Factors Associated with Medication Adherence among hypertensive Patients in a Tertiary Health Center: A Cross-Sectional Study

Introduction: Medication non-adherence is a major hindrance in the treatment of hypertension in Sub Saharan Africa. It is a major modifi able contributor to poor blood pressure control and complications of the disorder. An understanding of the factors that are associated with drug adherence in hypertension will contribute positively to the overall planning of public health educational programs on hypertension.


Introduction
Medication non-adherence is a major hindrance in the treatment of hypertension in Sub Saharan Africa (SSA) [1][2][3][4][5][6]. It is a major modifi able contributor to poor blood pressure control and complications of the disorder [7]. Other consequences of non-adherence include increased fi nancial burden (due to increased rate of hospitalization and loss of productivity), unwarranted change of medication, substantial negative effect on patients' quality of life and drop-out of treatment [7,8].
Non-adherence results from the complex interplay of several factors which may be grouped into patient-centered factors, therapy-related factors, healthcare system factors, social and economic factors, and disease related factors [9].
Currently, there are no data on adherence to anti-hypertensive medications in the South-East Nigeria. An understanding of the factors that are associated with drug adherence in hypertension will contribute positively to the overall planning of public health educational programs on hypertension.
The main aim of this study was to determine the prevalence of medication non-adherence among hypertensive patients and its correlates among patients attending medical outpatient clinic in Enugu south East Nigeria.

Setting
This study was cross sectional and descriptive in nature.

Study design
A semi structured questionnaire was used to collect data on selected socio-demographic characteristics and lifestyle behaviors including smoking, drinking and use of herbal drugs.
Measurements of weight, height were carried out and recorded.
Three 3 blood pressure measurements over the last 3 clinic visits were averaged. Fasting blood glucose was also recorded.
Minimal sample size was calculated using the formula n =Z 2 (pq)/e 2 where n = required sample size, p = 0.495 (non adherence rate reported in Nigeria), q =1-p, e = desired confi dence interval. The minimum sample size of 384 was selected.

Defi nition of terms
Hypertension was defi ned as a systolic blood pressure (SBP) of ≥140 mmHg and/or diastolic blood pressure (DBP) of ≥90 mmHg and/or use of an anti-hypertensive drug therapy or based on medical records of the subjects. Medical co-morbidities were defi ned using standard criteria or past medical history diagnosed by a qualifi ed personnel (doctors).
Level of education was the individual's highest educational (formal) attainment. Tobacco use was defi ned as the use of (any or all) cigarettes, snuff and chewing tobacco in the past 4 weeks. Alcohol use was defi ned as the consumption of any alcoholic beverage (beer, gin, stout, local brew) (greater than 14 units for females and 21 units for males in a week. Occupation was defi ned as the primary job which takes at least 50% of the working hours in a week. An artisan was defi ned as skilled manual labourers such as masons, mechanics, tailors, welders, metal workers and other crafts.

Study instruments
Anxiety and depressive symptoms were explored using the Hospital Anxiety and Depression Score questionnaire (HADS) [13]. The HADS questionnaire is a self-assessment scale Level of Physical disability was estimated using the modifi ed Barthel Index of Activities of Daily Living (BADL) score [14].
BADL score was used to measure functional disability by quantifying patient's performance in 10 activities of daily life.
These activities were grouped according to self-care (feeding, grooming, bathing, dressing, bowel and bladder care, and toilet use) and mobility (ambulation, transfers, and stair climbing) depending on what the patient was actually able to do. Sum the patient's scores for each item. Total possible scores range from 0-100, with lower scores indicating increased disability. Direct testing of the patient was done or with the help of their care giver when necessary.
The Morinsky-Green Medication adherence scale [15] was used to estimate medication adherence. Answers consistent with adherence were scored as 0 and answers consistent with non-adherence were scored as 1. For the purpose of the index study the scores are tallied and graded into high adherence (0-2), medium adherence (3)(4)(5) and low adherence (6)(7)(8). Medium and low adherence were grouped as 'non-adherent' while 0-2 and 'high-adherence'.

