Sweileh et al.

BMC Public Health 2014, 14:94

http://www.biomedcentral.com/1471-2458/14/94

RESEARCH ARTICLE Open Access

Influence of patients’ disease knowledge and

beliefs about medicines on medication
adherence: findings from a cross-sectional survey
among patients with type 2 diabetes mellitus in
Palestine
Waleed M Sweileh1*, Sa’ed H Zyoud2, Rawan J Abu Nab’a3, Mohammed I Deleq3, Mohammed I Enaia3,
Sana’a M Nassar3 and Samah W Al-Jabi2

Abstract
Background: Diabetes mellitus (DM) is a common serious health problem. Medication adherence is a key
determinant of therapeutic success in patients with diabetes mellitus. The purpose of this study was to assess
medication adherence and its potential association with beliefs and diabetes – related knowledge in patients with
type II DM.
Methods: This study was carried out at Al-Makhfia governmental diabetes primary healthcare clinic in Nablus,
Palestine. Main outcome of interest in the study was medication adherence. The Beliefs about Medicines
Questionnaire (BMQ) was used to assess beliefs. Morisky Medication Adherence Scale (MMSA-8©) was used to assess
medication adherence. The Michigan diabetes knowledge test (MDKT) was used to assess diabetes – related
knowledge. Univariate and multivariate analysis were carried out using Statistical Package for Social Sciences (SPSS 20).
Results: Four hundred and five patients were interviewed. The mean ± SD age of the participants was 58.3 ± 10.4
(range = 28 – 90) years. More than half (53.3%) of the participants were females. Approximately 42.7% of the study
sample were considered non-adherent (MMAS-8© score of < 6). Multivariate analysis showed that the following
variables were significantly associated with non-adherence: disease-related knowledge, beliefs about necessity of
anti-diabetic medications, concerns about adverse consequences of anti-diabetic medications and beliefs that
medicines in general are essentially harmful. Diabetic patients with high knowledge score and those with strong
beliefs in the necessity of their anti-diabetic medications were less likely to be non-adherent ([O.R = 0.87, 95% CI of
0.78 – 0.97] and [O.R = 0.93, 95% of 0.88 – 0.99] respectively). However, diabetic patients with high concerns about
adverse consequences of anti-diabetic medications and those with high belief that all medicines are harmful were
more likely to be non-adherent ([O.R = 1.09; 95% C.I of 1.04 – 1.16] and [O.R = 1.09, 95% C.I of 1.02 – 1.16] respectively).
Conclusions: Beliefs and knowledge are important factors in understanding variations in medication adherence
among diabetic patients. The BMQ can be used as a tool to identify people at higher risk of non-adherence.
Improving knowledge of patients about their illness might positively influence their medication adherence.
Keywords: Chronic illness, Beliefs about medicines, Adherence, Palestine

* Correspondence: waleedsweileh@yahoo.com
1
Department of Pharmacology/ Toxicology, College of medicine and health
sciences, An-Najah National University, Nablus, Palestine
Full list of author information is available at the end of the article

© 2014 Sweileh et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Sweileh et al. BMC Public Health 2014, 14:94 Page 2 of 8
http://www.biomedcentral.com/1471-2458/14/94

