I. INTRODUCTION

Despite the availability of these antihypertensive medications, global data suggest that less than half of those classified as hypertensive are aware of their problem[3]. Furthermore, less than a third of those who are treated for hypertension get their BPs controlled to currently recommended targets[3]. In Ghana, the prevalence of hypertension ranges from $19\mathrm{\%}$ to $48\mathrm{\%}$ between studies[4]. Old age, over-nutrition and alcohol consumption were some of the factors independently associated with hypertension[4]. According to a 2010 study by Bosu and others, less than one-third of hypertensives were aware of their condition and less than one-tenth had their blood pressures controlled even though there has been a trend towards improved awareness, treatment and control between 1975 and 2005[4]. Due to the asymptomatic nature at onset, diagnosis and adherence to treatment recommendation is often challenging. Treatment of hypertension requires lifestyle modifications and medications or a combination of both.

Adherence is defined as the extent to which a person's behavior corresponds with agreed recommendations from a health care provider[5]. Adherence to therapies is a primary determinant of treatment success and encompasses numerous health-related behaviours that extend beyond taking prescribed medication[5]. It reflects therapeutic behaviours which include seeking timely medical attention, taking prescribed medication appropriately, attending follow-up appointments, being immunized and adopting recommended behavioural modifications. Good outcomes from chronic disease management such as for hypertension, depends largely on the degree of patient adherence to medication and physician advice.

Despite worldwide improvement in diagnosis and treatment options for chronic diseases, a lack of adherence remains an impediment to improving patients' overall quality of life. A study by Criswell and others in 2010 reported that medication self-efficacy and social support were the most important determinants of medication and lifestyle adherence[6]. The same study also found that non-adherent patients experienced a lower level of social support from people outside their families and patients who were adherent to medication-and lifestyle recommendations reported higher medication self-efficacy as compared to non-adherent patients. There are various means of measuring adherence to chronic medication use in clinical practice, including direct and indirect measures. Direct means include the use of direct assays such as laboratory measurement of biological markers or drug metabolites. Indirect means of measurement include pill-counting, patient self-reporting and the use of modern electronic systems that record medication tablet dispensing[7], [8]. In Ghana, a study conducted at the out-patients department (OPD) of the Korle Bu Teaching Hospital amongst 413 patients found an adherence level of $47\mathrm{\%}$ [9]. Another Ghanaian study done in the Komfo Anokye Teaching Hospital between December 2001 and April 2002 found adherence level as low as $7\mathrm{\%}$ [10]. A study conducted in Malaysia observed good adherence in $53.4\mathrm{\%}$ of the 653 patients sampled[11]. It also found female participants to be more likely to adhere to their medication regime, compared to their male counterparts. Congestive heart failure and the presence of other comorbidities also led to higher adherence rates according to the study by Rao Chythra R. et al[11]. Other factors that had the strongest positive effect on adherence included duration of hypertension (the shorter the duration, the better the adherence) and the use of calcium antagonists, and angiotensin-converting enzyme (ACE). Patients taking two or more drugs and those prescribed more expensive drugs were found to be poorly adherent[11]. Good knowledge of the condition as well as of the medications prescribed has been associated with good adherence to medication in several studies[12]. A study conducted in Northern Ireland by Nazli MuzeyyenSencan found that $9.3\mathrm{\%}$ of participants were non-adherent with their antihypertensive medication when assessed using a self-reported adherence scale[13]. The same study also found that $37.9\mathrm{\%}$ of respondents had scores indicative of depressive symptoms, a significant comorbidity. It found age as the only significant predictor of medication adherence in that population. Another study conducted in Pennsylvania found the highest rates of adherence were associated with ACE inhibitors and calcium antagonists, and adherence was significantly higher than with diuretics and beta blockers [14]. The study also found poor adherence to be associated with a higher health care cost. Poor adherence to antihypertensive medication is a multifactorial challenge that affects both the patient and the health care system. As a multifactorial problem, adherence is usually influenced by different contextual factors to varying extents. Some of these factors are at the individual level whereas others may operate at the social and health system level.