Statistical Methods
For database management and statistical analyses, we Mean values were compared using the independent t-test. In all, p value of < 0.05 was regarded as statistically signifi cant.
Conclusions were drawn at 95% confi dence interval.

Results
A total of 436 cases (males 161(36.9%), females 275 (63.1%)) of hypertension (hypertensives) were surveyed in this study. The male to female ratio was 1:1.7. The ages ranged from 26 years to 95 years with a mean of 59.8 ± 12.8. The peak age was the 7th decade followed by the 6 th decade ( Table 1) women, 86(19.7%) and 85(19.5%) were retired and civil/public servants/offi ce workers. Most of the patients had a tertiary school education (completed more than 12 years of formal education) and came from within the city Table 1.

Behavioral risk factors Medical History and HADS scores
Means of weight and height measurements are shown in table 1. The mean (sd) BMI was 28.7(6.6) kg/m 2 signifi cantly higher in females than males. P=0.01. The mean fasting blood glucose (160.2 mg/dL) was equally high in males (154.3mg/ dL) and females 171(mg/dL). P=0.13. Figure 1. More males (26.1%) than females (9.1%) of the females reported current use of tobacco. P<0.001. Overall, 72% used herbal remedies in the last 12 months, 67% were taking NSAIDs and 28.7% drank alcohol at least occasionally. Table 1. There were no statistical signifi cant differences between mean depression and anxiety scores of males and females.

Blood pressure
The mean Systolic blood pressure (SBP) and diastolic blood pressure (DBP) are shown in table 1. There were no statistical differences between the mean SBP and DBP between males and females. Blood pressure control was achieved only in 17.2% (18.6% in males and 16.4% in females, p=0.54) of the patients Figure 1b.

Co-morbidity and barthel index of activities of daily living
During the survey a total of 131 (30%) participants had one co-morbidity and 132 (30.3%) had two co-morbidities Table 1. The frequency of various comorbidities is shown in table 2. The commonest co-morbidities were arthritis (48.6%), diabetes (42.4%) and headache (40.6%). Stroke and Parkinson's disease were more frequently found in males than females Table 2. Headache was slightly higher in females. P= 0.06. Sixty-nine patients (15.8%) reported a total of about 234 (53.7%) areas of disability giving an average of 3.4 per individual. Disability was reported more in males. P=0.04. Table 3 shows the responses based on the Morinsky-Green adherence scale. Most of the patients (90.1%) sometimes forget to take their medications or do not bring their medications along when they leave home (94.3%). About 88.1% cut back or stop taking their medications in the past without telling their doctors. However, 82.8% agreed that they took their medication the previous day. Just more than half said that forgetfulness is never/rarely a problem to medication adherence. Nonadherence defi ned as medium and low adherence in this study was 34.6% (151/436). The highest rates of non-adherence were reported in patients who were totally dependent (62.5%) although the total number of patients in this category was small. Civil servants/offi ce workers, artisans and pensioners had rates of non-adherence higher than 34.6% average found in the study.

Adherence
Out of the comorbidities considered peptic ulcer patients were the least adherent to medications (30.4%). p=0.02. Alcohol users were also non-adherent to medications Table 3.
The age distribution of non-adherence is shown in Table 4 and

Regression analysis
In bivariate correlation analysis (Table 5)