Background adherence and other factors [14-17]. The research ap-

Diabetes mellitus (DM) is a common health problem proach was based on using medication adherence as an
that has serious medical and economic consequences. outcome (adherence versus non-adherence) while using
The world prevalence of diabetes among adults (aged demographic, clinical, psychological factors as independ-
20–79 years) was estimated to be 6.4% in 2010 and will ent variables. The tools used in this study have been previ-
increase to 7.7% by 2030 [1]. Between 2010 and 2030, ously used by other investigators [14-21].
there will be a 69% increase in numbers of adults with
diabetes in developing countries and a 20% increase Study setting
in developed countries. It is striking that Arab world Nablus is the largest city in north West-Bank of Palestine.
(North Africa, Middle East, and Gulf area) will have the Residents of Nablus city are predominantly Arabs. The
second highest increase in percentage of people with study was carried out at Al-Makhfia governmental dia-
DM in 2030 compared to other parts of the world [1]. betes primary healthcare clinic in Nablus city. Al-Makhfia
No reliable data about treatment outcomes, complica- center is the main governmental center that provides
tions, and economic effects of diabetes mellitus are avail- care for diabetic patients with governmental insurance
able from Middle East in general and from Palestine in in Nablus city. During the study period, the investigators
particular [2]. made daily visits to the diabetic clinic to recruit and inter-
Medication adherence has important therapeutic and view potential participants.
economic consequences [3,4]. Medication adherence is
believed to be influenced by factors beyond the traditional Sample population
demographic and clinical factors [5,6]. For example, the This study included a convenience sample of adult
extended Self-Regulatory Model, which includes both ill- population. Participants were recruited from Al-Makhfia
ness and treatment beliefs, was successful in explaining diabetic clinics while waiting to be seen by their health
variations in medication adherence among patients with care providers. The inclusion criteria for this study were:
certain chronic diseases [7]. Diabetes-related knowledge 1) patients who reported having type 2 diabetes; 2) avail-
have also been reported to influence both medication ad- ability of a medical file at the diabetic clinic; 3) a history of
herence and glycemic control [8]. This suggests that there at least one year of diabetes mellitus; 4) currently being
is a complex model of demographic, clinical, knowledge under medical care for diabetes; and finally 5) willingness
and behavioral factors that affect medication adherence. to participate in this study. The main exclusion criterion
Several studies were carried out in Palestine and Arab was physical and/or mental conditions that could interfere
world about medication adherence in general and among with the participant’s ability to understand and/or answer
diabetic patients in particular [9,10]. Unfortunately, none questions in any of the used scales.
of these studies investigated the influence of factors such
as disease-related knowledge or behavioral aspects on Sample size
medication adherence. Such factors are important given A previous study indicated that medication non-adherence
the cultural differences between Arabs and Europeans or among Palestinian diabetic patients is in the range of 20 –
Asians where most studies about influence of behavioral 50% [9]. Therefore, the sample size was estimated based
and knowledge aspects on medication adherence were on the following assumptions: a descriptive study with di-
carried out [6,11,12]. Beliefs about medicines and extent chotomous outcome (adherence versus non-adherence),
of disease-related knowledge are different among differ- rate of medication non-adherence to be as low as 20%,
ent cultures [13]. Therefore, the objective of this study confidence interval width of 10% and confidence limit to
was to assess anti-diabetic medication adherence and its be 95%. Therefore, an estimated sample of 385 diabetic
potential association with beliefs and diabetes–related patients is needed for this study [22]. In order to minimize
knowledge among patients with type II DM attending a erroneous results and increase the study reliability, the in-
primary healthcare clinic in Palestine. vestigators recruited a total of 405 diabetic patients during
the study period between July 2012 and October 2012.
Methods
Study design Ethical approval
This was a cross-sectional study for the purpose of This study was approved by the Palestinian Ministry of
evaluating the association between beliefs about medi- Health and Institutional Review Board (IRB) at An-Najah
cines, diabetes-related knowledge, demographic and clin- National University. The interviewer explained the purpose
ical factors with medication adherence among Palestinians of the study to each participant and a verbal consent was
with type 2 diabetes mellitus. The approach we followed obtained from the participants prior to the commencement
in this study was similar to that used by other scholars of the study. The participants were also informed that their
who investigated the relationship between medication participation was voluntary and that they could withdraw
Sweileh et al. BMC Public Health 2014, 14:94 Page 3 of 8
http://www.biomedcentral.com/1471-2458/14/94

from the interview at any time without consequences. The Beliefs about medicines
participants were assured that their responses would be Beliefs about Medicines was assessed using Beliefs about
treated in confidence and they were assured anonymity Medicines Questionnaire (BMQ); [11]. Approval to use
through the use of strict coding measures. All information and translate BMQ was obtained from the developer.
was kept confidential. The study was carried out in full Translation of BMQ was also carried out according to
compliance with the guidelines of good clinical practice of the standard forward and backward method. BMQ con-
the world assembly declaration of Helsinki and was ap- sists of two sections, general and specific. The Specific sec-
proved by the university ethical committee. tion assesses patients’ beliefs about medications prescribed
for a particular illness and consists of two scales that
assess personal beliefs about the necessity and concerns
Recruitment procedure
of prescribed medication for controlling their illness (5
A brief screening was conducted by investigators to
statements for each scale). The General section deals
identify potential participants in the following manner:
with more general beliefs about medicines and consists
every person in the waiting area of the diabetic clinic
of two scales, the General-overuse scale which addresses
was asked if he/she is willing to talk to the investigator.
views about the way in which medicines are used by physi-
If the person agreed to talk to the investigator for pos-
cians (4 statements), and the General-harm scale which
sible participation, then an informed consent was read
assesses beliefs about the degree to which patients per-
and obtained by the investigators. Once the verbal con-
ceive medicines as essentially harmful. Each statement has
sent was obtained, verification of inclusion and exclusion
five potential answers (strongly disagree, disagree, uncer-
criteria took place. The questionnaires required for the
tain, agree, and strongly agree). The answers are scored
study were presented and explained during this interview.
from 1 (strongly disagree) to 5 (strongly agree). Points of
All participants completed the questionnaires in a private
each scale are summed to give a scale score. Higher scores
area in the clinic.
indicate stronger beliefs in the concepts of the scale.
The forms and questionnaires used in this study were:
There are five statements in Specific-necessity and
1) demographic and clinical information about the partici-
Specific-concerns scales and therefore the total sum of
pant, 2) Morisky Medication Adherence Scale (MMAS-8©)
possible scores in these scales would range from 5 to 25.
to determine level of medication adherence, 3) Beliefs about
Higher specific-necessity scores represent stronger per-
Medicines (BMQ) to assess beliefs of participants about
ceptions of personal need for the medication to maintain
their medications; and 4) Michigan Knowledge scale to
health now and in the future. Higher specific-concerns
gain information about diabetes-related knowledge among
scores represent stronger concerns about the potential
participants.
negative effects of the medication. There are four state-
ments in the General-overuse and the General-harm
Instruments scales and therefore the total sum of possible scores in
Medication adherence these scales would range from 4 to 20. Higher scores on
Adherence to anti-diabetic medications was measured the General-harm scale represent more negative views
using Morisky Medication Adherence Scale (MMAS-8©) about medicines as a whole and a tendency to see medi-
[23]. Approval to use and translate the (MMAS-8©) into cines as fundamentally harmful and addictive poisons.
Arabic language was obtained from the developer. The Higher scores on the General-overuse indicate more nega-
translation was carried out according to standard forward tive views about the way in which medicines are pre-
and backward method. The Arabic- translated version of scribed and beliefs that they are overused by physicians.
(MMAS-8©) was used in previous publication [9]. The
(MMAS-8©) consists of eight questions designed to meas- Diabetes knowledge test
ure medication adherence. The first 7 ones are Yes/No The brief diabetes knowledge test developed by the Michigan
questions while the last question is answered on a 5-point Diabetes Research and Training Center (MDRTC) known
Likert scale. One point is given for each sentence based on as Michigan diabetes knowledge test (MDKT) [24] was
the answer. In the first 7 questions, one point is given for translated into Arabic and used to assess the general
each “NO” answer except for question number 5 where knowledge of the participants. Approval to use and trans-
one point is given for the “YES” answer. For item number late MDKT was obtained from the developer. The general
8, one point is given for “never/rarely” item and zero point MDKT consists of 14 multiple choice questions with one
is given for “all the time” item. The total MMAS-8© score correct choice for each question. The Knowledge score is
is the summation of the scores for the 8 questions. The determined by giving one point for each correct answer
total score obtained ranges from 0–8. In this study, pa- and zero for the wrong answer or no response. Those who
tients with a total score of MMAS-8© < 6 were considered get the highest scores are the most knowledgeable about
non-adherent. diabetes. The total knowledge scores ranges from 0 to 14
Sweileh et al. BMC Public Health 2014, 14:94 Page 4 of 8
http://www.biomedcentral.com/1471-2458/14/94

