https://doi.org/10.4082/kjfm.19.

0081 • Korean J Fam Med 2021;42:150-158

Original Article

Determining the Prevalence of

eISSN: 2092-6715

and the Factors Associated with

Antihypertensive Medication
Non-Adherence in the Gaza Strip
Khalid Khadoura1,2, Elham Shakibazadeh3, Mohammad Ali Mansournia4, Yousef Aljeesh5, Akbar Fotouhi4,*

1
Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences International Campus, Tehran, Iran
2
Department of Nursing, Faculty of Health Professions, Israa University, Gaza, Palestine
3
Department of Health Education and Promotion, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
4
Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
5
Faculty of Nursing, Islamic University of Gaza, Gaza, Palestine

  See editorial commentary page on 89

Background: This study aimed to estimate the prevalence of and determine the factors associated with antihyper-
tensive medication (A-HTNM) non-adherence among hypertension care seekers attending primary health clinics
in the Gaza Strip.
Methods: A cross-sectional survey was conducted as the recruitment phase of a clustered randomized controlled
trial including 538 participants. The participants were randomly selected from 10 primary health care centers by
two-stage cluster random sampling. A structured questionnaire was used to collect data through face-to-face inter-
view. The questionnaire was developed based on the World Health Organization determinants for medication non-
adherence and the Health Belief Model. The main outcomes of this study were the prevalences of A-HTNM non-
adherence and its associated factors. Adherence status was assessed using the eight-item Morisky Self-Report
Medication Adherence Scale. Data were analyzed by STATA ver. 14.0 (Stata Corp., College Station, TX, USA)
using a standard complex survey, accounting for unresponsiveness and the clustering sampling approach.
Results: The overall prevalence of A-HTNM non-adherence was 65.8% (95% confidence interval [CI], 59.2–
71.8). Among all studied predictors, only self-efficacy of participants (odds ratio [OR], 3.8; 95% CI, 1.79–2.84)
and social support (OR, 2.26; 95% CI, 2.82–5.11) remained significantly associated with A-HTNM non-adherence
after adjust-ing for age, education level, number and frequency of A-HTNM doses per day, and comorbidities.
Conclusion: The high prevalence of non-adherence highlights the need for serious intervention to enhance the ad-
herence rate among hypertension patients. The associated factors can be considered when developing appropriate
interventions.

Keywords: Antihypertensive Drugs; Medication Non-Adherence; Prevalence

Received: June 15, 2019, Revised: August 23, 2019, Accepted: September 5, 2019
*Corresponding Author: Akbar Fotouhi https://orcid.org/0000-0002-6438-6833
Tel: +98-21-88992970, Fax: +98-21-88989664, E-mail: [email protected]

Copyright © 2021 The Korean Academy of Family Medicine

This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0)
which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence www.kjfm.or.kr   151

