2024;149(1). doi: 10.1161/CIR.0000000000001193 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.January CT, Wann LS, Calkins H, et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in Collaboration With the Society of Thoracic Surgeons. Circulation. 2019;140(2):e125–17. doi: 10.1161/CIR.0000000000000665 [DOI] [PubMed] [Google Scholar]
- 5.Lip GYH, Nieuwlaat R, Pisters R, Lane DA, Crijns HJGM. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137(2):263–272. doi: 10.1378/chest.09-1584 [DOI] [PubMed] [Google Scholar]
- 6.Tarn DM, Shih K, Tseng CH, Thomas A, Schwartz JB. Reasons for nonadherence to the direct oral anticoagulant apixaban. JACC Adv. 2023;2(1):100175. doi: 10.1016/j.jacadv.2022.100175 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Veale EL. Pharmacy-led management of atrial fibrillation: improving treatment adherence and patient outcomes. Integr Pharm Res Pract. 2024;13:101–114. doi: 10.2147/IPRP.S397844 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Capiau A, Grymonprez M, Scheire S, et al. Barriers and facilitators for adherence to direct oral anticoagulants in patients with atrial fibrillation: a qualitative approach. Thromb Res. 2023;229:164–169. doi: 10.1016/j.thromres.2023.06.012 [DOI] [PubMed] [Google Scholar]
- 9.Lee WK, Woo SI, Hyun DK, Jung SY, Kim MS, Lee J. Impact of treatment adherence on the effectiveness and safety of oral anticoagulants in patients with atrial fibrillation: a retrospective cohort study. Eur Heart J Qual Care Clin Outcomes. 2023;9(3):216–226. doi: 10.1093/ehjqcco/qcac023 [DOI] [PubMed] [Google Scholar]
- 10.Neiman AB, Ruppar T, Ho M, et al. CDC grand rounds: improving medication adherence for chronic disease management. Innov Opportunities. 2017;66(45). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rural-Urban Commuting Area Codes | economic Research Service. Available from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes. Accessed April 29, 2025.
- 12.Matza LS, Park J, Coyne KS, Skinner EP, Malley KG, Wolever RQ. Derivation and validation of the ASK-12 adherence barrier survey. Ann Pharmacother. 2009;43(10):1621–1630. doi: 10.1345/aph.1M174 [DOI] [PubMed] [Google Scholar]
- 13.Hahn SR, Park J, Skinner EP, et al. Development of the ASK-20 adherence barrier survey. Curr Med Res Opin. 2008;24(7):2127–2138. doi: 10.1185/03007990802174769 [DOI] [PubMed] [Google Scholar]
- 14.Takemura M, Nishio M, Fukumitsu K, et al. Optimal cut-off value and clinical usefulness of the Adherence Starts with Knowledge-12 in patients with asthma taking inhaled corticosteroids. J Thorac Dis. 2017;9(8):2350–2359. doi: 10.21037/jtd.2017.06.115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Singer SJ, Friedberg MW, Kiang MV, Dunn T, Kuhn DM. Development and preliminary validation of the patient perceptions of integrated care survey. Med Care Res Rev. 2013;70(2):143–164. doi: 10.1177/1077558712465654 [DOI] [PubMed] [Google Scholar]
- 16.Peek ME, Nunez-Smith M, Drum M, Lewis TT. Adapting the everyday discrimination scale to medical settings: reliability and validity testing in a sample of African American Patients. Ethn Dis. 2011;21(4):502–509. [PMC free article] [PubMed] [Google Scholar]
- 17.The PRAPARE Screening Tool | PRAPARE. Available from: https://prapare.org/the-prapare-screening-tool/. Accessed July 31, 2024.
- 18.Center for Medicare and Medicaid Services. Accountable Health Communities Model. Available from: https://innovation.cms.gov/innovation-models/ahcm. Accessed August 21, 2020.