with higher scores indicating higher level of diabetes Patients reported an average of 2.1 ± 1.6 additional
general knowledge. The diabetes knowledge test is an illnesses (median = 4; Q1 - Q3: 3 – 6). Hypertension
effective, efficient, and inexpensive way to obtain a gen- (187; 46.2%), and dyslipidemia (149; 36.8%) were the
eral assessment of a patient's knowledge about diabetes most frequently reported additional illnesses in the
and its care. Due to the fact that the test is short and its study sample. Patients reported an average of 1.8 ± 0.7
reading level is that of the 6th grade level, the diabetes (median = 2; Q1 – Q3: 1 – 2) anti-diabetic medications
knowledge test can usually be self-administered. More- and 4.3 ± 2.1 (median = 5; Q1-Q3: 4 – 7) different medica-
over, review of correct and incorrect items also can be tions taken on daily basis. One hundred and sixty eight pa-
used to provide feedback to patients and creates oppor- tients (41.5%) reported taking ≥ 5 medications on a regular
tunities for teachable moments [24]. basis. The mean duration of DM reported by the patients
was 7 ±6 (median = 8.5; Q1-Q3: 4 – 12 years). More than
Demographic and health information half of the participants had ≤ high school education (334;
This section of the questionnaire was designed to obtain 82.5%). Demographic and clinical characteristics of partici-
information about demographic data and health history pants are presented in Table 1.
of the participants. This section included information
about age, gender, years of education, income, medical
history, and current anti-diabetic medications. The med- Reported adherence and beliefs
ical history question was presented to the participant as The internal consistency of the MMAS-8© was satisfac-
a list of nine illnesses with dichotomous (yes/no) response. tory (alpha = 0.7). Two hundred and thirty two (57.3%)
All the participants were asked to report all the medica- patients were considered adherent (MMAS-8© adher-
tions that they use on chronic basis. The data reported by ence score ≥ 6) while 173 (42.7%) were non-adherent
the participants regarding their anti-diabetic medications (MMAS-8© adherence score < 6). More than one third
was validated through checking the computerized system (38%) of the participants reported that sometimes they
at the MOH which contained up-to-date information forgot to take their anti-diabetic medications; 24.0% re-
about the patients and their medications. ported that they did not take their medications on at
least one occasion in the two weeks before the inter-
Data management and statistical analysis view; 17.0% reported that they discontinued taking their
During the pre-analysis phase, the data were coded to medications without telling their doctor when they felt
maintain confidentiality for all participants. Summative worse upon taking their medications; 33.1% reported
score of the instruments was entered as a continuous that they sometimes forgot to take their medications
measure. Data were entered and analyzed using SPSS with them when they traveled or left home; 91.4.0%
(Statistical Package for the Social Sciences). Descriptive reported taking all their medications on the day before
statistics was carried out for all variables and expressed the interview; 17.0% reported that they stopped taking
as mean ± SD for continuous variables with normal distri- their medicines when they felt like their diabetic symp-
bution and as median (inter-quartile range: Q1-Q3) for toms were under control; 34.6% reported feeling hassled
non-normally distributed variables. Kolomogrov-Smirnov by their treatment plans; and finally 73.9% reported that
test was used to test normality of the variables. Factors they had difficulty remembering to take all their medicines
associated with non-adherence (MMAS-8 score < 6) were at least once in a while (Table 2). Internal consistency
analyzed with binary logistic regression followed by mul- for the BMQ four subscales was acceptable and showed
tiple logistic regression analysis. The dependent variable values between alpha = 0.63 – 0.82. Scores for Specific-
was non-adherence (coded as 1). Multiple logistic regres- necessity subscale were significantly higher than that for
sion analysis was carried out using variables that showed Specific-concern scores [median (interquartile): 20 (16 – 21)
significance in univariate analysis. A p value of < 0.05 was versus 14 (10–17); p < 0.001]. The medians (interquartile)
considered statistically significant. The relationship be- for General-overuse and General-harm subscale were 12
tween non-adherence and any particular variable of inter- (10 – 14) and 10 (8 – 13) respectively. Participants en-
est was evaluated by calculating an odds ratio (O.R) at dorsed the belief that their medications are necessary for
95% confidence interval (C.I). their current health but they were concerned about be-
coming too much dependent on their medications. Fifty
Results one percent of the participants who endorsed belief that
General characteristics of the study sample medications are harmful were non-adherent. Finally, ana-
During the study period, a total of 405 consecutive pa- lysis of MDKT scores showed that the majority of the
tients were recruited. The mean ± SD age of the partici- participants (327; 80.7%) scored ≥ 7 out of a total score
pants was 58.3 ± 10.4 (range = 28 – 90) years. More than of 14. The mean ± SD of the MDKT scores was 8.2 ± 2
half of the participants were females (216; 53.3%). and median (Q1 – Q3) of 8 (7 – 10).
Sweileh et al. BMC Public Health 2014, 14:94 Page 5 of 8
http://www.biomedcentral.com/1471-2458/14/94

