According to the most recently available report of the Centers for Disease Control and Prevention of the United States (U.S.) Department of Health and Human Services (2012), approximately 117 million people in the U.S. suffer from at least one chronic illness (Centers for Disease Control and Prevention, 2016). Chronic illness is defined as a medical condition that persists longer than three months and includes persistent and recurring health problems (e.g., heart disease, stroke, cancer, diabetes, arthritis, and chronic pain issues) (Goodman, Posner, Huang, Parekh, & Koh, 2013). Medications are often used to manage and treat these diseases.

Medication adherence is defined as the patient’s ability to effectively manage... all medications prescribed by a healthcare provider. Medication adherence is defined as the patient’s ability to effectively manage timing, dosing, and frequency of all medications prescribed by a healthcare provider (Brown & Bussell, 2011; Viswanathan et al., 2012). Adhering to a prescribed course of medication, however, is often difficult for patients who take multiple prescription medications with varying dosage times. When patients have difficulty taking daily medications as prescribed, health complications can occur and may even lead to costly, distressing, and unnecessary recurrent hospitalizations (Ho et al., 2007; Krousel-Wood,, Islam, Webber, Re, Morisky, & Munter, 2009). Patients who are able to adhere to their prescribed course of medications have the best opportunity for optimal outcomes (Bosworth et al., 2011).

Patient education does not guarantee medication adherence. Nurses play a key role in educating patients about their medications and complex regimens. Patient education does not guarantee medication adherence. The prudent nurse knows that there are many factors that support optimal outcomes, including patient characteristics; complexity of the regimen; communication barriers; coordination of care by multiple providers; lack of communication about potential adverse effects; lack of health information technology (Brown & Bussell, 2011) as well as patient beliefs about medication (Al Hewiti, 2014).

In this article, we report on the current challenge of medication adherence in a specific population. We emphasize the need for nurses to remain aware of patients who do not adhere to their medication regimen and to work with these patients to enhance medication adherence. Next, we report how our study explored the prevalence of medication adherence among patients in a cardiology ambulatory care setting. We share and discuss our results and study limitations and offer implications for practice.

The Medication Adherence Challenge

The challenge of medication adherence among patients with chronic illness is widespread and has been well-studied (Bosworth et al., 2011). Researchers who have studied people living with various chronic diseases have reported that approximately 50%, nationwide, do not adhere to their medication regimen (Brown & Bussell, 2011). While there may be many barriers to medication regimen adherence, often it is the complexity and confusion of taking multiple medications with varying dosage times that contributes to non-adherence (Caldeira, Vaz-Carneiro, & Costa, 2014; George, Phun, Bailey, Kong, & Stewart, 2004).

Many remedies to medication regimen non-adherence have been proposed. Many remedies to medication regimen non-adherence have been proposed (Mendys, Zullig, Burkholder, Granger, & Bosworth, 2014). Caldeira, et al. (2014), in reporting a meta-analysis, noted that reducing the actual dosing time of medications to once daily influenced adherence and supported a 56% decrease in non-adherence. Davis, Packard and Jackevicius (2014) recommended pharmacy support at discharge to help improve adherence. A number of studies have indicated that age may be a factor in medication adherence. Gellad, Grenard, and Marcum (2011) found older adults have the greatest difficulty with adhering to their medication regimen, so age-appropriate assistance with medication adherence may be needed.

Previous research reports have indicated that medication non-adherence among patients with chronic cardiac illness is similar to that of patients with other chronic illnesses (Ho et al., 2008). We conducted a review of the literature noted in the Cumulative Index of Nursing and Allied Health Literature (CINAHL) and PubMed from 2010 to 2017, using the terms medication non-adherence, medication compliance, cardiovascular disease, clinic, outpatient, ambulatory, and health center. We identified only two articles that reported on the extent of medication non-adherence among patients seen in outpatient cardiology clinics in the U. S. (Dunlay, Eveleth, Shah, McMillan, & Roger, 2011; Zullig, McCant, Melnyk, Danus, & Bosworth, 2014).The purpose of the following study was to explore medication adherence in patients who were being treated at a large, ambulatory, suburban, cardiology practice. Determining the scope of adherence to a prescribed medication regimen in this population is necessary to develop effective interventions that could be useful to help these patients adhere to their medication regimen.

Research Method

This section will describe the study design; setting and sample; measures and outcomes; and the data collection process. We will also address validity and reliability.