Discussion
Non-adherence to medication is determined by several factors and potentially a modifi able risk factor for cardiovascular complications of hypertension [1][2][3][4][5][6][7]. Improving medication adherence has the potential to reduce the economic burden of hypertension in an already impoverished region of the world.
In the index study, the rate of non-adherence was 34.7%, signifi cantly higher in patients with higher depression scores in HADS, lower BADL scores (higher disability), current alcohol users and those with peptic ulcer disease. High HADS-D scores, low BADL scores and presence of peptic ulcer diseases not only correlated with adherence scores but were also signifi cant predictors.
The rate of non-adherence (34.7%) among hypertensives in the index study was lower than 66.7% reported from Boima et al. [4], but within the range of 32.7-49.3% reported in previous studies in Nigeria [2,3,16]. Outside the continent patients' adherence to anti-hypertensive therapy vary between 50% and 70% [9]. Methodological differences and population characteristics may account for the wide range of fi ndings.
Nevertheless, collecting data from a referral center in patients with multi-morbidity places some hospital bias on our fi ndings.
Furthermore, the rates of tobacco and alcohol use were higher than previously reported [17]. The use of these products may worsen adherence as reported by previously [6].
Medical comorbidities were common among our patients especially the elderly. Comorbidities increase pill load leading to drug fatigue and hence non-adherence to medications.     age and adherence has been contradicting and varied with the age group studied [9]. Similar to the index study some have reported a non-signifi cant correlation in adherence with age [26][27][28] (Table 6,7).
The effect of gender and marital status on adherence may bi-directional [9]. While some studies documented positive effect on adherence others did not. Similar to the index study several other studies did not fi nd a relationship between gender and adherence [29,30]. The positive effect of marital status on adherence may be related to the ability of the individual to create social networks [9,29]. The social support inherent in marriage could be the reason why married patients were more compliant to medication than single patients [29]. However, marital status was not found to be related to patient's adherence in the index study similar to previous studies [31,32].
Item by item analysis of Morinsky-Green scores shows that forgetfulness, travelling, worry over side effects were the most common reasons for non-adherence. Forgetfulness is a widely reported factor that causes non-adherence [1,[3][4][5][6][8][9][10][11][12]. Adherence is usually better with medications that require less frequent administration especially for frequent travelers [19]. In the index study, about 94.3% of the patients sometimes do not travel with their medications. With changing lifestyles and ever busy schedules especially in younger patient; drug regimens should be made to suit individual lifestyles.
A Japanese study in elderly home-care recipients found an association between meal frequency and compliance.
Suggesting that timing drugs with regular meals may decrease non-adherence [33].
About 90.1% patients in our study stop their medication when they had concerns about potential adverse effects of antihypertensives. These concerns may include not only effects, but as fear of dependence and drug-drug interactions [9,29,34,35]. The effect of side effects on compliance may beexplained in terms of physical discomfort, skepticism about the effi cacy of the medication, and decreasing the trust in physicians.
Several studies have reported that formal education signifi cantly improves adherence [4,36]. However, the extent to which level of education remains controversial and at best equivocal with some studies reporting better adherence in patients with lower levels of education [30,35] while like the index study others reported no association [25,29]. It is expected that patients with higher educational level should have better knowledge about hypertension and its treatment and therefore be more compliant. However, DiMatteo [37] found that even highly educated patients may not understand their conditions or believe in the benefi ts of being compliant to   their medication regimen suggesting that patients with lower educational level might have more trust in physicians' advice.
Our study did not fi nd any signifi cant difference between medication adherence and levels of income. This is in keeping with previous studies [4,32,[38][39][40]. One explanation may be few high-income earners patronize public health institutions like ESUTH although other factors such high levels of poverty may be contributory. Unlike in the index study, other studies have reported signifi cant rates of non-adherence among those that use herbal remedies [9,32].
Strengths of our study include the large number of outpatients studied and the use of simple, easy to use instrument to assess medicationadherence, depression and disability. The study also sought to correlate adherence to disability and comorbidity which has not been undertaken in our region.
Limitations: This was a hospital-based study, therefore patients on multiple medications may be over represented. This may affect this study in two possible ways. Firstly, level of adherence may be lower than what may be obtainable in the community. Secondly, patient oriented health education in tertiary hospitals may have a positive impact on adherence levels thus suggesting the possibility of better adherence among these patients. A more objective measure such as urine antihypertensive drug assay have been advocated to demonstrated level of adherence.
Notwithstandingthese shortcomings, and in view of the external consistency of our data with those of similar studies, our results may well be a representative of the actual condition of patients with chronic medical disorders especially hypertension in our region and can be used to formulate local health policies, at least for the age groups studied.

Conclusions
There is a high rate of non-adherence among hypertensive patients attending tertiary care centers in the South East. Measures targeted towards improving adherence like information on the benefi ts of medication adherence and modalities of coping with disabilities should be developed for these centers. There is also the need to involve mental health practitioners in the care of such patients.