INTRODUCTION 2. Eligibility Criteria

Palestinian citizens attending government primary health centers in
Hypertension (HTN) was the sixth leading cause of death in Palestine Gaza, aged above 18 years, registered as an HTN patient since at least
in 2016,1) with a prevalence rate of 27.6%. 2) HTN is associated with 1 year, and taking at least one A-HTNM were eligible for this study.
community and individual burden through increased risk of stroke, Pa-tients who were diagnosed with cognitive impairment or a serious
heart, and kidney diseases. 3)
The goal for patients with HTN is to medical condition as reported by their primary care physician were
4)
achieve a blood pressure (BP) lower than 130/80 mm Hg. The avail- excluded from the study.
ability of a large number of antihypertensive medications (A-HTNM)
indicate the progress in the management of HTN in the last 5 years, 5) 3. Measures
although the rates of BP control and outcomes in HTN patients A structured questionnaire was used to collect data from the partici-
remain suboptimal as most of the efforts are directed to study the pants via a 20-minute interview during clinic hours (8 AM to 2 PM, 5
effects of medications, while little attention is paid on determining days a week). The questionnaire variables in this study were exposure
whether pa-tients take these medications as prescribed or not. 6) In and outcome variables. The exposure variables included participants’
2003, the World Health Organization (WHO) advocated the term characteristics (age, sex, marital status, employment, education level)
“adherence” to de-scribe patients’ active, voluntary, and collaborative and health status variables (smoking status, comorbidities, weight,
involvement in a mutually acceptable course of behavior to produce height, body mass index [BMI], BP measurement, number and fre-
therapeutic results. According to the WHO, the average rate of quency of A-HTNM doses, and duration of HTN).
adherence to long-term therapy in developed countries is 50%, and In addition, other exposure variables of interest (beliefs about
the rates are even lower in developing countries. 7) Adherence to medi-cation, self-efficacy, intrinsic motivation, social support, and
medical treatment is a crucial mediator of antihypertensive treatment, health system-related factors) were also assessed. The outcome
while non-adherence is a strong barrier against effective BP control. 8) variable was A-HTNM adherence.
Evidence indicates that non-adherence to chronic disease medications BP was measured on the right arm in sitting position using a mercu-ry
is common, and about 45.2% of HTN patients are non-adherent to sphygmomanometer after completing the interview; the result was
prescribed medication regimens.9) The WHO has conceptualized recorded as the person’s BP expressed in terms of mm Hg. Weight and
factors related to medica-tion adherence in five domains (patient- height were measured using a mechanical weighing machine with a height
related factors, socioeconom-ic factors, health condition-related rod (Health o Meter, McCook, IL, USA). Consequently, BMI was
factors, therapy-related factors, and healthcare system-related calculated using the WHO chart based on weight and height.
factors).10) Furthermore, the Health Belief Model (HBM) has been
widely used in medication adherence studies under the assumption 4. Instruments
that patients are able to make decisions about their health. 11) An instrument was developed based on the WHO determinants for
In this study, we aimed to estimate the prevalence of and to deter- medication non-adherence and HBM. The first part of the question-
mine the associated factors of A-HTNM non-adherence based on the naire consisted of questions about the demographic characteristics
WHO determinants and HBM among HTN care seekers in primary and clinical history of participants. The second part involved the de-
health care centers in the Gaza Strip. termination of patients’ adherence status. The third part involved the
evaluation of patients’ understanding and perception of HTN (HBM
METHODS constructs: perceived susceptibility, perceived severity, and perceived
threat). The fourth domain involved the assessment of the partici-
1. Study Design and Participants pants’ beliefs about medications (HBM constructs: perceived barriers
A cross-sectional survey was conducted between August 1 and De-cember and benefits). The fifth domain involved the determination of the par-
30, 2018. The cross-sectional survey data were derived from the baseline ticipants’ active participation in monitoring and decision making in
data of the recruitment phase of a clustered randomized controlled trial. A relation to disease management (HBM construct: self-efficacy), and
total of 538 HTN care seekers from 10 primary health care centers across the sixth domain involved the evaluation of the participants’ intrinsic
the Gaza Strip governorates were enrolled. Partici-pants were recruited by motivation (HBM construct: internal cues to action). The last two do-
two-stage cluster random sampling. In the first stage, 10 centers were mains involved the assessment of healthcare system-related factors
randomly selected by a stratified simple random sampling approach to such as relationships between patients and physicians and health sys-