Design
This study used a descriptive, correlational and comparative design to answer the following research questions: In patients who had been diagnosed with at least one chronic illness and who were prescribed four or more different routine medications each day:

(a) What was the prevalence of adherence to the prescribed medication regimen as measured with the Morisky Medication Adherence Scale (MMAS-8^©)?
(b) Was there a difference in adherence to the prescribed medication regimen based on age, gender, education level, and level of home assistance?
(c) Was there a difference in adherence based on the complexity of their medication regimen according to the Medication Regimen Complexity Index (MRCI)?

Setting and Sample
The setting for this study was a large, suburban, cardiology ambulatory care setting with approximately 10,000 patient visits per year. Inclusion criteria consisted of adults who were able to hear, read, and speak English; prescribed a minimum of four daily medications; and able to complete a paper survey and information sheet or willing to allow a researcher to administer the survey and complete the demographic information sheet. Patients were excluded from participation if they had a history of dementia or mental health changes that would prevent them from completing the survey. Of the 634 patients invited to participate, 132 did not meet inclusion criteria; 197 declined to participate; and one did not complete the survey, leaving 304 subjects for analysis. The characteristics of the final sample are listed in Table 1.

Table 1.

Sample characteristics (N = 304)

Measures and Outcomes
The outcome variables measured in this study included adherence to prescribed medication regimen, complexity of the medication regimen, and demographic information. Each of these variables is described below.

1) Adherence to prescribed medication regimen was measured using the Morisky Medication Adherence Scale (MMAS-8^©), an 8-item questionnaire with a demonstrated internal consistency (Cronbach’s alpha) of 0.83 (Morisky, Ang, Krousel'Wood, & Ward, 2008). The MMAS-8^© was designed to measure both intentional and unintentional medication adherence, including forgetting, and not taking medications if feeling worse or because of feeling better. Questions were specifically worded to capture these circumstances (Morisky et al., 2008).

2) Complexity of the medication regimen was measured with the Medication Regimen Complexity Index (MRCI), a 65-item, valid tool that scores the level of medication complexity based on the multiple variables, including: number of individual medications; dosing frequency; additional directions for medication use (e.g., crush or dissolve medication, taper doses, take with food); and route of administration. The MRCI includes three opportunities for positive scoring.

• The first addresses the route (e.g., oral, topical, ear, eye, nose, inhalation, with each weighing on a score of one to five).
• The second takes into account dosing frequency in which a number is assigned to the times a medication is prescribed each day. Points ranging from 0.5 to 12.5 are assigned to varying times each medication is prescribed to be taken daily. For example if a medication is ordered daily as needed (PRN), the score is lower than if ordered three times daily PRN.
• The third component of the MRCI takes into consideration 10 types of additional dosing instructions and adds one to two points for such conditions as medication needing to be crushed or dissolved, or dose tapering.

The MRCI has inter-rater and test-retest reliability for the total score and for scores of individual sections ≥0.9 and also significant correlation with the number of drugs in the regimen, Spearman’s Rho = 0.9; p < 0.0001 (George et al., 2004).

3) Demographic information gathered included age, gender, education level, and level of home assistance four or more days a week. Assistance varied from none; spouse or live-in family assistance; family or friend assistance but not live-in; and hired assistance including self-pay and covered by health plan. This information was important because other studies have demonstrated that medication adherence may be influenced by these variables (Crowley et al., 2014; Roth & Ivey, 2005).

Data Collection
Following Institutional Review Board (IRB) approval, subjects were recruited from the cardiology practice ambulatory care setting. Those who provided written consent to be in the study completed a brief demographic information sheet and the MMAS-8^©. Each participant’s medication regimen complexity score was calculated using the MRCI and based on information abstracted from records immediately following the outpatient visit at which the most current medication records had been reviewed and updated with the patient. Study data were collected and managed by the research team using the Research Electronic Data Capture (REDCap) electronic data tool (Harris et al., 2009), a secure, web-based application designed to support data capture for research studies.

Validity and Reliability
A sample size of 300 was needed to achieve a power of 0.80, therefore the size of the sample collected was adequate for this study (N=304) and contributed to its validity. An acceptable level of reliability on the MMAS-8^© scale was achieved with an internal consistency Cronbach’s alpha score of 0.60 for the total test.

Results

In this section, we will first describe our data analysis process. Then we will share our findings.

