Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Archives of Psychiatric Nursing

journal homepage: www.elsevier.com/locate/apnu

Exploring predictors of medication adherence among inpatients with

schizophrenia in Singapore's mental health settings: A non-experimental
study
⁎
Xiang Cong Thama, , Huiting Xiea, Cecilia Mui Lee Chngc, Xin Yi Seahd, Violeta Lopezb,
Piyanee Klainin-Yobasb
a
Institute of Mental Health, Singapore
b
Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
c
National University Hospital, Singapore
d
Singapore General Hospital, Singapore

A R T I C L E I N F O A B S T R A C T

Keywords: Schizophrenia is a mental disorder, which is marked by frequent relapses. The main reason for relapse is non-
Schizophrenia adherence to antipsychotics. A cross-sectional, correlational research study was conducted with a convenience
Adherence sample of 92 participants. The primary aim of this study was to explore the predictors of medication adherence
Factors among inpatients with schizophrenia hospitalised at tertiary hospitals in Singapore. Post-hoc analysis revealed
Predictors
that insight, religion, side effects, types of antipsychotics, social support from significant others, nurse-client
relationship, were significant predictive factors. Results from this study added knowledge to the nursing lit-
erature about medication adherence of schizophrenia patients and in Singapore setting.

Introduction Singapore is 0.7% (Ministry of Health, 2009). In Singapore, schizo-

Background
Schizophrenia is a psychotic disorder (American Psychiatric
Association, 2013). Persons with schizophrenia may experience hallu-
cinations, delusions, disorganized behaviours, negative symptoms and
disturbance in social functioning (American Psychiatric Association,
2013). Schizophrenia consists of a range of features. Patients may dis-
play improper affect and delusions and as a result, fear and worry may
follow (American Psychiatric Association, 2013). Cognitive impair-
ments may occur in individuals and are usually seen through beha-
viours such as inability to remember and communicate (American
Psychiatric Association, 2013). Lack of insight is also common among
individuals, but it is known as a symptom rather than a coping me-
chanism, and this symptom is one of the predictors of medication non-
adherence (American Psychiatric Association, 2013). Individuals may
sometimes be aggressive such that it will lead to violence; however, this
feature is presented only in the minority (American Psychiatric
Association, 2013).
Schizophrenia has a prevalence rate of 0.5%–0.7% globally
(American Psychiatric Association, 2013), while the prevalence rate in
phrenia is the 10th leading cause of disability, comprising of 2.7% of
disability-adjusted life years (Ministry of Health, 2014), indicating that
it is a crippling disorder.
Schizophrenia requires long-term treatment, which is unique to
individuals and usually comprised of antipsychotics and psychosocial
interventions (Ministry of Health, 2011). Antipsychotics comprise of
two classes: typical and atypical (Meltzer, 2013). Typical antipsychotics
are first-generation antipsychotics such as haloperidol and fluphena-
zine, while atypical antipsychotics are second-generation anti-
psychotics such as risperidone and olanzapine. Psychosocial interven-
tions for schizophrenia patients are cognitive-behavioural therapy,
compliance therapy, and psychoeducation (Velligan et al., 2009). De-
spite the available treatment, schizophrenia is also marked by frequent
relapses and remissions (American Psychiatric Association, 2013).
The main reason for relapse is nonadherence of antipsychotics
(American Psychiatric Association, 2013) which is estimated to be
about 50% (Lacro, Dunn, Dolder, Leckband, & Jeste, 2002).
World Health Organisation (2003) definition of adherence as “the
extent to which a person's behaviour taking medication, following a
diet, and/or executing lifestyle changes, corresponds with agreed re-
commendations from a healthcare provider” (p. 3). Therefore, in this

⁎
Corresponding author at: Nursing Training Department, Institute of Mental Health, 10 Buangkok View, Buangkok Green Medical Park, 539747, Singapore.
E-mail address: tham.xiangcong@gmail.com (X.C. Tham).

https://doi.org/10.1016/j.apnu.2018.02.004
Received 24 January 2017; Received in revised form 18 January 2018; Accepted 11 February 2018
0883-9417/ © 2018 Published by Elsevier Inc.