select two centers from each governorate. In the second stage, we tem support (HBM constructs: perceived barriers and benefits).
proportionally selected participants from each center using systematic Adherence status was assessed using the eight-item Morisky Medi-
random sampling based on the eligibility criteria and their agreement to be cation Adherence Scale (MMAS-8), a known validated and reliable
involved in the study. When any of the selected patients refused to self-report medication adherence scale.12)
participate, the next patient was selected. The Belief about Medication Questionnaire (Arabic version) was used
to test the participants’ beliefs about the use of A-HTNMs and the

https://doi.org/10.4082/kjfm.19.0081
152  www.kjfm.or.kr Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence

degree of acceptance of the validity and reliability observed in different 5. Definition of Antihypertensive Medications Adherence
cultures with some cross-cultural adaptation of the scale.11,13,14) In this study, we depended on a self-report medication adherence
Self-efficacy was defined as a behavior-specific predictor, which was scale to determine the adherence status. The MMAS-8 by Morisky et
assessed using the modified version of the existing scale.15) Intrinsic al.12) was used for this purpose. It can identify the exact reasons for pa-
motivation was defined as the degree of one’s motivation for a particu- tients’ non-adherence and why the individual is exhibiting this behav-
lar behavior, which was scored using the Treatment Self-Regulation ior. It consisted of eight questions that are assigned points ranging
Questionnaire (TSRQ); both of these tools were previously validated from 0 (“no” answer) to 1 (“yes” answer); a lower score indicates more
and are widely used.16) adherence (0/8). The score has its own cut-off points: 0 for perfect ad-
An Arabic validated and reliable version of the Patient-Doctor Rela- herence, 1–2/8 for medium adherence, and ≥3/8 for low adherence. 12)
tionship Questionnaire-9 was used to assess the relationship between To define the status of adherence, the MMAS-8 scale responses were
patients and doctors.13,17) Likewise, a health care system support ques- dichotomized into optimal and suboptimal adherence. Responses of
tionnaire was used with some modifications. 13,18) 0/8 were classified as optimal adherence, while responses indicating
The Morisky Scale, TSRQ, and Self-Efficacy Questionnaires were non-adherence (≥3/8) and moderate adherence (1–2/8) were defined
translated to Arabic based on the five-step process (forward transla- as suboptimal adherence (A-HTNM non-adherent).
tion, expert panel, back translation, editing, and pre-testing) of trans- The cut-off point for optimal adherence in HTN patients was 80% or
19)
lating and adapting an instrument recommended by the WHO. Oth- more than that in patients showing A-HTNM. However, the decision to
er covariates such as medical comorbidities were assessed using the use self-report measures to identify patients with different levels of ad-
Charlson Comorbidity Index, which is a validated and widely used herence was based on the suggested cut-off point of the scale used.
weighted-index designed to evaluate disease comorbidity. 20)
Content validity of the Arabic questionnaires were reviewed by pan- 6. Sample Size
el of experts. Required changes were made to clarify any ambiguity The sample size was calculated to be 165 participants, with an estimat-
and to ensure the comprehension of Palestinian participants after the ed non-adherence rate of 70%.22) The sample size increased to 291
pilot study. participants considering the effect of a cluster design by 1.5 and allow-
Reliability tests were performed by test-retest within a period of 2 ing a nonresponse rate of 15%. However, the sample size was increased
weeks. Intra-class correlation coefficients (ICC) and 95% confidence again to 538 participants for the purpose of conducting a controlled
intervals (CIs) were computed for the three questionnaires (Morisky trial, as these data were the baseline data of a clustered randomized
Scale, TSRQ, and Self-Efficacy Questionnaire). Table 1 reveals the controlled trial.
strong agreement level and statistically significant ICC; an ICC agree-
ment level of 0.75–0.9 was considered good, while an ICC level of >0.90 7. Data Analysis
indicated excellent reliability.21) The lowest and highest ICC in each IBM SPSS ver. 23.0 (IBM Corp., Armonk, NY, USA) was used for ICC
questionnaire ranged between 0.63 and 0.95. However, the ICC for all test and Cronbach’s α. One-way random effects model was performed,
other items was more than 0.75. and the ICC was reported.
Internal consistency was tested by determining the Cronbach’s α for For the main analysis, a standard complex survey data analysis was
MMAS-8 adherent, TSRQ and Self-efficacy domains, which give the performed by STATA ver. 14.0 (Stata Corp., College Station, TX, USA).
values of 0.92, 0.81, and 0.91 respectively, and 0.81 for the whole ques- We accounted for clustering using the STATA PSU option and for un-
tionnaire, which is considered almost good. equal probability of selection using sample weight variable analysis.