In spite of the improvement in diagnosis and treatment options for hypertension, poor adherence remains an impediment to reducing end-organ damage and improving patients' overall quality of life especially in sub-Saharan Africa. This study sought to investigate adherence rates and the local factors that contribute to nonadherence amongst hypertensive patients attending a specialist-led clinic in Kumasi-Ghana in order to improve patient outcome.

II. METHODS

a) Study design and site

This hospital-based cross-sectional study was conducted at the out-patient clinic of County Hospital, a large privately-owned multi-department urban health facility in Kumasi, Ghana. The specialist-led hypertension clinic caters to a diverse population of patients with different personal and socio-demographic characteristics.

b) Study population

The study involved hypertensive patients, aged 18 years and above and who have been enrolled in the Hypertensive Clinic for at least 12 months. The study was conducted between October 2019 and January 2020. County hospital runs two (2) hypertension clinics per week and sees an average of thirty-five (35) patients per clinic day. The hospital runs a counseling, dietary education session and short fitness exercise for patients on every clinic day.

c) Data collection, Processing and Analysis

Data for the study was collected using an electronic questionnaire developed in open data kit (ODK ${}^{\circledR }$ ). The research team collected data directly from patients and medical records. The electronic questionnaire included variables on sociodemographic characteristics, clinical characteristics such as class of drugs taken by the patients and the Morinsky adherence scale (MMAS) to assess adherence of respondents to their medications.

The MMAS-4 is the original four item scale has a reliability score of 0.61 as a measure of internal consistency [15]. The MMAS-4 has been significantly revised since its introduction in 1986 by Morisky DE, Green LW and Levine DM. A higher MMAS-4 score significantly correlated with the presence of a drug metabolite marker[15]. MMAS-8 is a modification of MMAS-4 into an 8-item scoring scale and has a higher reliability score of 0.83. The MMAS-8 has proven reliable in indirectly measuring the medication-taking behavior of patients with chronic diseases such as hypertension and diabetes mellitus[16]–[18]. The MMAS-8 was used as a measure of adherence in this study.

d) Assessment of Medication Adherence

The MMAS-8 questionnaire adopted scoring algorithm, where negative response for each item was coded as 1, except for the question asking if the patient took their medications yesterday (where a positive response was coded as 1). The total MMAS-8 score was calculated by summing the values from all the 8 question items. Adherence was defined as having a MMAS-8 score more than 6 out of a total of 8 scores. Cronbach's alpha test of internal consistency was calculated at 0.79 for the 8 items in MMAS-8 score.

e) Data analysis

The data was analyzed using Stata/SE 14.0 statistical software (StataCorp. 4905 Lakeway Drive Station, Texas 77845, USA). Descriptive statistics was performed for all variables and expressed as means and standard deviation for continuous variables.

Bivariate analysis (logistic regression) was done to measure the strength of the association between socio-demographic, BP level, antihypertensive drug used and medication adherence score. These were presented as crude (unadjusted) Odds ratio. Multivariate logistics regression model was fitted using forward stepwise approach to adjust for the effect of other confounding factors in order to unravel the true factors associated with medication adherence score. The regression models controlled or adjusted for age, gender, educational level, occupation, cigarette smoking and family history of hypertension.

All statistical analysis was done at a $95\mathrm{\%}$ significance level with p values considered as statistically significant. Cronbach's alpha test of internal consistency was calculated at 0.79 for the 8 items in MMAS-8 score.

f) Operational definitions

Three seated Blood Pressure (BP) measurements were obtained at 5-minute intervals from each participant. An average was calculated for each of the three systolic and diastolic measurements taken. Hypertension was defined as an average systolic BP (SBP) ≥140 mm Hg and/or an average diastolic BP (DBP) ≥ 90 mm Hg. Controlled hypertension was defined as having an average SBP <140 mm Hg and/or an average DBP < 90 mm Hg, whilst on medication.