Table 1 Univariate analysis of factors associated with non-adherence

Variable Total N = 405 Non-adherent N = 173 Adherent N = 232 Odds ratio with 95% CI P-value
Age 58.3 ± 10.4 59.0 ± 10.7 57.7 ± 10.2 1.0 (0.99 – 1.0) 0.21
Gender
Male 189 (46.7%) 78 (45.1%) 111 (47.8%) Reference 0.58
Female 216 (53.3%) 95 (54.9%) 121 (52.2%) 1.1 (0.8 – 1.7)
Education
Illitrate 49 (12.1%) 23 (13.3%) 26 (11.2%) Reference
Primary 168 (41.5%) 73 (42.2%) 95 (40.9%) 0.9 (0.5 – 1.6) 0.54
Secondary 117 (28.5%) 52 (30.1%) 65 (28.0%) 0.9 (0.5 – 1.8)
College 71 (17.5%) 25 (14.5%) 46 (19.8%) 0.6 (0.3 – 1.3)
Marital Status
Single 105 (25.9%) 55 (31.8%) 50 (47.6%) Reference 0.021
Married 300 (74.1%) 118 (68.2%) 182 (60.7%) 0.6 (0.4 – 0.9)
Disease knowledge score 8 (7 – 10) 8 (6 – 9) 9 (7 – 10) 0.8 (0.7 – 0.9) < 0.001
Duration of Diabetes Mellitus 8.5 ± 6.0 8.5 ± 5.8 8.5 ± 6.2 1.0 (0.96 – 1.0) 0.96
Presence of chronic diseases
Yes 284 (70.1%) 132 (46.5%) 152 (53.5%) 1.7 (1.1 – 2.6) 0.02
No 121 (29.9%) 41 (33.9%) 80 (66.1%) Reference
Anti-diabetic therapy
Monotherapy 144 (35.6%) 57 (32.9%) 87 (37.5%) Reference 0.34
Combination 261 (64.4%) 116 (67.1%) 145 (62.5%) 1.2 (0.8 – 1.8)
Insulin Use
Yes 132 (32.6%) 61 (35.3%) 71 (30.6%) Reference 0.32
No 273 (67.4%) 112 (64.7%) 161 (69.4%) 0.8 (0.5 – 1.2)
Number of medications 4.3 ± 2.1 4.5 ± 2.1 4.1 ± 2.1 1.1 (1.0 – 1.2) 0.042
Specific -necessity score 18.5 ± 4.0 17.9 ± 4.1 19.0 ± 3.8 0.94 (0.9 – 1.0) 0.009
Specific- concern score 14.0 ± 4.3 15.2 ± 3.9 13.1 ± 4.3 1.1 (1.1 – 1.2) < 0.001
General – overuse score 12.0 ± 3.3 12.0 ± 3.3 12.0 ± 3.3 1.0 (0.94 – 1.1) 0.95
General – harm score 10.5 ± 3.7 11.5 ± 3.7 10.0 ± 3.7 1.1 (1.1 – 1.2) < 0.001
Abbreviations: C.I confidence interval.