Table 1. ICC agreement level

Questionnaire No. of test (%) No. of retest (%) ICC single measure 95% confidence interval P-value Cronbach’s α
MMAS-8 0.91 0.85–0.95 <0.001 0.92
Adherent 24 (47) 23 (45.1)
Moderately adherent 17 (33.4) 13 (25.5)
Non-adherent 10 (19.6) 15 (29.4)
TSRQ* 0.90 0.84–0.94 <0.001 0.81
Mean >3 50 (98.0) 49 (96.0)
Mean ≤3 1 (2.0) 2 (4.0)
Self-efficacy* 0.84 0.73–0.90 <0.001 0.91
Mean >3 44 (86.3) 44 (86.3)
Mean ≤3 17 (13.7) 17 (13.7)
Cronbach’s α for all 0.81
ICC, intra-class correlation; MMAS-8, eight-item Morisky Medication Adherence Scale; TSRQ, Treatment Self-Regulation Questionnaire.
*On the Likert scale.

https://doi.org/10.4082/kjfm.19.0081

Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence

 www.kjfm.or.kr   153
Data were checked for quality and statistical assumption before
conducting the main analysis. None of the tested variables violated
justed effect of explanatory variables on the medication adherence
the statistical assumptions; hence, there was no need to performed
rate. All significant variables at a level of 0.1 were included in the
non-parametric tests.
mul-tiple regression analysis. Variables that disturbed the model were
Data were described using descriptive statistics, and groups were
ex-cluded by the backward stepwise elimination method. A two-sided
compared using the chi-square test for categorical variables. Under the
P-value of less than 0.05 was considered significant.
standard complex survey data setting, univariable logistic regres-sion was
performed to assess the association between medication ad-herence state
and participants’ characteristic variables and other pre-dictors. A multiple
8. Ethical Considerations
Prior to conducting this research study, ethical approval from Tehran
logistic regression model was used to assess the ad-
University of Medical Sciences Ethical Committee (code no., IR.TUMS.
SPH.REC.1396.4828) was obtained. Approval from the Palestinian Health
Research Council (Helsinki committee) (PHRC/HC/322/18) was gained.
Table 2. Participants’ characteristics (n=538)
The purpose of the study was explained to all participants, and they were
Characteristic % or mean 95% confidence interval
reassured that their data will remain confidential; each participant was
Categorical variable (%)
asked to sign a consent form prior to participation.
Governorate
North Gaza 15.34 11.6–19.9
Gaza city 28.92 25.6–32.4 RESULTS
Middle zone 25.08 20.4–30.3
Khan Yunis 21.86 15.5–29.9
1. Participants’ Characteristics
Rafah 8.80 7.1–10.8
A total of 538 participants were included in the survey, with a 94% re-
Age groups (y)
28–39 6.98 2.9–15.4
40–59 49.71 40.5–58.9
≥60 43.31 30.7–56.7
Table 3. Participants’ health status (n=538)
Sex
Male 39.05 30.9–47.8 95% confidence
Female 60.95 52.1–69.08 Characteristic % or mean
Marital status interval
Married 90.43 83.0–94.8
Single 1.87 0.85–4.1
Categorical variable
Divorced 1.21 0.50–2.8
Duration of hypertension since diagnosis (y)
Widowed 6.49 2.8–14.1
Employment 1–5 42.24 31.2–54.1
Employed 13.53 10.7–16.9 6–20 52.57 43.1–61.9
Jobless 27.93 23.2–33.1 >20 5.19 3.3–7.8
Retired 13.18 7.2–22.9 No. of antihypertensive medications
House wife 45.35 36.5–54.0 One medication 64.42 59.6–68.9
Level of education Two medication 27.50 23.5–31.8
Illiterate 9.78 5.5–16.7 Three and more medications 8.08 54.6–11.7
Elementary school 43.10 38.4–47.9
Frequency of antihypertensive
Secondary school 27.02 24.4–29.7
medications taken per day
University 20.09 19.1–21.1
Once 64.84 55.7–72.9
BMI (kg/m2)
Normal (18.5–24.9) 7.56 6.52–9.87 Twice and more 35.16 27.0–44.2
Overweight (25–29.5) 27.23 25.58–28.95 Smoking status
Obese (≥30) 65.20 62.15–68.14 Never 81.42 76.5–85.4
BP Former 8.78 5.3–14.2
Controlled 54.41 39.6–68.5 Current 9.79 7.9–12.1
Uncontrolled 45.59 31.4–60.4 Other comorbidities
Continuous variable (mean) Low 85.57 71.8–93.2
Age (y) 57.14 53.11–61.17 High 14.43 6.7–28.1
BMI (kg/m2) 32.25 31.86–32.63 Perception of blood pressure control
Weight (kg) 87.39 86.04–88.75
Controlled 58.96 50.4–66.9
Height (cm) 164.43 163.90–164.97
Systolic BP (mm Hg) 131.66 126.78–136.53 Uncontrolled 34.02 28.0–40.6
Diastolic BP (mmHg) 83.27 81.53–85.02 Don’t know 7.01 3.1–15.1
Self-rated health
BMI, body mass index BP, blood pressure.
Very poor 3.80 0.9–13.9
Poor 14.87 8.6–24.4
Fair 39.41 34.2–44.8
Good 36.17 25.2–48.7
Excellent 57.36 34.1–92.0
Continuous variable (mean)
Duration of hypertension (y) 8.46 6.68–10.25
 https://doi.org/10.4082/kjfm.19.0081
154  www.kjfm.or.kr Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence

sponse rate. More than half (60.95%) were women with an overall smokers (81.4%). Obesity was the most common comorbidity among the
mean age of 57.1 years (95% CI, 53.1–61.2 years). The majority were study population, with a mean BMI of 32.25 kg/m 2 (95% CI, 31.86– 32.63
educated (90.2%), married (90.4%), unemployed (86.5%), and non- kg/m2). Of the total participants, 45.6% had uncontrolled BP with

Table 4. Univariate analysis of adherent status and participants’ characteristics

Unadjusted odds ratio

Characteristic Adherent % Non-adherent % P-value (two sided)
(95% confidence interval)

Age (y) 0.012* 1.03† (1.01–1.06)

18–35 2.6 3.6
36–50 16.5 27.8
51–65 47.3 50.4
>66 33.6 18.2
Sex 0.301
Female 43.6 36.7 Reference
Male 56.4 63.3 1.33 (0.71–2.5)
Marital status 0.702
Not married 10.2 9.2 Reference
Married 89.8 90.8 0.89 (0.42–1.89)
Employment 0.303
Unemployed 89.1 85.1 Reference
Employed 10.9 14.9 0.69 (0.31–1.56)
Level of education 0.028†
Illiterate 6.9 11.3 Reference
Literate 93.1 88.7 1.7 (1.08–2.67)
Body mass index (kg/m2) 0.380
Normal (18.5–24.9) 8.7 6.9 Reference
Above normal (>25.0) 91.3 93.1 0.78 (0.40–1.5)
*Statistically significant variables in multivariable regression analysis. †Odds ratio for age as a continuous variable.

Table 5. Univariate analysis of adherent status and participants’ health status

Unadjusted odds ratio

Variable Adherent % Non-adherent % P-value (two sided)
(95% confidence interval)

Duration of hypertension (y) 0.039*

1–5 32.4 47.4 Reference
>5 67.6 52.6 1.88 (1.04–3.37)
No. of antihypertensive medications 0.042*
One medication 59.0 67.2 Reference
Two medications and more 41.0 32.8 1.43 (1.02–2)
Frequency of antihypertensive medications taken per day 0.087*
Twice and more 31.6 37.0 Reference
Once 68.4 63.0 1.27 (0.95–1.71)
Smoking status 0.266
Nonsmoker 86.9 91.9 Reference
Smoker 13.1 8.1 1.7 (0.57–5.08)
Other comorbidities 0.098*
High 11.8 15.8 Reference
Low 88.2 84.2 1.40 (0.91–2.14)
Perception of BP control 0.055*
Uncontrolled 26.1 48.8 Reference
Controlled 73.9 51.2 2.7 (0.97–7.56)
Self-rated health 0.121
 Poor 46.5 64.1 Reference
Good 53.5 35.9 2.05 (0.76–5.51)
BP measurement 0.036*
Uncontrolled 34.7 51.3 Reference
Controlled 65.3 48.7 1.98 (1.06–3.68)
BP, blood pressure.
*Statistically significant variables at a level of 0.1 included in multivariable regression analysis.