- 19.Huang YM, Shiyanbola OO. Investigation of barriers and facilitators to medication adherence in patients with type 2 diabetes across different health literacy levels: an explanatory sequential mixed methods study. Front Pharmacol. 2021;12:745749. doi: 10.3389/fphar.2021.745749 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fetters MD, Curry LA, Creswell JW. Achieving integration in mixed methods designs—principles and practices. Health Serv Res. 2013;48(6 Pt 2):2134–2156. doi: 10.1111/1475-6773.12117 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Creswell JW, Fetters MD, Ivankova NV. Designing A mixed methods study in primary care. Ann Fam Med. 2004;2(1):7–12. doi: 10.1370/afm.104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Guetterman TC, Fetters MD, Creswell JW. Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Ann Fam Med. 2015;13(6):554–561. doi: 10.1370/afm.1865 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ozaki AF, Choi AS, Le QT, et al. Real-world adherence and persistence to direct oral anticoagulants in patients with atrial fibrillation. Circ Cardiovasc Qual Outcomes. 2020;13(3):e005969. doi: 10.1161/CIRCOUTCOMES.119.005969 [DOI] [PubMed] [Google Scholar]
- 24.Bodenheimer T, Lorig K, Holman H, Grumbach K. Patient self-management of chronic disease in primary care. JAMA. 2002;288(19):2469–2475. doi: 10.1001/jama.288.19.2469 [DOI] [PubMed] [Google Scholar]
- 25.Pacleb A, Lowres N, Randall S, Neubeck L, Gallagher R. Adherence to cardiac medications in patients with atrial fibrillation: a pilot study. Heart Lung Circ. 2020;29(7):e131–e139. doi: 10.1016/j.hlc.2019.11.012 [DOI] [PubMed] [Google Scholar]
- 26.Imboden MT, Jones KG, Roth SE, et al. Treatment adherence patterns for patients with atherosclerotic cardiovascular disease: patient and provider perspectives. Prog Cardiovasc Dis. 2025;90:131–141. doi: 10.1016/j.pcad.2025.06.004 [DOI] [PubMed] [Google Scholar]
- 27.Hamilton JE, Blanco E, Selek S, et al. Patient and provider perspectives on medication non-adherence among patients with depression and/or diabetes in diverse community settings – a qualitative analysis. Patient Prefer Adherence. 2022;16:1581–1594. doi: 10.2147/PPA.S328785 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Scott AB, McClure JE. Engaging providers in medication adherence: a health plan case study. Am Health Drug Benefits. 2010;3(6):372–380. [PMC free article] [PubMed] [Google Scholar]
- 29.Kalsi N, Gordon D, Geske J. Identifying and addressing social determinants of health to improve patient-centered care. J Clin Transl Sci. 2024;8(1):e73. doi: 10.1017/cts.2024.511 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Teshale AB, Htun HL, Dalli L, Freak-Poli R. Social Determinants of Cardiovascular Health: understanding their impact and transitioning to a holistic care approach. Eur Heart J. 2024;45(18):1582–1585. doi: 10.1093/eurheartj/ehae012 [DOI] [PubMed] [Google Scholar]
- 31.Schoenbaum AE, Stephenson L, Evans J. Addressing Social Determinants of Health (SDOH) in Healthcare Part 2 | HIMSS. 2023. Available from: https://gkc.himss.org/resources/addressing-social-determinants-health-sdoh-healthcare-part-2. Accessed October 22, 2024.
- 32.Titus S. 7 ways to address financial barriers to medication adherence and contain costs. HFMA; 2019. Available from: https://www.hfma.org/operations-management/cost-reduction/contain-costs-by-addressing-financial-barriers-to-medication-adh/. Accessed July 23, 2024. [Google Scholar]
- 33.Biddle MJ, Moser DK, Pelter MM, Robinson S, Dracup K. Predictors of adherence to self-care in rural patients with heart failure. J Rural Health. 2020;36(1):120–129. doi: 10.1111/jrh.12405 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chen J, Tian Y, Yin M, et al. Relationship between self-efficacy and adherence to self-management and medication among patients with chronic diseases in China: a multicentre cross-sectional study. J Psychosom Res. 2023;164:111105. doi: 10.1016/j.jpsychores.2022.111105 [DOI] [PubMed] [Google Scholar]
- 35.Wang W, Luan W, Zhang Z, Mei Y. Association between medication literacy and medication adherence and the mediating effect of self-efficacy in older people with multimorbidity. BMC Geriatr. 2023;23(1):378. doi: 10.1186/s12877-023-04072-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Self-efficacy for medication management: a systematic review of instruments - PubMed. Available from: https://pubmed.ncbi.nlm.nih.gov/30050290/. Accessed October 22, 2024. [DOI] [PMC free article] [PubMed]
- 37.Martin LR, Feig C, Maksoudian CR, Wysong K, Faasse K. A perspective on nonadherence to drug therapy: psychological barriers and strategies to overcome nonadherence. Patient Prefer Adherence. 2018;12:1527–1535. doi: 10.2147/PPA.S155971 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Farinha JM, Jones ID, Lip GYH. Optimizing adherence and persistence to non-vitamin K antagonist oral anticoagulant therapy in atrial fibrillation. Eur Heart J Suppl. 2022;24(Supplement_A):A42–A55. doi: 10.1093/eurheartj/suab152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Young HN, Len-Rios ME, Brown R, Moreno MM, Cox E. How does patient-provider communication influence adherence to asthma medications? Patient Educ Couns. 2017;100(4):696–702. doi: 10.1016/j.pec.2016.11.022 [DOI] [PubMed] [Google Scholar]
- 40.Althubaiti A. Information bias in health research: definition, pitfalls, and adjustment methods. J Multidiscip Healthc. 2016;9:211–217. doi: 10.2147/JMDH.S104807 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Olson K. Survey participation, nonresponse bias, measurement error bias, and total bias. Public Opin Q. 2006;70(5):737–758. doi: 10.1093/poq/nfl038 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Aggregate data may be made available with reasonable request to the corresponding author.