Factors affecting non-adherence variables were significantly associated with non-adherence:

Univariate analysis (Table 1) showed that there was no sig- disease-related knowledge, beliefs about necessity of
nificant difference between adherers and non-adherers in anti-diabetic medications, concerns about adverse con-
age, duration of illness, gender, education, anti-diabetic sequences of the chronic use of anti-diabetic medications
regimen and use of insulin. However, there was a signifi- and beliefs that medicines in general are essentially
cant association between non-adherence and marital sta- harmful. Diabetic patients with high knowledge score
tus, presence of other chronic illnesses, diabetes-related and those with strong beliefs in the necessity of
knowledge, total number of chronic medications, specific- anti-diabetic medications were less likely to be non-
necessity, specific-concern and general-harm scales. Mar- adherent ([O.R = 0.87, 95% CI of 0.78 – 0.97] and
ried participants and those with high diabetes-related [O.R = 0.93, 95% of 0.88 – 0.99] respectively). How-
knowledge score were less likely to be non-adherent ever, diabetic patients with high concerns about ad-
([O.R = 0.6; 95% C.I of 0.4 – 0.9] and [O.R = 0.8; 95% C.I verse consequences of anti-diabetic medications and
of 0.7 - 0.9] respectively). Adherers have significantly those with high belief that all medicines are essen-
higher specific-necessity, lower specific-concern, and lower tially harmful were more likely to be non-adherent
general-harm scores compared to non-adherers. Multi- ([O.R = 1.09; 95% C.I of 1.04 – 1.16] and [O.R = 1.09,
variate analysis (Table 3) showed that the following 95% C.I of 1.02 – 1.16] respectively).
Sweileh et al. BMC Public Health 2014, 14:94 Page 6 of 8
http://www.biomedcentral.com/1471-2458/14/94

Table 2 Self-reported medication adherence behavior of all medicines are essentially harmful. Our findings regard-
study participants as determined by the Morisky 8-Item ing association of medication adherence with knowledge
Medication Adherence Scale (MMAS-8©) and belief are in agreement with those obtained by other
Item Number (%) of patients investigators. Most studies about medication adherence
who answered yes
concluded that negative beliefs about medications is a
Do you sometimes forget to take your 154 (38.0%) powerful barrier to successful adherence [25-31]. There-
[health concern] pills?
fore, healthcare providers should address patient’s beliefs
People sometimes miss taking their 97 (24.0%) about medications in the hope of improving medication
medications for reasons other than
forgetting. Thinking over the past two weeks, adherence and therapeutic outcome. Furthermore, health-
were there any days when you did not take care workers need to assess and educate patients about
your [health concern] medicine? diabetes mellitus to improve the level of medication ad-
Have you ever cut back or stopped taking 69 (17.0%) herence and consequently therapeutic outcome.
your medication without telling your doctor,
because you felt worse when you took it?
The results of our study showed that most diabetic pa-
tients strongly believe that anti-diabetic medications are
When you travel or leave home, do you 134 (33.1%)
sometimes forget to bring along your [health necessary for their current and future health. However,
concern] medication? diabetic patients expressed their concerns about the ad-
Did you take your [health concern] medicine 370 (91.4%) verse consequences of taking anti-diabetic medications
yesterday? on regular basis. Similar findings were also reported in
When you feel like your [health concern] is 69 (17.0%) hypertensive patients [32]. Concerns of about long-term
under control, do you sometimes stop taking effects of taking anti-diabetic medications should be
your medicine?
addressed by healthcare workers to minimize concerns
Taking medication everyday is a real 140 (34.6%) and consequently improve medication adherence. Pa-
inconvenience for some people. Do you ever
feel hassled about sticking to your [health tients with diabetes mellitus need to know that their
concern] treatment plan? medications are not addictive and that medications have
How often do you have difficulty an acceptable safety profile for long-term use. Studies
remembering to take all of your medicine? have reported that concerns about adverse consequences
Never/rarely 125 (30.9%) medications is independent of patients’ age and level of
Once in a while 174 (43%) education [33]. Therefore, concerns about anti-diabetic
medications need to be addressed regardless of the age
Sometimes 86 (21.2%)
or level of education of the patient. It is important to
Usually 14 (3.5%)
note that not all studies endorsed the finding that belief
All the time 6 (1.5%) about medications is an important factor in medication
adherence. For example, a Swedish study concluded that
Discussion about one-third of the migraineurs did not adhere to
In the current study, medication adherence and its poten- their prophylactic drugs and that belief about medicines
tial association with beliefs about medicines and diabetes– and medication-related factors could not predict non-
related knowledge was assessed in a selected sample of adherence among those patients. The authors recom-
Arab Palestinians patients with type II DM. Our results mended further research on medication-related variables
showed that non-adherence was significantly associated in relation to adherence among migraineurs [34]. These
with diabetes-related knowledge, beliefs about necessity findings might suggest that the influence of beliefs on
of the anti-diabetic medications, concerns about adverse medication adherence is dependent on the type of the
consequences of anti-diabetic medications, and beliefs that chronic disease.