https://doi.org/10.4082/kjfm.19.0081
Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence www.kjfm.or.kr 155

mean systolic and diastolic BP of 131.66 mm Hg (95% CI, 126.78– tory variables. Tables 4 and 5 show that participants’ ages (odds ratio
136.53 mm Hg) and 83.27 mm Hg (95% CI, 81.53–85.02 mm Hg), re- [OR], 1.03; 95% CI, 1.01–1.06), education level (OR, 1.7; 95% CI, 1.08–
spectively. More than half of the participants (57.25%) had been diag- 2.67), number of years since HTN diagnosis (OR, 1.88; 95% CI, 1.04–
nosed with HTN for more than 5 years (mean, 8.46 years; 95% CI, 3.37), number of A-HTNMs (OR, 1.43; 95% CI, 1.02–2.00), and BP mea-
6.68–10.25 years). Almost two-thirds of the participants (64.4%) were surement (OR, 1.98; 95% CI, 1.06–3.68) were the only factors associated
treated with only one A-HTNM once a day (64.84%), while 35.6% were significantly with adherent status.
treated with two or more medications twice or several times a day In the prediction of other independent variables, which could be as-
(35.16%). Only 14.43% of participants had high comorbidities. Approx- sociated with medication adherence status, Table 6 shows that self-ef-
imately 58.96% of the participants considered that their BP is con- ficacy of the participants (OR, 4.47; 95% CI, 3.28–6.09) and social sup-
trolled, while 41.9% rated themselves as having a good health status in port (OR, 2.87; 95% CI, 2.66–3.09) were the only predictors that were
the self-rated health questionnaire (Tables 2, 3). significantly associated with adherent status among all other studied
factors.
2. Prevalence of Adherence and Non-adherence A multiple logistic regression model was used to assess the adjusted
The participants were divided into three groups according to their association of predictors with adherence status. All statistically signifi-
MMAS-8 scores: non-adherent (≥3/8), moderately adherent (1–2/8), cant variables from univariate analysis at a level of 0.1 were included in
and adherent (0/8). Approximately 32% (95% CI, 28%–36%) and 32% the multiple logistic regression analysis. The 11 factors that were in-
(95% CI, 26%–42%) of the respondents were classified as moderately cluded in the model were age, education level, duration of HTN, fre-
adherent and non-adherent, respectively, while 34% (95% CI, 28%– quency of anti-HTN medication, number of anti-HTN medications
41%) were adherent. The MMAS-8 scale responses were dichotomized per day, other comorbidities, BP measurement, perception of BP con-
into optimal adherence (0/8) and suboptimal adherence (A-HTNM trol, self-efficacy of participants, social support, and intrinsic motiva-
non-adherent) (≥1/8). The overall prevalence of A-HTNM non-adher- tion. Four of them (duration of HTN, perception of BP control, BP
ence was 65.8% (95% CI, 59.2%–71.8%). measurement, and intrinsic motivation) disturbed the model; hence,
they were excluded by the backward stepwise elimination method.
3. Relationship between Adherence Status and Predictors The other seven factors remained statistically significant and were
Under the standard complex survey data setting, univariable logistic found to be associated with the risk of A-HTNM non-adherence (Table
regression was used to predict the association between adherence sta- 7).
tus and baseline data of the participants and other interesting explana- These significant predictors included the following: age (OR, 1.04;

Table 6. Univariate analysis of adherent status and explanatory variables

Unadjusted odds ratio

Predictor Adherent % Non-adherent % P-value (two sided)
(95% confidence interval)

Mean score of participants’understanding and perception 0.132

about hypertension
Good 98.2 94.9 3.03 (0.62–14.78)
Poor 1.8 5.1 Reference
Mean score of beliefs about medication 0.455
Positive 74.4 77.6 0.84 (0.48–1.46)
Negative 25.6 22.4 Reference
Mean score of self-efficacy of participants <0.001*
Good 84.8 55.5 4.47 (3.28–6.09)
Poor 15.2 44.5 Reference
Mean score of intrinsic motivationparticipants 0.070*
High 58.1 43.1 1.82 (0.93–3.59)
Low 41.9 56.9 Reference
Mean score of relationships between patients and physician 0.244
Good 2.1 3.8 1.85 (0.56–6.09)
Poor 97.9 96.2 Reference
Mean score of health system support 0.385
Good 46.3 42.1 1.18 (0.75–1.86)
Poor 53.7 57.9 Reference
Mean score of social support <0.001*
Good 87.3 70.5 2.87 (2.66–3.09)
Poor 12.7 29.5 Reference
*Statistically significant variables at a level of 0.1 included in multivariable regression analysis.