Please cite this article as: Tham, X.C., Archives of Psychiatric Nursing (2018), https://doi.org/10.1016/j.apnu.2018.02.004
X.C. Tham et al.
study, medication adherence is determined as behaviour to follow
medication prescriptions (antipsychotics) as ordered by physicians.
Empirical evidence suggested that medication adherence is asso-
ciated with various factors and these factors can be classified into four
categories: patient-related, medication-related, illness-related, and ex-
ternal/environment factors (Abdel-Baki, Ouellet-Plamondon, & Malla,
2012). Firstly, patient-related factors are elements or behaviours
knowingly or unknowingly caused by the patients themselves. Ex-
amples of these factors are age (Sweileh et al., 2012), education status
(Linden, Scheel, & Eich, 2006), and insight (Moritz et al., 2013). Sec-
ondly, medication-related factors are factors caused by or linked to
psychiatric or nonpsychiatric medications such as number of medica-
tions (Mullins, Obeidat, Cuffel, Naradzay, & Loebel, 2008), and side
effects of medications (McCann, Boardman, Clark, & Lu, 2008). Thirdly,
illness-related factors are any disabilities, disorders or clinical features
which are explainable by science such as psychotic symptoms (Yang
et al., 2012) and chronic medical illnesses (Sweileh et al., 2012).
Fourthly, external/environmental factors are extrinsic forces which
may affect an individual well-being. Examples of these factors are social
support (Yang et al., 2012), access to psychiatrists (McCann, Clark, &
Lu, 2009), living area (McCann, Deans, Clark, & Lu, 2008), and
healthcare plan (Farley, Hansen, Kristina, & Maciejewski, 2012).
Literature review
A literature search for published studies was performed to identify
the knowledge gaps and methodological limitations in the literature
before the research study was conducted. Results from the literature
review revealed six factors of medication adherence such as greater
awareness of illness (insight), previous history of medication ad-
herence, positive attitude towards medication, types of atypical anti-
psychotics, less severe psychotic symptoms, and social support (Tham
et al., 2016). Knowledge gaps identified were lack of research studies
conducted in inpatient settings; no studies used theoretical frameworks
which are essential in guiding the progress of the studies, and relating
the findings to the available nursing knowledge and real life applica-
tions (Burns & Grove, 2009); no studies explored the therapeutic re-
lationship of nurses and their clients affecting medication adherence;
unclear conclusion about the extent of social support affecting medi-
cation adherence; not enough studies conducted in Asian countries
especially in Singapore. Moreover, the methodological limitations in
the literature were that no studies used power analysis to determine
sample size; few studies used normality tests; most studies used one
categorical factor instead of four categories. Therefore, this research
study was designed to address the aforementioned knowledge gaps and
methodological limitations in the literature.
Study's aims
The primary aim of this study was to explore the predictors of
medication adherence among inpatients with schizophrenia at tertiary
hospitals in Singapore. The secondary aim was to examine the re-
lationship between nurse-client relationships and patients' insight.
Hypotheses
The researcher hypothesized:
a) Patient-related factors had significant predictive effects on medica-
tion adherence among inpatients with schizophrenia in Singapore.
b) Medication-related factors had significant predictive effects on
medication adherence in Singapore.
c) Illness-related factors had significant predictive effects on medica-
tion adherence in Singapore.
d) External/environmental-related factors had significant predictive
effects on medication adherence in Singapore.
2
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
e) Nurse-client relationship had a significant relationship with patients'
insight among inpatients with schizophrenia in Singapore.
Two tertiary hospitals in Singapore were selected as research set-
tings in this study.
Data was collected from 5 November 2014 to 11 March 2015 from
these hospitals. Post-hoc analysis was conducted on 5 December 2016.
Theoretical Framework
The Socio-ecological Model (SEM) (McLeroy, Bibeau, Steckler, &
Glanz, 1988) was used to guide this study. The SEM is a theoretical
framework that focuses on health behaviours as outcomes of interests.
The SEM consists of five constructs including intrapersonal factors
(IRPs), interpersonal processes and primary groups (IPs), institutional
factors (IFs), community factors (CFs) and public policies (PPs)
(McLeroy et al., 1988), which health interventions can focus on. IRPs
are personal characteristics of the individual found within self and in-
volve physiological processes. Both patient-related and illness-related
factors coincide with intrapersonal factors. CFs refer to geographical
boundaries and the only factor identified is the living areas in which the
patients live in. IFs are organisations or institutions in which in-
dividuals reside; these factors concur with “access to psychiatrists” and
medication-related factors. IPs emphasize on relationships between
individuals and coincide with social support and nurse-client relation-
ship. Lastly, PPs are policies that are implemented within the country
and are related to healthcare plan and co-payment policies. CFs, IFs, IPs
and PPs correspond with external/environmental factors. Each of these
five constructs has the ability to effect a change in health behaviour.
When changes are made in the five constructs together as whole, the
change in health behaviour is much more significant compared to
changes to one of the five constructs individuals (McLeroy et al., 1988).
Fig. 1 has been constructed for better illustration.
The secondary aim of this study is to examine the relationship be-
tween nurse-client relationships (an external/environmental factor)
and patients' insight (a patient-related factor). This aim coincides with
Hypothesis (e). Based on the SEM, the researcher postulates that IRPs
and IPs will also have a relationship with each other. There are no
factors belonging to CFs, IFs, and PPs, explored in this study. As such
their relations with external/environmental factors are in long dash dot
lines instead of solid lines. Fig. 2, which incorporates Hypothesis (e),
has been constructed for better illustration.
Methodology
Research design
This study adopted a cross-sectional and correlational research de-
sign. Convenience sampling was used to recruit potential participants.
Data in this study was collected using self-reported questionnaires.
Participants were recruited from two Singapore tertiary hospitals
(Hospital A and Hospital B). Hospital A is a 2000-bed public mental
health institution which has 50 wards. Data in this study were collected
from nine acute general psychiatric wards (Hospital A) which manage
patients with schizophrenia, and other psychiatric disorders such as
depression, bi-polar disorders, behavioural disorders and nonorganic
psychosis. Five of them were male wards while the rest were female.
Hospital B is a public general hospital, which only has one 23-bed
psychiatric ward admitting patients with schizophrenia, mood dis-
orders, and anxiety disorders and among others.
Study samples
Potential participants were invited to the study if they were: a) adult
inpatients (aged between 21 and 60) who had been diagnosed with
schizophrenia for this hospitalisation by their attending psychiatrists, b)
X.C. Tham et al. Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