Data Analysis
When data collection was complete, all data (including demographic data, MMAS-8^© answers, and the scores calculated from the MRCI) were cleaned, double checked, and entered into Statistical Package for Social Sciences (SPSS) version 21 for analysis. There was very little missing data (total of 6 unanswered items) and those items were handled through the use of Expectation Maximization (EM) rather than deleting missing data and cases (Musil, Warner, Yobas, & Jones, 2002). Data were determined to be normally distributed; thus parametric testing was used to answer the research questions. Comparisons were made using Pearson’s correlation coefficient; the t-test was used to determine the differences between the variables. Data re-grouped into categories for analysis were analyzed using non-parametric tests.

Findings
The first research question asked, “What was the prevalence of adherence to the prescribed medication regimen as measured with the Morisky Medication Adherence Scale (MMAS-8^©)?” The MMAS-8^© adherence score categories are: extremely low adherence (0 to 3.99), low adherence (4.00-5.99), medium adherence (6.00 to 7.99), and high adherence (8). The Figure illustrates that scores were generally low with 12% of subjects (36) scoring in the extremely low adherence category and 74% of subjects (224) scoring in the low adherence category. Only seven subjects (2%) scored high adherence.

Figure. MMAS-8 Score prevalence in subjects

Table 2 provides the questions asked on the MMAS-8^©, the number of respondents that responded affirmatively, and the mean score for each question. The total mean score of 5.903 for all subjects demonstrates overall low adherence.

Table 2. Prevalence of adherence to the prescribed medication regimen as measured with the Morisky Medication Adherence Scale (MMAS-8^©-Item) (N = 304)

Note: Permission to use the ©MMAS was obtained from the author, Dr. Donald Morisky. Use of the MMAS^© is protected by U.S. copyright laws. Permission for use is required. A license agreement is available from: Donald E. Morisky ScD, ScM, MSPH, Professor, Department of Community Health Sciences, UCLA School of Public Health, 650 Charles E. Young Drive South, Los Angeles, CA 90095-1772.

The second research question asked, “Was there a difference in adherence to the prescribed medication regimen based on age, gender, education level, and level of home assistance?" There was no significant correlation between the total MMAS-8^© adherence score and age, gender, education level, and level of home assistance. However, a statistically significant inverse correlation was found between the responses to the question “How often do you have difficulty remembering to take all your medications?” and age (p < 0.01) with those greater than age 65 answering rarely or never having difficulty remembering to take their medications.

...a more complex medication regimen is more likely to be viewed as an inconvenience or hassle. The final research question asked, “Was there a difference in adherence scores based on the complexity of their medication regimen according to the Medication Regimen Complexity Index (MRCI)?” There was no statistically significant correlation between the total MRCI score and the total MMAS-8^© adherence score, even though the mean number of medications prescribed per day for the sample was very high at 10.35 (range of medications per day, 4 – 31; SD=4.41). However, a statistically significant positive relationship existed between the responses to the MMAS-8^© question “Do you ever feel hassled about sticking to your treatment plan?” and the total MRCI score (p < 0.05) indicating that a more complex medication regimen is more likely to be viewed as an inconvenience or hassle. A higher MRCI score was also positively and significantly correlated with the total number of medications a person was taking (p < 0.01), and the level of home care assistance that the participant received (p < 0.05). The variables with significant relationships are listed on Table 3.

Table 3. Correlations between total MMAS-8^©-Item score, and difficulty remembering, age, education level, presence of home health assistance, total meds taken, and medication complexity index score (N = 304)

p < .05* (two-tailed)

Discussion

Analyses of the data revealed some findings that were expected and consistent with previous research conducted on medication adherence in patients with chronic illness. However, the overall extremely low MMAS-8^© scores (85.5% scored ‘‘low adherence,’ i.e. a score below 6) indicating very poor adherence, was remarkable. Due to the personal experiences of the research team members and anecdotal discussions among healthcare providers at the site where the research was conducted, we expected the scores of non-adherence to be problematic but closer to the 50-56% as described in other studies (Brown & Bussell, 2011; Al Hewiti, 2014). Thus, the low adherence scores in this study were surprising. These scores indicated that the problem of medication non-adherence in this study population was greater than that reported in previous research and was a source of great concern to our team members.

Another surprising result was the significant inverse correlation found between age and difficulty remembering to take medications. Another surprising result was the significant inverse correlation found between age and difficulty remembering to take medications. Although there was no significant relationship between age and overall MMAS-8^© scores, those who reported ‘usually’ having difficulty remembering to take their medications were younger (< 65 years). Patients over the age of 65 in this study often reported ‘rarely or never’ having difficulty remembering to take their medications, as presented on Table 2, number 8.