https://doi.org/10.4082/kjfm.19.0081

156  www.kjfm.or.kr port, poor physician-patient relationship, and health care support af-fect
adherence in different populations.5,7,23-25)
Table 7. Predictors that remained significant in multiple regression analysis Addressing these factors is very important for successful control and
management of HTN. To the best of our knowledge, this is the first
Adjusted odds ratio
Variable P-value
(95% confidence interval)
https://doi.org/10.4082/kjfm.19.0081

Age 1.04 (1.03–1.06) 0.002

Education level 0.030
Illiterate Reference
Literate 2.23 (1.12–4.45)
Frequency of antihypertensive medication 0.010
≥Twice Reference
Once 2.12 (1.31–3.42)
No. of antihypertensive medications <0.001
One medication Reference
≥2 medications 2.27 (1.91–2.71)
Other comorbidities 0.016
High Reference
Low 1.60 (1.14–2.24)
Self-efficacy of participants <0.001
Poor Reference
Good 3.80 (2.82–5.11)
Social support <0.001
Poor Reference
Good 2.26 (1.79–2.84)
By the Hosmer and Lemeshow strategy.

95% CI, 1.03–1.06), with the odds of non-adherent status increased by

0.04 times for a 1-year decrease in age; education level (OR, 2.23; 95%
CI, 1.12–4.45), with the odds of adherent status increased by 2.23 times
among literate participants compared with illiterate participants; and the
number of anti-HTN medications per day (OR, 2.27; 95% CI, 1.91– 2.71),
with the odds of adherent status increased by 2.27 times among patients
using two or more medications compared with those taking one
medication. The odds of adherent status also increased by 2.12 times
among patients using medications once a day (OR, 2.12; 95% CI, 1.31–
3.42) compared with those taking the medication twice or more per day.
On comparing between patients with low and high comorbid-ity, other
comorbidities (OR, 1.60; 95% CI, 1.14–2.24) were likely to in-crease the
odds of adherent status by 1.6 times. Other significant ex-planatory factors
were self-efficacy of participants (OR, 3.8; 95% CI, 1.79–2.84) and social
support (OR, 2.26; 95% CI, 2.82–5.11). The adher-ent rates were
increased by 3.8 and 2.26 among patients with good self-efficacy and
good social support compared with those who had poor self-efficacy and
social support, respectively.

DISCUSSION
HTN complications can be prevented by patients’ adherence to A-HT-
NM. Factors such as age, sex, number of pills per day, side effects of
medication, comorbid medical conditions, patient’s inadequate un-
derstanding about the complications of A-HTNM, lack of social sup-
Khalid Khadoura, et al.  • Antihypertensive Medication Non- intentional and unintentional reasons for medication non-adherence.
Adherence By contrast, indirect measures that in-volve self-report measures have
the potential to identify the exact rea-sons for patients’ non-adherence
study to investigate adherence to A-HTNM and its associated and why the individual is exhibiting this behavior.
factors among HTN patients attending primary health care centers Many other indirect measures are available: pill count, pharmacy
in the Gaza Strip. We investigated the levels of medication refill records, BP measurements, and Medication Events Monitoring
adherence and its predictors among 538 HTN patients attending System (MEMS). The pill count method has been criticized by re-
primary health care centers in the Gaza Strip governorates. searchers because many patients did not return with their pill contain-
Self-reported measures were used in this research to determine ers on each clinic visit. In addition, some patients might combine
the patients’ perceived response in relation to their medication some pills from different containers into one medication container.
adherence behavior, which met the purpose of the study and can Moreover, some patients do not bring the real bottles, which have
serve as a basis for conducting an interventional controlled trial on some remaining pills that indicate their poor adherence to medica-
behavioral change among non-adherent participants. tions or discard the remaining pills before their clinic visit to show
their perfect adherence to the prescribed medications. Although the
The MMAS-8 proved to be reliable, with good concurrent and pre-
effectiveness of MEMS has been shown, it is an expensive measure
dictive validity in primarily low‐income HTN patients and might func-
and has a practical problem: the removal of the dosage unit is equiva-
tion as a screening tool in outpatient settings based on its own cut-off
lent to taking the medication. Thus, the patient could open the con-
points. The scale was shown to correlate strongly with sustained be-
havioral change for individuals who attended counseling sessions. 12,26) tainer without taking the medications just to increase the adherence
Actually, there is no gold standard measure for medication adher- rate.
ence among the widely available measures of medication adherence Meanwhile, because a perfect measure does not exist, a multi-mea-
used in this research. The self-report scales have been classified as sure approach can be a good solution. Still, the decision regarding
an indirect and subjective measure of medication adherence. It is a choosing a suitable approach should balance reliability and practicali-
low cost and simply applicable measure. Determination of ty, especially cost-effectiveness and the purpose of the study. 26)
medication ad-herence can be objective or subjective, also called as The overall prevalence of A-HTNM non-adherence was 65.8% (95%
direct or indirect. The direct measure of medication adherence CI, 59.2%–71.9%). Approximately 32% and 34% of the respondents were
depends on the biologi-cal assay of the drug in body fluids (blood moderately adherent and non-adherent, respectively. These val-ues are
or urine); however, it does not provide information about almost lower than those of a previous study conducted in the
Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence high comorbidities were more likely to adhere to multiple medica-