Intrapersonal Community Institutional Interpersonal Public Policies

factors (IRPs) factors (CFs) factors (IFs) processes and (PPs)
primary groups
+ + + (IPs) +

Behaviour
change=Medication
adherence

Patient-related factors Illness-related factors Medication-related External/ Environmental-

factors related factors
Examples: Age, Examples: Psychotic
education status, insight symptoms, chronic Examples: Number of CFs: Living Area

medical illnesses medications, side effects IFs: Access to psychiatrists

+ + + IPs: Social Support

PPs: Healthcare plan

Fig. 1. Socio-ecological Model on factors of medication adherence among patients with schizophrenia with examples from the literature.

in stable condition as assessed by their psychiatrists (determined by the
absence of psychotic symptoms, suicidal ideation, severe depression,
such that they would not harm self and others), c) able to speak, read
and write English as the questionnaires used for this study were in
English Language. Participants were excluded if they: a) were aged <
21 years or older than 60 years, b) had psychiatric co-morbid disorders
such as affective disorder, and substance abuse.
Determination of sample size
Power analysis was used to determine the expected sample size for
this study. Sufficient sample size means that there is enough power to
detect a significant difference or a significant relationship among study
variables when it exists in reality (Burns & Grove, 2009). This means
that Type II error (a null hypothesis which is false, fails to be rejected)

Intrapersonal Community Institutional Interpersonal Public Policies

factors (IRPs) Factors (CFs) factors (IFs) processes and (PPs)
primary groups
+ + + (IPs) +

Hypothesis (e)

Behaviour
change=Medication
adherence

Patient-related factors Illness-related factors Medication-related factors External/ Environmental-

Examples: Age, gender,
ethnicity, religion,
educational level, marital
status, and insight
Examples: Number of types of related factors
Examples: Non- medications, route of medication
psychiatric co- administration, types of IPs: Social support,
morbidities and duration antipsychotics, medication doses, nurse-client relationship
types of non-anpsychoc
+ of illness + psychiatric medicine prescribed,
+
and side effects

Hypothesis (e)

Fig. 2. Socio-ecological Model on factors of medication adherence (examples are factors to be explored in the study).