Because older adults make up a large part of the population in the outpatient cardiology practice used for this study, it was expected that the problem of remembering to take medication would likely be higher among adults as they got older. The findings in this study contradict some previous study findings that demonstrate that patients following their seventh decade may become less adherent as regimens become more complex, cognition and physical status may decline, and support mechanisms dwindle (Krueger, Berger, &Felkey, 2005). While it is typically assumed that older adults would be more forgetful and less adherent with their medication regimen as they aged, this study found that older patients actually reported being more likely to remember to take their medications. However, it is important to keep in mind that in this study, medication adherence was measured by self-report; patients were not enrolled if limitations from cognitive ability as found in dementia could have affected ability to answer questions, or sensory limitations hindered communication of the responses to the MMAS-8^©. One explanation for this finding may have been that this older population in a suburban cardiology practice may have had adequate resources to obtain their medications and adequate support systems in place to promote higher levels of adherence.

We discussed several other possible explanations for this phenomenon of younger patients reporting ‘usually’ having difficulty remembering to take their medications. We speculated that younger populations may have more distractions including technology overload, careers, child-rearing responsibilities, and caring for aging parents. Balancing a fast-paced, stressful lifestyle leaves less time to focus and manage healthcare regimen and medication management. Hence it could be that older adults, without these other time constraints, are less likely to forget their medications. More research is needed to confirm/disconfirm such explanations.

Another explanation is proposed by Vandermause and colleagues (2016) who reported the findings of a recent qualitative study that looked at the medication-taking practices of older adults with multiple chronic medical diagnoses. These researchers found that adding a new prescription to an older adult’s regimen could be seen as a sign of impending mortality, adding a burden to their own health management. In the current study, it may be that the older adult subjects are like those interviewed by Vandermause and colleagues: their chronic conditions and numerous prescriptions have a way of reminding them of their own limitations, providing an incentive to take their medications for self-preservation (Vandermause et al., 2016).

The high correlation found between younger patients and difficulty remembering to take their medications corroborates recent findings from other studies that have investigated medication adherence among populations of different ages and education levels (Al Hewiti, 2014; Wheeler, Roberts, & Neiheisel, 2014). In one study, when age groups were separated into quartiles, the lower-age group comprised of subjects 18-49 years old were less adherent (Rolnick, Pawloski, Hedblom, Asche, & Bruzek, 2013). In a prior study conducted by Horne and Weinman (1999), younger age was also found to be associated with lower medication adherence, suggesting this phenomenon may be population specific.

Additional research is necessary to determine interventions that will be effective for these population groups. The assumption that those with more advanced education have a higher level of medication adherence was not supported in this study. Some have suggested behavioral- and/or educational-focused interventions as approaches to promote medication adherence (Krousel-Wood, Munter, et al., 2009). Although simplifying medication regimens by changing medication dosages to once daily may be effective in increasing medication adherence in some populations, few other recommendations or strategies have been successful in addressing medication adherence at any age (Bosworth et al., 2011).

Future research could focus on examining different methods, such as inclusion of technology to improve adherence to the prescribed medication regimen... Future research could focus on examining different methods, such as inclusion of technology to improve adherence to the prescribed medication regimen and help patients to remember to take their medications. Currently, more than 66.6% of the world population has access to a cell phone (Free et al., 2013), making text messaging available through the U.S. Centers for Disease Control and Prevention a viable aid in health management (Free et al., 2013). Other solutions would need to be explored to aid persons who do not have access to these devices. Longitudinal studies examining medication adherence of patients across decades would provide insight as to the importance of long-term medication adherence (Morisky & DiMatteo, 2011).

Limitations

This study included only the population from one large, suburban, cardiology ambulatory care practice. It may be that patients seen in this practice were a homogenous group, all living in the same geographical area. If replicated, other studies should include a population sample from a wide variety of demographic areas in order for findings to be more generalizable.

Implications for Practice

...non-adherence to one’s medication regimen remains a problem for those with a chronic illnesses... The findings of this study contribute evidence that non-adherence to one’s medication regimen remains a problem for those with a chronic illnesses, such as a cardiac illnesses. The inverse relationship between age and remembering to take one’s medication must be explored. This knowledge can be used to create more effective, tailored adherence interventions.

Nurses who provide medication instructions must not take for granted that younger patients understand and will adhere to their prescribed medication regimen. Although not a topic for this study, subjects often voluntarily shared with the data collectors their methods for remembering to take their medications. When taking multiple medications, many reported using pill box organizers and relied on reminders from family members to facilitate their adherence. There is likely no single solution to address medication non-adherence for all patients. An individual, patient-focused approach is necessary to improve medication adherence (Krousel-Wood, Munter, et al., 2009).