tions.13,28)
West Bank of Palestine in 2013, which revealed that 28.9% and 54% had
Social support is a construct that describes the structure of a per-
medium and poor adherence, respectively. 27) The values reported in this son’s social environment and the perceptible instrumental and emo-
study fall within the range of the recorded results in the United Arab tional support the social environment provides. Self-efficacy is the
Emirates (54.4%), Saudi Arabia (72.1%),13) and Pakistan (77%).24) ability of patients to take their medications in good and bad
In our study, age and education level were found to be significant situations and their ability to integrate medication management into
independent factors associated with adherence, with better adherence their daily life.29) This study illustrates the significant effect of self-
observed in older and more educated people; this finding is in line efficacy and so-cial support on adherence to antihypertensive
with those of other studies. 13,24,27) Older patients could be more sensi- treatment. However, only a few studies in the Arabic region have
tive to the side effect and show deterioration when their medications addressed these factors, and the finding is consistent with those of
are ignored; these reasons probably motivated them to increase their several other studies con-ducted in different countries. 30)
adherence to their treatment. Therefore, counseling for patients in the Although this study is the first study in the Gaza Strip that used
younger age groups should consider full and detailed explanation the survey method to assess medication adherence, several
about the complications of HTN and the risks of poor adherence to limitations were observed. First, this study did not consider all
medications. Patients with greater levels of education may have a bet- factors associated with A-HTNM adherence. Therefore, a
ter understanding regarding the goal of controlling their BP and the qualitative approach and ex-ploration of more factors may provide
potential complications associated with A-HTNM non-adherence. additional information for A-HTNM adherence. Second, we did not
This study found a positive relationship between adherence and the determine the number of other medications that the patients were
number of pills prescribed. Patients taking only one tablet are less ad- taking beside A-HTNM. Third, we did not discuss the classes of A-
herent compared with those taking multiple pills per day. This is in- HTNM and were only concerned about the number of A-HTNM.
consistent with what has been advocated about one tablet dosing to In conclusion, adherence to antihypertensive treatment is low
increase adherence. However, this finding is consistent with those of
several other studies.13,23,24) One reason for this finding could be that
patients taking two or more pills recognize the severity of their
disease and hence become more attentive to their treatment. In
addition, tak-ing multiple pills probably helped them avoid forgetting
their medica-tions.
We also found that high comorbidities were associated with a de-
crease in adherence among these patients. However, previous studies
in Saudi Arabia and the United States have reported that patients with
 www.kjfm.or.kr   157

among HTN patients attending primary health care clinics in the Gaza
Strip, Palestine. Therefore, screening for non-adherence to A-HTNM
should be a part of the routine care in primary health care clinics.
Older age, higher education level, multiple pills per day, low comor-
bidities, good self-efficacy, and social support were considered as pre-dictors
of higher adherence. Thus, more attention should be given to younger and
less educated patients. As the same as, attention should be paid to
participants who take only one pill a day, had high comor-bidities and
complain of poor self-efficacy and social support. Due to the multiple factors
associated with non-adherence, interventions for improving adherence rate
should be mainly directly toward the indi-viduals themselves to increase
their ability to challenge and cross this gap regardless of the life situations
they are facing.

CONFLICT OF INTEREST
No potential conflict of interest relevant to this article was reported.

ACKNOWLEDGMENTS
The researchers would like to thank the Palestinian Ministry of Health
and the General Administration of Primary Health Care. This study was
approved by the Tehran University of Medical Sciences Ethical
Committee and the Palestinian Health Research Council. The re-
searchers are grateful to them.

ORCID
Khalid Khadoura: https://orcid.org/0000-0002-9538-2296
Elham Shakibazadeh: https://orcid.org/0000-0002-1320-2133
Mohammad Ali Mansournia: https://orcid.org/0000-0003-3343-2718
Yousef Aljeesh: https://orcid.org/0000-0003-0657-426X
Akbar Fotouhi: https://orcid.org/0000-0002-6438-6833

REFERENCES
1. Mosleh M, Aljeesh YI, Dalal K. Burden of chronic disease in the Pales-
tinian healthcare sector using Disability-Adjusted Life Years (DALY),
Palestine. Divers Equal Health Care 2016;13:261-8.
2. Khdour MR, Hallak HO, Shaeen M, Jarab AS, Al-Shahed QN. Preva-
lence, awareness, treatment and control of hypertension in the Pales-tinian
population. J Hum Hypertens 2013;27:623-8.
3. James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C,
Handler J, et al. 2014 Evidence-based guideline for the management of high
blood pressure in adults: report from the panel members ap-pointed to the
Eighth Joint National Committee (JNC 8). JAMA 2014;311:507-20.
4. Whelton PK, Carey RM, Aronow WS, Casey DE Jr, Collins KJ, Dennison
Himmelfarb C, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/
ASH/ASPC/NMA/PCNA guideline for the prevention, detection, eval-uation, and
management of high blood pressure in adults: a report of the American College of
Cardiology/American Heart Association Task

https://doi.org/10.4082/kjfm.19.0081
158  www.kjfm.or.kr Khalid Khadoura, et al.  • Antihypertensive Medication Non-Adherence