3
X.C. Tham et al.
may occur when there is not enough sample (Burns & Grove, 2009).
There are four factors essential for power analysis: 1) sample size, 2)
significant level, 3) power, and 4) effect size (Cohen, 1988). If sig-
nificant level, effect size and power are known, the sample size can be
calculated by power analysis (Burns & Grove, 2009). Suresh and
Chandrashekara (2012) recommended large values of power with at
least 80%. Hence, power (1 - β) used was 0.8.
Significant level (α) is the probability of having a Type I error (a
null hypothesis is true, but it is rejected) (Burns & Grove, 2009). It is
important as it is a cut-off point to determine if the sample reflects the
population of interest or another different population (Burns & Grove,
2009). Alpha level of 0.05 or 0.01 is a common value (Burns & Grove,
2009). For this study, the alpha of 0.05 was used.
There were no other similar studies that used regression analyses.
To calculate an effect size, the research used the findings from a study
by Misdrahi, Petit, Blanc, Bayle, and Llorca (2012) for a number of
reasons: a) the participants were inpatients, b) the participants were
schizophrenia and schizoaffective patients, c) one of the aims of their
study was to determine the association between insight and therapeutic
alliance and medication adherence, and d) similar data collection tools
were used to assess insight, therapeutic alliance, and medication ad-
herence.
Pearson Correlation rho (r) values between measures of insight and
medication adherence were −0.44, −0.47, −0.48, and measures of
therapeutic alliance and medication adherence was −0.66, served as
effect sizes and thus, they were used to calculate the sample size (two-
tail) with an aid of an online calculator (https://www.statstodo.com/
SSizCorr_Pgm.php#Multiple calculations: sample size) based on the
formulae by Machin, Campbell, Fayers, and Pinol (1997) and Altman,
Machin, Bryant, and Gardner (2000). Following the calculation, the
resulting sample sizes were 38, 33, 31, and 15 with an average of 29.25.
This indicated that a minimally adequate sample size for this study
would be 30. However, the researcher recruited as many as participants
as possible to enhance external validity of research findings.
In order to minimise researcher's selection biases, the researcher
visited all wards and approached all eligible patients. In a day, morn-
ings were dedicated to male wards while afternoons were dedicated to
female wards. This ensured that both genders had equal chances of
getting selected (Grimes & Schulz, 2002).
Measurements
Dependent variables
Medication adherence was measured with Morisky Medication
Adherence Scale-8 (MMAS-8) (Morisky, Ang, Krousel-Wood, and Ward,
2008; Morisky & DiMatteo, 2011), which is a self-administered 8-item
generic adherence scale. For this study, the term “health concern” in the
scale is replaced with “schizophrenia.” There are two response cate-
gories: yes (0) and no (1). Scores of MMAS-8 ranged from 0 to 8 and
scores obtained are categorized into one of the following: high ad-
herence (score 8), medium adherence (score 6 to < 8), and low ad-
herence (< 6) (Morisky, Ang, Krousel-Wood, & Ward, 2008). This
MMAS-8 was an improved version of the 4-item and has a high relia-
bility of an internal consistency of 0.83 (Morisky et al., 2008). The
MMAS-8 is also strongly associated with antihypertensive drug phar-
macy refill adherence (Krousel-Wood et al., 2009). This information
supports the construct validity of the scale. Furthermore, the MMAS-8
has a high internal consistency with total correlations > 0.30 for each
of the eight items (Morisky et al., 2008). Moreover, the internal con-
sistency score of a Spanish version of MMAS-8 on psychiatric out-
patients was 0.75 (De las Cuevas, and Peñate, 2015). Concurrent va-
lidity of the MMAS-8 was supported by its significant correlation
(r = 0.64) with Morisky Medication Adherence Scale (4 items)
(Morisky, Green, & Levine, 1986) (Morisky et al., 2008). The MMAS-8 is
not time specific to the present moment, instead, this scale asks ques-
tions about the past, present and future intentions to be adherent to
4
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