Table 3 Multivariate analysis of factors associated with non-adherence

Variable β S.E. Wald p-value Odds ratio with 95% C.I
Marital status (single) 0.46 0.25 3.54 0.060 1.59(0.98-2.56)
High knowledge score −0.14 0.06 6.27 0.012 0.87(0.78-0.97)
Presence of other co-morbid disease 0.42 0.31 1.83 0.177 1.53(0.83-2.81)
High number of medications consumed per day 0.08 0.07 1.33 0.249 1.08(0.95-1.24)
High Specific – necessity score −0.07 0.03 5.77 0.016 0.93(0.88-0.99)
High Specific – concern score 0.07 0.03 11.86 0.001 1.10(1.04-1.16)
High General – harm score 0.07 0.03 7.29 0.007 1.09(1.02-1.16)
Abbreviations: C.I confidence interval, β coefficient of predictor variables, S.E. Standard error.
Sweileh et al. BMC Public Health 2014, 14:94 Page 7 of 8
http://www.biomedcentral.com/1471-2458/14/94

An interesting finding in our study is the effect of Therefore, caution should be exercised in generalizing
total number of medications consumed daily on medi- our findings. Fourth, the selection method might have
cation adherence among diabetic patients. Univariate created bias toward positive beliefs since patients who
analysis showed weak significant association between attend to the clinic are those who usually care about
total number of medications and medication adherence. their health. These limitations might be responsible for
However, this weak significant association disappeared the small observed odds ratio. Finally, no glycemic con-
when strong factors like beliefs and disease-related know- trol data (HbA1C) were obtained. If such piece of infor-
ledge were entered into the regression model. Further- mation was available, then we would be able to link
more, anti-diabetic regimen (mono versus combination adherence, beliefs and knowledge with glycemic control.
therapy) showed no significant association with medica-
tion adherence in univariate analysis. This emphasizes the Conclusions
idea that it is not the total number of medications that As a conclusion, beliefs in one’s medications and diabetes-
determines level of adherence, rather, it is the belief related knowledge were significantly associated with ad-
about the importance and necessity of medicines for pa- herence. Assessment of beliefs and knowledge can be used
tient’s life which determines level of adherence. Similar to understand variations in adherence among diabetic pa-
result was obtained with regard to marital status. Univari- tients. The Beliefs about Medicines Questionnaire may be
ate analysis showed that married patients were less likely an appropriate tool to assess beliefs about medicines. Fi-
to be non-adherent compared to single patients. However nally, improving knowledge of diabetic patients about their
this association disappeared in multiple regression model. illness can positively influence their medication adherence
This means that family factors are important for medica- and therapeutic outcome.
tion adherence when considered as single factor but such
family factors become insignificant when considered in Abbreviations
DM: Diabetes mellitus; MMAS-8©: Morisky Medication Adherence Scale;
the presence of other stronger factors like beliefs and BMQ: Beliefs about medicines questionnaire; MDKT: Michigan diabetes
disease-related knowledge. knowledge test; IRB: Institutional review board; SPSS: Statistical Package for
In our study we used self-reporting method to measure Social Sciences; Q1-Q3: Lower – Upper quartiles; SD: Standard deviation;
O.R: Odds ratio; C.I: Confidence interval.
medication adherence because it is considered the sim-
plest and the least expensive method. George et al. found Competing interests
that when a valid scale [35], like Morisky questionnaire, is The authors declare that they have no competing interests.
used to assess medication adherence, the obtained scores
Authors’ contributions
will be accurate with both sensitivity and specificity of
All authors were involved in drafting the article and all authors approved the
over 70%. The 8-item Morisky Medication Adherence final version to be submitted for publication. All authors have added an
Questionnaire (MMAS-8©) has been translated into differ- intellectual significant value to the manuscript. RA, MD, ME and SN were
involved in subject recruitment and interview, data collection, data coding
ent languages and has been validated in patients with
and entry, literature review, and manuscript editing. This was done as a
different types of chronic illnesses [23,36,37]. In our study, Pharm D graduation project at An-Najah National University. WS, SA and SZ
we used “Beliefs about Medicines Questionnaire (BMQ)” were involved in concept, study design, data analysis, data interpretation,
manuscript writing and editing, manuscript submission, manuscript revision.
to measure patients’ beliefs about medicines. The BMQ
The project was initially and originally conceptualized and designed by the
has been translated into several languages to assess beliefs Clinical Pharmacology/ Toxicology Research Group at An-Najah National
about medicines across a wide range of diseases like dia- University (WS, SZ, and SA). The project was then assigned to Pharm.
D senior students as a graduation project and was academically supervised
betes mellitus, mental health illnesses, rheumatoid arth-
by W.S and S.Z in adherence to An-Najah University regulations with regard
ritis and others [30, 33]. Therefore, tools used in our study to academic supervision for graduation projects.
are considered appropriate tools to achieve the stated
objective. Authors’ information
Professor Waleed M. Sweileh is the head of a research group (S.Z, S.A, A.S
Our study had few limitations. First, a self-report method and A.A) which has published in the field of clinical pharmacology,
was used to assess medication adherence. Therefore, toxicology, pharmacoepidemiology, social and community pharmacy, clinical
overestimation of adherence may have occurred. More pharmacy and medicine. The research group has also supervised many
students in the fields of nursing, public health and pharmacy.
precise estimates of medication adherence can be ob-
tained through direct methods. However, self-reported Acknowledgment
assessment of medication adherence is practical and The authors would like to express many thanks and gratitude to An-Najah
inexpensive. Second, the validity of the Arabic version University and the Palestinian MOH for their help and ethical approval to
conduct this study.
of adherence and belief scales has not been tested and
further studies are needed to confirm validity of both Author details
1
scales among patients in the Arab world. Third, al- Department of Pharmacology/ Toxicology, College of medicine and health
sciences, An-Najah National University, Nablus, Palestine. 2Department of
though the sample size was relatively large, it was not clinical pharmacy and Pharmacotherapy, College of medicine and health
representative of Palestinian or Arab diabetic patients. sciences, An-Najah National University, Nablus, Palestine. 3Pharm. D Program,
Sweileh et al. BMC Public Health 2014, 14:94 Page 8 of 8
http://www.biomedcentral.com/1471-2458/14/94