III. ETHICAL CONSIDERATIONS

Participation Informed Consent Forms (ICF) were designed based on the principles of Good Clinical Practice (GCP). The content of the ICF were clearly explained to the understanding of the potential participants. Literate participants were allowed to write and sign the ICF while non-literate participants thumb printed their consent, assisted by a third party, preferably an accompanying relative (as witness) who countersigned the ICF. Ethical approval was obtained from the Committee on Human Research Publication and Ethics (CHRPE) from the Kwame Nkrumah University of Science and Technology after administrative approval from County Hospital.

This study posed minimal or no anticipated risks to participants since it was a non-invasive study, with no identifying information collected. There was no cost to the participant nor was there compensation to participate in this study

IV. RESULTS

a) Sociodemographic of the hypertensive patients

The study involved 361 patients who consented to participate in the study. All participants were included in the analysis of the study. Out of this, 296 (81.99%) were females and 65 (18.01%) were males, giving a female-male ratio of 4.6:1. More than half of the patients (n=219; 60.66%) were over 55 years old with a mean age of 59.74. The majority (n=296; 81.99%) of respondents had some level of education but only 11% had achieved tertiary education. Further details of participant socio-demographics can be found in Table 1.

b) Clinical Characteristics of Participants

Out of the 361 hypertensive patients, more than half 234 (64.32%) reported a family history of hypertension. Almost half (n=167; 46.26%) of participants reported a comorbidity, out of which 82.63% were Diabetic. A little over three quarters (n=272;

More than half (n=207; 57.34%) were adherent (MMAS-8 score 6 to 8) to antihypertensive medication prescribed (Table 2).

c) Class of Antihypertensive Drugs used among the Hypertensive Patients

The most common class of drugs prescribed for participants were calcium channel blockers (n=244;

67.59%) and angiotensin receptor blockers (n=238; 65.93%). Other classes of medications prescribed included statins (30.47%), beta-blockers (26.04%), ACE inhibitors (10.53%) and diuretics (8.31%). (Table 3).

Bivariate analysis of medication adherence and patient demographic factors, clinical characteristics and drug classes taken was performed. Significant association was demonstrated between age of patient, existence of other comorbidities and certain classes of antihypertensive medications. Older patients (> 55 years) were more likely (aOR=2.74, CI=1.60- 4.68;

V. DISCUSSION

a) Blood pressure control

Blood pressure control amongst respondents was optimum. Most of our study participants had their blood pressures controlled with a mean systolic and diastolic of 125.25mmHg and 75.54mmHg amongst hypertensives who had been on treatment for at least a year. This is in contrast to a systematic review by Bosu and colleagues which found that less than ten percent (10%) of hypertensives had their blood pressures controlled according to most Ghanaian studies[4]. This could be due to the effective management by cardiology specialists and the relatively high adherence of most patients to treatment recommendations. Our study site also runs a counselling, dietary session and short fitness exercise for patients on every clinic day. Also, our study was conducted in an urban setting and this might contribute to the better control of hypertension similar to a study by Chow and others which found increased awareness, treatment, and control of hypertension in urban communities compared to rural communities[3]. In addition, this could it be due to the nature of people who may opt for care in private clinics, they may have better socioeconomic profile, be better motivated or may be more likely to have better support systems.

b) Adherence rates

The overall adherence rate found in this study falls within 52% to 92% reported from western studies[19]. It is similar to that reported by an Ethiopian study [20] even though it is considerably less than the adherence rate of 82.2% reported by a Malaysian study[21]. Previous studies conducted on adherence in Ghana found lower rates of 47% in Korle Bu[9] and 7% in Komfo Anokye Teaching Hospital[10]. The rates of good adherence to antihypertensive medications may vary due to a host of factors such as the study design and population, method of adherence measurements, biases, scoring systems used etc. Population wise, adherence rates may vary due to perception of orthodox medications, poverty, cost of medications, insurance policies, self-care attitudes and social support systems in place. Our study site is a specialist led clinic and this might explain the reason for a high adherence rate comparable to that of developed countries. Our clinic setting employs a patient-centered care approach which focuses on true partnership between patients and healthcare staff. It involves counselling and education on hypertension and its end-organ damage, diet and exercise. In addition, patient communal coping in our clinic setting might also be a contributing factor to the relatively high adherence rate observed in this study.