medications. Therefore, this scale is suitable for inpatient participants
for this study.
Insight was measured with 17-item Self-Appraisal of Illness
Questionnaire (SAIQ) (Marks, Fastenau, Lysaker, & Bond, 2000). The
SIAQ comprises 4- point Likert scale, ranging from 0 (strongly disagree)
to 3 (strongly agree). Lower scores of SAIQ indicate poorer insight.
Possible scores range from 0 to 51. The internal consistency of the SAIQ
was 0.87 while test-retest reliability was 0.82 (Marks et al., 2000)
suggesting good reliability. The SAIQ also had a significant correlation
with insight scale, Scale to Assess Unawareness of Mental Disorder
(r = 0.54) (Amador et al., 1991), and one of the Positive and Negative
Syndrome Scale item, “Lack of Insight and Judgment” (r = 0.55) (Kay,
Flszbein, & Opfer, 1987). This information supports concurrent and
construct validity of the SAIQ. It was designed to measure patient's
insight towards his/her medical condition. SAIQ was first tested with
people with schizophrenia or schizoaffective disorder.
Independent variables
Participants' demographics (age, gender, ethnicity, religion, educa-
tional level, and marital status) were collected. In addition to demo-
graphics, insight would also be examined for causality to medication
adherence.
Medication-related factors (number of types of medications, route of
administration of antipsychotics, types of antipsychotics, name and
dosage of antipsychotics, types of non-antipsychotic medications pre-
scribed, and side effects) were collected. The researcher collected
clinical data from the study site's electronic Inpatient Medication
Record System (eIMR) with the assistance of staff nurses. An open
ended question was placed after the MMAS-8 to illicit responses about
medication side effects that the participants were experiencing.
Information about duration of illness and the co-morbidities was
collected. This clinical data was collected by the researcher. The co-
morbidities in this study referred to chronic illnesses. Hence, the re-
searcher had selected a few common chronic conditions identified by
Ministry of Health (2010a) under Chronic Management Programme.
Based on the medications that the patients were having at the point of
data collection, the researcher would infer the illness that the partici-
pants were having. To ensure the accuracy of the data collected, the
researcher also confirmed the co-morbidities with the staff-nurse in-
charge.
In this study, social support and nurse-client relationship were used
to represent external/environmental-related factors. Social support was
measured with the modified Social Support Network Inventory (SSNI)
(Richman, Rospenda, & Cloninger, 2009). This modified SSNI was
adapted from the original 11-item SSNI by Flaherty, Gaviria, and
Pathak (1983). The modified SSNI comprises of a 7-point scale, ranging
from 0 (very uncomfortable) to 7 (very comfortable). The higher the
score, the more social support they have (Richman et al., 2009). Pos-
sible scores can range from 16 to 112, depending the type of support
each respondent has. Respondents are expected to answer four ques-
tions regarding their support from (1) significant others (spouse/sig-
nificant other/girlfriend/boyfriend), (2) coworker, (3) friend outside of
the workplace, and (4) relative. The possible scores from each category
of social support include 4 to 28. Cronbach's alpha coefficients were
0.91 for women and 0.90 for men. The SSNI has a convergent validity
with a significant correlation with clinician's ratings (r = 0.68). A study
was conducted by comparing a closely-linked religious community
sample with an urban sample with the SSNI; the religious community
was revealed to have higher scores (Flaherty et al., 1983). Hence, this
scale has a concurrent validity.
Another external/environmental-related factor to be explored is
nurse-client relationship, which is a therapeutic relationship between
the nurse and the client (the patient) in which the client develops a trust
in the nurse to solve issues expressed by the client (Forchuk et al.,
2000). The 4-Point ordinal Alliance Self-report (4PAS; Misdrahi,
Verdoux, Lançon, & Bayle, 2009) has 11 items with response categories
X.C. Tham et al. Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