College of medicine and health sciences, An-Najah National University, 20. Sweileh WM, Ihbesheh MS, Jarar IS, Sawalha AF, Abu Taha AS, Zyoud SH,
Nablus, Palestine. Morisky DE: Differences in medication adherence, satisfaction and clinical
symptoms in schizophrenic outpatients taking different antipsychotic
Received: 31 July 2013 Accepted: 29 January 2014 regimens. Curr Drug Saf 2011, 6(5):285–290.
Published: 30 January 2014 21. Sweileh WM, Ihbesheh MS, Jarar IS, Sawalha AF, Abu Taha AS, Zyoud SH,
Morisky DE: Antipsychotic medication adherence and satisfaction among
Palestinian people with schizophrenia. Curr Clin Pharmacol 2012,
References 7(1):49–55.
1. Shaw JE, Sicree RA, Zimmet PZ: Global estimates of the prevalence of 22. Daniel WW: Biostatistics: Basic Concepts and Methodology for the Health
diabetes for 2010 and 2030. Diabetes Res Clin Pract 2009, 87(1):4–14. Sciences. New Jersey, USA: John Wiley & Sons; 2010.
2. Husseini A, Abu-Rmeileh NM, Mikki N, Ramahi TM, Ghosh HA, Barghuthi N, 23. Morisky DE, Green LW, Levine DM: Concurrent and predictive validity of a
Khalili M, Bjertness E, Holmboe-Ottesen G, Jervell J: Cardiovascular diseases, self-reported measure of medication adherence. Med Care 1986,
diabetes mellitus, and cancer in the occupied Palestinian territory. Lancet 24(1):67–74.
2009, 373(9668):1041–1049. 24. Fitzgerald JT, Funnell MM, Hess GE, Barr PA, Anderson RM, Hiss RG, Davis
3. Sokol MC, McGuigan KA, Verbrugge RR, Epstein RS: Impact of medication WK: The reliability and validity of a brief diabetes knowledge test.
adherence on hospitalization risk and healthcare cost. Med Care 2005, Diabetes Care 1998, 21(5):706–710.
43(6):521–530. 25. Gatti ME, Jacobson KL, Gazmararian JA, Schmotzer B, Kripalani S:
4. Ho PM, Rumsfeld JS, Masoudi FA, McClure DL, Plomondon ME, Steiner JF, Relationships between beliefs about medications and adherence.
Magid DJ: Effect of medication nonadherence on hospitalization and Am J Health Syst Pharm 2009, 66(7):657–664.
mortality among patients with diabetes mellitus. Arch Intern Med 2006, 26. Chummun H, Boland D: How patient beliefs affect adherence to
166(17):1836–1841. prescribed medication regimens. Br J Nurs 2013, 22(5):270–276.
5. Vermeire E, Hearnshaw H, Van Royen P, Denekens J: Patient adherence to 27. Clifford S, Barber N, Horne R: Understanding different beliefs held by
treatment: three decades of research. A comprehensive review. adherers, unintentional nonadherers, and intentional nonadherers:
J Clin Pharm Ther 2001, 26(5):331–342. application of the Necessity-Concerns Framework. J Psychosom Res 2008,
6. Horne R, Weinman J: Patients’beliefs about prescribed medicines and 64(1):41–46.
their role in adherence to treatment in chronic physical illness. 28. Iihara N, Tsukamoto T, Morita S, Miyoshi C, Takabatake K, Kurosaki Y: Beliefs
J Psychosom Res 1999, 47(6):555–567. of chronically ill Japanese patients that lead to intentional non-
7. Horne R, Weinman J: Self-regulation and self-management in asthma: adherence to medication. J Clin Pharm Ther 2004, 29(5):417–424.
exploring the role of illness perceptions and treatment beliefs in 29. Mardby AC, Akerlind I, Jorgensen T: Beliefs about medicines and
explaining non-adherence to preventer medication. Psychol Health 2002, self-reported adherence among pharmacy clients. Patient Educ Couns
17(1):17–32. 2007, 69(1–3):158–164.
8. Al-Qazaz H, Sulaiman SA, Hassali MA, Shafie AA, Sundram S, Al-Nuri R, 30. Menckeberg TT, Bouvy ML, Bracke M, Kaptein AA, Leufkens HG, Raaijmakers
Saleem F: Diabetes knowledge, medication adherence and glycemic JA, Horne R: Beliefs about medicines predict refill adherence to inhaled
control among patients with type 2 diabetes. Int J Clin Pharm 2011, corticosteroids. J Psychosom Res 2008, 64(1):47–54.