Force on Clinical Practice Guidelines. J Am Coll Cardiol primary care: development and psychometric evaluation. Gen Hosp
2018;71:e127-248. Psychiatry 2004;26:115-20.
5. Gosmanova EO, Kovesdy CP. Adherence to antihypertensive medica- 18. Ogedegbe C, Nyirenda T, Delmoro G, Yamin E, Feldman J. Health
tions: is prescribing the right pill enough? Nephrol Dial Transplant care workers and disaster preparedness: barriers to and facilitators of will-
2015;30:1649-56. ingness to respond. Int J Emerg Med 2012;5:29.
6. Blackwell B. Drug therapy: patient compliance. N Engl J Med 1973; 19. World Health Organization. Process of translation and adaptation of
289:249-52. instruments. Geneva: World Health Organization; 2018.
7. Sabate E. Adherence to long-term therapies: evidence for action. Ge- 20. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of
neva: World Health Organization; 2003. classifying prognostic comorbidity in longitudinal studies: develop-ment
8. Elliott WJ. Improving outcomes in hypertensive patients: focus on ad- and validation. J Chronic Dis 1987;40:373-83.
herence and persistence with antihypertensive therapy. J Clin Hyper-tens 21. Koo TK, Li MY. A guideline of selecting and reporting intraclass
(Greenwich) 2009;11:376-82. corre-lation coefficients for reliability research. J Chiropr Med
9. Abegaz TM, Shehab A, Gebreyohannes EA, Bhagavathula AS, Elnour 2016;15:155-63.
AA. Nonadherence to antihypertensive drugs: a systematic review and 22. Abed Y, Abu-Haddaf S. Risk factors of hypertension at UNRWA
meta-analysis. Medicine (Baltimore) 2017;96:e5641. prima-ry health care centers in Gaza governorates. ISRN Epidemiol
10. World Health organization. World Health Day 2013. Geneva: World 2013;2013:720760.
Health Organization; 2013. 23. Choi HY, Oh IJ, Lee JA, Lim J, Kim YS, Jeon TH, et al. Factors
11. Kamran A, Sadeghieh Ahari S, Biria M, Malepour A, Heydari H. affecting adherence to antihypertensive medication. Korean J Fam Med
Deter-minants of patient’s adherence to hypertension medications: 2018; 39:325-32.
applica-tion of health belief model among rural patients. Ann Med Health 24. Hashmi SK, Afridi MB, Abbas K, Sajwani RA, Saleheen D, Frossard
Sci Res 2014;4:922-7. PM, et al. Factors associated with adherence to anti-hypertensive
12. Morisky DE, Ang A, Krousel-Wood M, Ward HJ. Predictive validity treatment in Pakistan. PLoS One 2007;2:e280.
of a medication adherence measure in an outpatient setting. J Clin Hyper- 25. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med
tens (Greenwich) 2008;10:348-54. 2005;353:487-97.
13. Alsolami F, Correa-Velez I, Hou XY. Factors affecting 26. Lam WY, Fresco P. Medication adherence measures: an overview.
antihypertensive medications adherence among hypertensive patients in Biomed Res Int 2015;2015:217047.
Saudi Arabia. Am J Med Med Sci 2015;5:181-9. 27. Al-Ramahi R. Adherence to medications and associated factors: a
14. Sjolander M, Eriksson M, Glader EL. The association between pa- cross-sectional study among Palestinian hypertensive patients. J Epi-
tients’ beliefs about medicines and adherence to drug treatment after demiol Glob Health 2015;5:125-32.
stroke: a cross-sectional questionnaire survey. BMJ Open 2013;3: 28. An J, Nichol MB. Multiple medication adherence and its effect on
e003551. clin-ical outcomes among patients with comorbid type 2 diabetes and hy-
15. Warren-Findlow J, Seymour RB, Brunner Huber LR. The association pertension. Med Care 2013;51:879-87.
between self-efficacy and hypertension self-care activities among Afri-can 29. Criswell TJ, Weber CA, Xu Y, Carter BL. Effect of self-efficacy and
American adults. J Community Health 2012;37:15-24. social support on adherence to antihypertensive drugs. Pharmacotherapy
16. Pelletier LG, Tuson KM, Haddad NK. Client motivation for therapy 2010;30:432-41.
scale: a measure of intrinsic motivation, extrinsic motivation, and 30. Magrin ME, D’Addario M, Greco A, Miglioretti M, Sarini M, Scrignaro
amotivation for therapy. J Pers Assess 1997;68:414-35. M, et al. Social support and adherence to treatment in hypertensive patients: a
17. Van der Feltz-Cornelis CM, van Oppen P, van Marwijk HW, de Beurs E, meta-analysis. Ann Behav Med 2015;49:307-18.
van Dyck R. A patient-doctor relationship questionnaire (PDRQ-9) in

https://doi.org/10.4082/kjfm.19.0081