ranging from 1 (No it's wrong) to 4 (Yes it's true). Possible scores range Table 1
from 11 to 44, with higher score indicating positive therapeutic alliance Descriptive statistics of study variables.
(Misdrahi et al., 2009). Internal consistency of the 4PAS was 0.91
Variables n X SD Variance Skewness Kurtosis α
(Misdrahi et al., 2009). Furthermore, the 4PAS has a significant cor-
relation with the visual analogue items of the Helping Alliance Scale Medication adherence (n = 92)
(r = 0.62) (Priebe & Gruyters, 1993). This information supports con- High adherencea 8 8 0 0 NA NA 0.54
current validity of the scale. It is important to note that the original Medium adherencea 30 6.61 0.46 0.21 −0.32 −1.32 0.54
Low adherencea 54 3.91 1.23 1.52 −0.30 −0.82 0.54
4PAS scale measures the therapeutic alliance between medical doctors Total medication 92 5.15 1.81 3.29 −0.24 −0.91 0.54
and patients. Therefore, all the items contain the term “doctor’. With adherencea
permission from the scale developer (David Misdrahi), the researcher Insight 92 28.79 6.32 39.97 −0.61 −0.23 0.78
changed the term "doctor” to “nurse” in all questionnaire items so that Nurse-client 92 36.53 7.27 52.82 −1.28 1.81 0.93
relationship
they suit the purpose of this study.
Social supportc (n = 92)
Ethical considerations Significant othersb 54 10.72 10.57 111.83 0.33 −1.43 0.93
Co-workersb 50 9.32 9.85 96.97 0.48 −1.21 0.93
Friendsb 66 11.36 8.84 78.21 0.02 −1.21 0.93
Ethical approval was first obtained from relevant hospital and na- Relativesb 79 15.90 9.61 92.42 −0.33 −1.13 0.93
tional ethics committee before data collection began. Every eligible Total social supportb 92 47.29 25.89 670.32 0.23 −0.33 0.93
participant had equal opportunities to participate in this study. None of
a
Subgroups of medication adherence as determined by MMAS-8 (Morisky et al., 2008).
the potential participants was turned down because of other reasons not b
Subgroups of social support as determined by modified SSNI (Richman et al., 2009).
stated in the eligibility criteria. The purpose of this study, together with c
Some participants had up to four types of social support.
risks and benefits, were explained to the potential participants. These
participants had to verbalise an understanding of the study before
only medication adherence scale (MMAS-8) had a low reliability score
proceeding to sign an informed consent. The study was completely
of 0.54.
voluntary and the participants were allowed to withdraw from this
study at any time. Upon the completion of questionnaire, the partici-
pant was given $5.00 cash as a token of appreciation. Moreover, the Patient-related factors
questionnaires were completely anonymous as any personal informa-
tion such as name was not required in the questionnaires. Furthermore, The demographic profile of the participants was outlined in Table 2.
participants' answers were not revealed to anybody outside the study The participants had a mean age of 43.96 years old (SD = 9.60). The
team. Therefore, their treatment plans were not altered because of this majority of the participants were male (n = 63, 68.50%) and Chinese
study. This study also provided the participants the flexibility regarding (n = 57, 62.00%). During the data collection, some patients claimed
the time to complete the questionnaires. They were allowed extra time that they had more than one religion and were unable to identify a main
to complete if they requested so, and were allowed to stop filling the religion. Therefore, a participant could have up to four religions. The
questionnaires should they feel stressed while doing. participants mostly had Christianity as their first religion (n = 44,
47.80%). The majority had an education qualification of an ‘O' Level
Data analyses certificate (n = 21, 22.80%) and were single (n = 66, 71.70%). The
participants had mean insight of 28.79 (SD = 6.32) (see Table 1). In-
The data was entered using IBM's SPSS (Version 23) (IBM Corp., ferential analyses indicated no predictive factors of medication ad-
2015). Descriptive analyses were used to describe the study variables herence.
and participants. Open-ended questions were analysed with content
analyses. The data also underwent normality tests, specifically, ob- Medication-related factors
servation of skewness and kurtosis, normal Q-Q plot, histograms, and
Kolmogorov-Smirnov test, so as to better compare sample data to po- The common route of medication administration was oral (n = 44,
pulation parameters, and also to meet the assumptions of the ex- 47.80%). The majority of the participants were prescribed with atypical
ploratory analytical methods. Pearson Correlation Test, Chi-square test antipsychotics only (n = 42, 45.70%). The top five non-antipsychotics
for association, linear and logistic regression analyses were used to test medicine prescribed for the participants were trihexyphenidyl HCl
the study's hypotheses. Post-hoc analysis focused on the use of logistic (n = 40, 43.48%), lorazepam (n = 28, 30.43%), hydroxyzine HCl
regression to test the hypotheses of study variables on individual items (n = 26, 28.26%), sodium valproate (n = 24, 26.09%), and lactulose
of MMAS-8. syrup (n = 24, 26.09%). Majority of the participants were prescribed
five or more types of medicine (n = 59, 64.10%). The majority did not
Results complain of any side effects experienced (n = 44, 47.80%). The top
three side effects experienced were drowsiness/giddiness/sleepiness/
Five potential participants from Hospital B were referred whereas a tiredness (n = 18, 19.60%), tremors (n = 3, 3.30%), and musculoske-
total of 135 people were recommended from Hospital A. Ninety-two letal problem (n = 3, 3.30%).
questionnaires were given out and returned. Therefore, the response To ensure the same unit of measurement for this study, a chlor-
rate was calculated to be 65.71%. promazine HCl equivalent dosage (CPZ) was used. This means that each
The majority of the participants had low adherence (n = 54, antipsychotic's dosage was converted to CPZ. To facilitate the conver-
58.7%), with a mean score of 3.9 (SD = 1.23). Overall, the total ad- sion of units, recommended antipsychotic equivalences were calculated
herence score (n = 92) was 5.15 (SD = 1.81) (see Table 1). by using guidelines (National Health Service, 2012) and published lit-
erature articles (Bishara, 2010; Leucht et al., 2014; Woods, 2003). A
Reliability of measurements participant could be prescribed up to three antipsychotics. Table 3
presented the descriptions of the antipsychotics prescribed and the
Internal consistency reliability (Cronbach's alpha) of measurements corresponding dosage. The majority of the participants had risperidone
was presented in Table 1. Measurements of insight, nurse-client re- as their first antipsychotics (n = 23, 25.00%), flupentixol decanoate
lationship, and social support had Cronbach's alpha of above 0.7, in- (n = 12, 13.00%) as second, and chlorpromazine (n = 2, 2.20%) as
dicating acceptable reliability (Tavakol & Dennick, 2011). However, third.