33(6):1028–35. Epub 2011 Nov 15 2011. 31. Sirey JA, Greenfield A, Weinberger MI, Bruce ML: Medication beliefs and
self-reported adherence among community-dwelling older adults.
9. Jamous RM, Sweileh WM, Abu-Taha AS, Sawalha AF, Zyoud SH, Morisky DE:
Clin Ther 2013, 35(2):153–160.
Adherence and satisfaction with oral hypoglycemic medications: a pilot
32. Ruppar TM, Dobbels F, De Geest S: Medication beliefs and
study in Palestine. Int J Clin Pharm 2011, 33(6):942–948.
antihypertensive adherence among older adults: a pilot study. Geriatr
10. Sweileh WM, Ihbesheh MS, Jarar IS, Taha AS, Sawalha AF, Zyoud SH, Jamous
Nurs 2012, 33(2):89–95.
RM, Morisky DE: Self-reported medication adherence and treatment
33. Neame R, Hammond A: Beliefs about medications: a questionnaire survey
satisfaction in patients with epilepsy. Epilepsy Behav 2011, 21(3):301–305.
of people with rheumatoid arthritis. Rheumatology (Oxford) 2005,
11. Horne R: Patients’beliefs about treatment: the hidden determinant of
44(6):762–767.
treatment outcome? J Psychosom Res 1999, 47(6):491–495.
34. Hedenrud T, Jonsson P, Linde M: Beliefs about medicines and adherence
12. Horne R, Clatworthy J, Polmear A, Weinman J: Do hypertensive
among Swedish migraineurs. Ann Pharmacother 2008, 42(1):39–45.
patients’beliefs about their illness and treatment influence medication
35. George CF, Peveler RC, Heiliger S, Thompson C: Compliance with tricyclic
adherence and quality of life? J Hum Hypertens 2001, 15(Suppl 1):S65–68.
antidepressants: the value of four different methods of assessment.
13. Al-Saeedi M, Elzubier AG, Bahnassi AA, Al-Dawood KM: Patterns of belief
Br J Clin Pharmacol 2000, 50(2):166–171.
and use of traditional remedies by diabetic patients in Mecca, Saudi
36. Korb-Savoldelli V, Gillaizeau F, Pouchot J, Lenain E, Postel-Vinay N, Plouin PF,
Arabia. East Mediterr Health J 2003, 9(1–2):99–107.
Durieux P, Sabatier B: Validation of a French version of the 8-item Morisky
14. Mann DM, Ponieman D, Leventhal H, Halm EA: Predictors of adherence to
medication adherence scale in hypertensive adults. J Clin Hypertens
diabetes medications: the role of disease and medication beliefs.
(Greenwich) 2012, 14(7):429–434.
J Behav Med 2009, 32:278–284.
37. Sakthong P, Chabunthom R, Charoenvisuthiwongs R: Psychometric
15. Al-Qazaz HK, Hassali MA, Shafie AA, Sulaiman SA, Sundram S, Morisky DE:
properties of the thai version of the 8-item morisky medication
The eight-item Morisky Medication Adherence Scale MMAS: translation
adherence scale in patients with type 2 diabetes. Ann Pharmacother 2009,
and validation of the Malaysian version. Diabetes Res Clin Pract 2010,
43(5):950–957.
90(2):216–221.
16. Al-Qazaz HK, Hassali MA, Shafie AA, Syed Sulaiman SA, Sundram S:
doi:10.1186/1471-2458-14-94
Perception and knowledge of patients with type 2 diabetes in Malaysia
Cite this article as: Sweileh et al.: Influence of patients’ disease
about their disease and medication: A qualitative study. Res Soc Adm
knowledge and beliefs about medicines on medication adherence:
Pharm 2011, 7(2):180–191. findings from a cross-sectional survey among patients with type 2
17. Fawzi W, Abdel Mohsen MY, Hashem AH, Moussa S, Coker E, Wilson KC: diabetes mellitus in Palestine. BMC Public Health 2014 14:94.
Beliefs about medications predict adherence to antidepressants in older
adults. Int Psychogeriatr 2012, 24(1):159–169.
18. Zyoud SH, Al-Jabi SW, Sweileh WM, Morisky DE: Relationship of treatment
satisfaction to medication adherence: findings from a cross-sectional
survey among hypertensive patients in Palestine. Health Qual Life
Outcomes 2013, 11(1):191.
19. Zyoud SH, Al-Jabi SW, Sweileh WM, Wildali AH, Saleem HM, Aysa HA,
Badwan MA, Awang R, Morisky DE: Health-related quality of life associated
with treatment adherence in patients with hypertension: a
cross-sectional study. Int J Cardiol 2013, 168(3):2981–2983.