Conclusion

There is likely no single solution to address medication non-adherence for all patients. It is important for clinicians practicing in the ambulatory care setting to understand that the rate of medication non-adherence remains high. It is important that patient age and lifestyle be addressed during any medication adherence instruction. Exploring self-reported methods for successful adherence and age would provide insight into a more tailored approach to age-related interventions.

It is not only older patients with whom nurses need be concerned, but younger populations as well. Gearing interventions for adherence may include user-friendly electronic devices and other techniques easily accessible and popular among today’s youth (Becker et al., 2015). The results of this study provide the foundation for a future intervention study using novel means to individually assist patients with a complex medication regimen to comply with their prescribed course of treatment.

Acknowledgment:

Research funded in part by an internal grant, Nursing Research Fund Award (NURF) 2014, Cleveland Clinic Institute of Nursing. Special thanks to the staff of the Mayfield Cardiology practice, Mayfield Heights, Ohio, a division of the Heart and Vascular Institute of the Cleveland Clinic, Cleveland, Ohio, for their cooperation in making this study possible. Additionally, the research team would like to thank Dr. Donald Morisky for allowing them the complimentary use of the Morisky Medication Adherence Scale^© (MMAS-8-Item).

Authors

Jayne Rosenberger, BSN, RN, CCRN
Email: jrosenbe@ccf.org

Ms. Rosenberger, the principal investigator in this research study, is part of the Nursing Institute at the Cleveland Clinic Hillcrest Hospital (Mayfield Heights, OH). She is a hospital-based, clinical nurse with 36 years of direct patient care experience, including the provision of care in the critical care, emergency room, medical surgical, oncology, and postpartum/newborn nursery areas. She holds certification in critical care nursing and has served on the Hospital Nursing Research Council for the past three years. Ms. Rosenberger currently works with hospitalized patients in critical care and the emergency room, formerly with newly admitted hospitalized patients from the cardiology practice involved in this research study in which her duties included daily assessments and review of current medications of patients in this practice. This review included asking patients whether or not they were taking their medications prior to their current hospital stay. The discrepancies that she observed regarding medication adherence during these interviews prompted her investigation of medication non-adherence in this specific population.

Esther Bernhofer, PhD, RN-BC, CPE
Email: bernhoe@ccf.org

Dr. Bernhofer is a Senior Nurse Researcher at the Cleveland Clinic (Cleveland, OH) in the Office of Nursing Research and Innovation. In addition to her own program of research, she mentors clinical nurses as they conduct and disseminate their research projects. Prior to her current position, Dr. Bernhofer worked as a medical/surgical registered nurse for 25 years, after which time she worked as a pain education specialist. Her additional research interest includes patient-centered outcomes, such as medication management.

Susan McCrudden, BSN, RN
Email: smccrudd@ccf.org

Ms. McCrudden, a co-investigator on this research study, is a hospital-based clinical nurse with 35 years of clinical experience. She currently works in the Nursing Institute at the Cleveland Clinic Hillcrest Hospital (Mayfield Heights, OH), specializing in critical care bedside nursing in the intensive care unit. She has had prior experience in the areas of emergency, medical surgical, and school nursing. Her involvement in shared governance and hospital research councils has led to additional research opportunities. One such opportunity included serving as co-investigator for a study that explored differences between arms in blood pressure measurements in patients admitted to a critical care unit. Additionally Ms. McCrudden has published a chapter describing the building and sustaining of a hospital-based nursing research program.

Rosa Johnson, MS, RN-BC, BA, BSN, OCN
Email: rjohnso1@ccf.org

Ms. Johnson, a hospital-based clinical nurse with 50 years of nursing experience, currently works in the Nursing Institute at the Cleveland Clinic Hillcrest Hospital (Mayfield Heights, OH). She holds a Master of Science in Management, a Bachelor of Arts in Communications, and a Bachelor of Science in Nursing. She is board certified in gerontology and oncology nursing. Ms. Johnson became involved in this research through her work on the hospital shared governance council, and served as a co-investigator this research study. She is an adjunct clinical instructor for Ursuline College and works at Hillcrest Hospital on the medical-surgical and telemetry units. Previously Ms. Johnson has held positions as a Nurse Manager, an Assistant Nurse Manager in oncology, and as a Staff Nurse in various nursing specialties of nursing. She has enjoyed bringing her clinical and academic expertise to this investigation.