5
X.C. Tham et al. Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

Table 2 Table 3
Demographic profile of the participants. Descriptive statistics of prescribed antipsychotics.

Demographic profile n % Antipsychoticsa n % CPZ (mg)

Gender X SD Range
Female 29 31.50
Male 63 68.50 No medicine 1 1.10 NA NA NA
Race First antipsychotic (n = 91)
Chinese 57 62.00 Asenapine 1 1.10 250 NA 250
Malay 14 15.20 Zuclopenthixol decanoate 16 17.40 358 143 100–600
Indian 12 13.00 Ziprasidone 1 1.10 133 NA 133
Eurasian 6 6.50 Amisulpride 8 8.70 500 302 200–1200
Others 3 3.30 Haloperidol 8 8.70 552 566 100–1670
Religiona Risperidone 23 25.00 245 147 66.67–533
First religion (n = 92) Flupentixol decanoate 8 8.70 563 200 400–1000
Christianity 44 47.80 Chlorpromazine HCl 1 1.10 150 NA 150
Buddhism 14 15.20 Olanzapine 17 18.50 265 117 100–500
Islam 23 25.00 Clozapine 3 3.30 417 245 275–700
Hinduism 5 5.40 Fluphenazine decanoate 1 1.10 333 NA 333
Free thinker 6 6.50 Paliperidone 1 1.10 1000 NA 1000
Second religion (n = 5) Quetiapine 3 3.30 689 192 467–800
Buddhism 2 2.20
Second antipsychotic (n = 49)
Islam 1 1.10
Zuclopenthixol decanoate 1 1.10 250 NA 250
Hinduism 1 1.10
Ziprasidone 1 1.10 250 NA 250
Taoism 1 1.10
Amisulpride 1 1.10 133 NA 133
Third religion (n = 1)
Haloperidol 2 2.20 500 0 500–500
Hinduism 1 1.10
Risperidone 5 5.40 280 110 100–400
Fourth religion(n = 1)
Flupenthixol decanoate 12 13.00 306 164 100–533
Taoism 1 1.10
Olanzapine 7 7.60 548 329 133–1200
Education status
Clozapine 3 3.30 467 115 400–600
Primary 7 7.60
Fluphenazine decanoate 3 3.30 425 130 275–500
PSLE 4 4.30
Paliperidone 3 3.30 367 252 100–600
Secondary 17 18.50
Aripiprazole 2 2.20 417 401 133–700
‘O’ Level 21 22.80
Quetiapine 2 2.20 267 94.3 200–333
‘N’ Level 6 6.50
Sulpiride 1 1.10 133 NA 133
NITEC 3 3.30
Higher NITEC 2 2.20 Third antipsychotic (n = 6)
‘A’ Level 6 6.50 Amisulpride 1 1.10 250 NA 250
Diploma 15 16.30 Haloperidol 1 1.10 133 NA 133
Degree 9 9.80 Chlorpromazine HCl 2 2.20 500 0 500–500
Masters and above 1 1.10 Olanzapine 1 1.10 333 NA 333
Others 1 1.10 Trifluoperazine 1 1.10 1200 NA 1200
Marital status
Single 66 71.70 a
Some participants had up to three antipsychotics.
Married 12 13.00
Separated 2 2.20
Divorced 10 10.90
Illness-related factors
Widowed 2 2.20
Number of years diagnosed with schizophrenia The majority of the participants had duration of illness with
Less than one year 16 17.40 10 years and above (n = 50, 54.30%) (see Table 2). In addition, more
1–4 years 14 15.20
than half of participants (n = 47, 51.10%) had no co-morbidities. Most
5–9 years 12 13.00
10 years and above 50 54.30 of them had only one co- morbidity (n = 27, 29.30%). Of which, the
majority of the participants had lipid disorders as their first co-mor-
a
Some participants had up to five religions. bidity (n = 30, 32.60%), hypertension (n = 10, 10.90%) as second,
diabetes (n = 7, 7.60%) as third.
Each of the participants might have been prescribed more than one Inferential analyses indicated no illness-related predictive factors of
antipsychotic by their physicians. In this study, some participants were medication adherence.
prescribed up to three antipsychotics each. The mean dose for the first
antipsychotic was 369.43 (SD = 268.59) mg; second was 202.45
(SD = 231.27) mg; third was 37.32 (SD = 176.65) mg (See Table 3). External/environmental-related factors
Univariate logistic regression analysis ascertained that only first
antipsychotics dosage had significant predictive effects on medication 4PAS (Misdrahi et al., 2009) indicated a mean score of 36.53
adherence (p = 0.04, OR = 1.00). (SD = 7.27) for nurse-client relationship (see Table 1). Types of social
Multivariate analysis was conducted to confirm the predictive effect support were categorized into significant others, colleagues, friends,
of antipsychotics dosage on medication adherence. These variables and relatives. Of the four categories, social support from relatives had
(first antipsychotic dosage, significant others, first religion, marital the highest frequency (n = 79, 85.90%) with a mean score of 15.90
status) were entered as co-variants in the logistic model. This model (SD = 9.61) compared to social support from significant others with a
was statistically significant, χ2 (10.84), p = 0.03. This model explained lower frequency (n = 54, 58.70%) and a lower mean score of 10.72
15.0% of the cases and appropriately classified 69.6% of the cases. (SD = 10.57). The mean score of the total social support received by
Based on Hosmer and Lemeshow Test, the model is a good fit, χ2 (4.55), the participants was 47.29 (SD = 25.89) (See Table 1). Inferential
p = 0.80. Antipsychotics dosage had no predictive effect based on this analyses indicated no external/environmental-related predictive factors
regression analysis (p = 0.06) (see Table 4). Hence, there were no of medication adherence.
medication-related predictive factors of medication adherence.

6
X.C. Tham et al. Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

Table 4
Multivariate Logistic Regression Analysis of medication adherence.

Variables b SE Wald P-value OR 95% CI for OR

Lower Upper

First antipsychotics dosage −0.002 0.001 3.45 0.06 1.00 1.00 1.00
Significant others 0.03 0.02 2.03 0.15 1.03 0.99 1.08
Religion −0.52 0.47 1.24 0.27 1.69 0.67 4.25
Marital status 0.92 0.54 2.95 0.09 0.40 0.14 1.14
Model χ2= 10.84 P < 0.05
Pseudo R2= 0.15
n= 92