c) Factors affecting medication adherence

Worldwide, there are several studies that have associated several factors to medication adherence. Medication adherence has been associated with a host of factors such as demography, psychological factors as well as disease and medication-taking behaviors[22]. Based on the results of our study, adherence to antihypertensive medication is influenced by factors such as older age, significant comorbidities and the use of calcium channel blockers.

As found in other studies[22], younger adults had lower adherence rates compared to older adults. Older people are more accepting of their diagnosis and hence may be more inclined to follow through with their medications. Also, younger people have more distractions in terms of work and meeting responsibilities compared to those over 55 who may be retired, have less work commitments or may be settled into a more routine life.

The presence of other comorbidities such as atrial fibrillation, diabetes mellitus and glaucoma were found to be significantly associated with poorer adherence amongst hypertensives. Comorbidities can lead to a higher pill burden which can negatively influence adherence to medications. This finding is however in contrast to studies by Rao Chythra R. and others who found congestive heart failure and the presence of other comorbidities led to higher adherence rates $[11]$. Perhaps, one can reason that the presence of severe life-threatening comorbidities tend to increase reliance on medications for survival and hence may make patients somewhat more adherent. The conflicting evidence with regards to comorbidities and medication adherence suggests a more nuanced relationship that requires further contextual examination.

The use of calcium channel blockers was associated with lower rates of adherence amongst hypertensive patients. This could be as a result of the side effects of pedal edema and headaches. This was also in contrast with a Pennsylvanian study which found the highest rates of adherence were associated with ACE inhibitors and calcium antagonists, significantly higher than with diuretics and beta blockers[14].

VI. CONCLUSION

From our study, blood pressure control among respondents at the specialist led clinic was optimum. Our study also found a relatively high adherence rate to medications which was positively associated with older age, and negatively associated with factors such as comorbidities and the use of calcium channel blockers. The relatively high level of adherence found in this study can be attributed to our specialist-led setting which uses a patient-centered care approach involving adequate counseling and education of patients and communal coping mechanisms used by patients. The high prevalence of BP control found in this study can be linked to the relatively high level of adherence found amongst the participants. Better adherence leads to better control which ultimately leads to improved outcomes, prolonged survival and reduction in the incidence of end organ damage from hypertension. The extent to which such adherence is influenced by attendance by specialist physicians in the local context should be the subject of further enquiry.

What is already known about this topic

• Worldwide and particularly in sub-Saharan Africa there is an increasing prevalence of hypertension and its complications;

• In spite of increasing availability of modern treatment options for hypertension, adherence to treatment modalities remain an impediment to the control of hypertension;

• Poor adherence to antihypertensive medications is multifactorial and affects not only the patient but the entire health care system.

What this study adds

• A relatively high level of adherence was found amongst hypertensive patients being managed at a specialist-led clinic

• Specialized-clinics that adopt a holistic and patient-centered care approach such as counseling on diet, exercise and complications of hypertension, may be an important factor in ensuring adherence

• Relatively high levels of adherence to antihypertensives correlate with achieving optimum blood pressure targets, which could in turn lead to improved patient outcomes.

ACKNOWLEDGEMENTS

The authors would like to send our sincerest appreciation to the hospital management of County hospital in Kumasi Ghana for granting us permission to conduct the study in their facility. We would also like to thank everyone who has contributed to this study, mention can be made of Mr Kaba AdoaChrisdo, Dr Gibil Reuben, Dr Emmanuel Opoku-Mensah and Dr Prince Samuel Osei.

Conflict of interest

All authors declare no conflict of interest

Author's contribution

All authors contributed in idea generational, data collection, analysis, writing and proofreading of this case report and are in agreement with its content before submission.