Relationship between insight and nurse-client relationship Table 5

Univariate Logistic Regression Analysis of each of items of MMAS-8.
Simple linear regression analysis was performed to ascertain the
Variables b SE Wald P- OR 95% CI for OR
relationship. However, nurse-client relationship had no significant re- value
lationship with patients' insight among inpatients with schizophrenia in Lower Upper
Singapore (F(1, 90) = 0.038, p = 0.85). This finding does not support
the hypothesis that nurse-client relationship had a significant re- Item 1: Do you sometimes forget to take your schizophrenia pills?
Insight 0.07 0.04 4.21 0.04⁎ 1.08 0.999 1.15
lationship with patients' insight among inpatients with schizophrenia in
Singapore. Item 2: People sometimes miss taking their medications for reasons other than
forgetting. Thinking over the past two weeks, were there any days when you
did not take your schizophrenia medicine?
Post Hoc analyses Significant others −0.04 0.02 4.32 0.04⁎ 0.96 0.91 0.99
Second 0.002 0.001 4.13 0.04⁎ 1.002 1.00 1.01
antipsychotic
Each of the independent variables was further analysed with each of dosage
the eight items of MMAS-8 using logistic regression analysis. Univariate Item 3: Have you ever cut back or stopped taking your medication without telling
logistic regression analysis was first conducted to determine predictive your doctor, because you felt worse when you took it?
factors of adherence (see Table 5). Multivariate analysis was then Route of −1.099 0.46 5.74 0.02⁎ 0.33 1.14 0.82
medication
conducted on items of MMAS-8 which had more than one predictive
administration
factor (see Table 6) to ensure that extraneous variables were controlled. (Both oral and
Variables (social support from significant others and second anti- injection)
psychotics dosage) were entered as co-variants in the logistic model Second 0.002 0.001 5.26 0.02⁎ 1.002 1.00 1.004
antipsychotic
(Item 2). This model was statistically significant, χ2 (10.16), p = 0.01.
dosage
This model explained 14.7% of the cases and appropriately classified
68.5% of the cases. Based on Hosmer and Lemeshow Test, the model is Item 4: When you travel or leave home, do you sometimes forget to bring along
your schizophrenia medication?
a good fit, χ2 (8.21), p = 0.41. Only social support from significant No predictive factors
others has predictive effect based on this regression analysis (p = 0.02)
Item 5: Did you take your schizophrenia medicine yesterday?
(see Table 6).
Insight −0.11 0.05 6.00 0.01⁎ 0.90 0.82 0.98
Route of medication administration and second antipsychotics do- Number of side 1.24 0.49 6.32 0.01⁎ 3.46 1.32 9.12
sage were entered as co-variants in the logistic model (Item 3). This effects
model was statistically significant, χ2 (7.50), p = 0.02. This model ex- First side effects 1.24 0.49 6.32 0.01⁎ 3.46 1.32 9.12
plained 10.6% of the cases and appropriately classified 65.2% of the Item 6: When you feel like your schizophrenia is under control, do you sometimes
cases. Based on Hosmer and Lemeshow Test, the model is a good fit, χ2 stop taking your medicine?
(4.24), p = 0.37. There were no factors which have predictive effect No predictive factors

based on this regression analysis (see Table 6). Item 7: Taking medication everyday is a real inconvenience for some people. Do
Insight and number of side effects were entered as co-variants in the you ever feel hassled about sticking to your schizophrenia treatment plan?
First religion 1.17 0.44 6.98 0.01⁎ 3.23 1.36 7.71
logistic model (Item 5). This model was statistically significant, χ2
(Christianity)
(12.06), p = 0.02. This model explained 17.7% of the cases and ap-
Item 8: How often do you have difficulty remembering to take all your
propriately classified 73.9% of the cases. Based on Hosmer and
medications?
Lemeshow Test, the model is a good fit, χ2 (4.15), p = 0.84. Both fac- Insight 0.11 0.04 7.68 0.01⁎ 1.11 1.03 1.20
tors (insight, p = 0.03; number of side effects, p = 0.03) have pre- Types of 1.17 0.47 6.11 0.01⁎ 3.23 1.28 8.19
dictive effects based on this regression analysis (see Table 6). antipsychotics
These variables (insight, types of antipsychotics, nurse-client re- (Both typical
and atypical)
lationship) were entered as co-variants in the logistic model (Item 8).
Nurse-client 0.08 0.03 5.02 0.03⁎ 1.08 1.01 1.15
This model was statistically significant, χ2 (19.92), p < 0.001. This relationship
model explained 26.0% of the cases and appropriately classified 70.7%
⁎
of the cases. Based on Hosmer and Lemeshow Test, the model is a good p < 0.05.
fit, χ2 (13.03), p = 0.11. The three factors (insight, p = 0.03; types of
antipsychotics, p = 0.03; nurse-client relationship, p = 0.02) have of medication adherence. Participants whose religion is Christianity,
predictive effects based on this regression analysis (see Table 6). were 3.23 times more likely to stick to their schizophrenia treatment
plan (p < 0.05) (Item 7).
Participants with better insight, were 1.08 times more likely to re-
Patient-related factors
member to take their schizophrenia pills (p < 0.05) (Item 1).
Moreover, those with poor insights were 0.90 times less likely to take
Religion and insight had shown to be significant predictive factors

7
X.C. Tham et al. Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx

Table 6
Multivariate logistic regression analysis of items of MMAS-8.

Variables b SE Wald P-value OR 95% CI for OR

Lower Upper

Item 2: People sometimes miss taking their medications for reasons other than forgetting. Thinking over the past 2 weeks, were there any days when you did not take your
schizophrenia medicine?
Significant others −0.05 0.02 5.29 0.02⁎ 0.95 0.91 0.99
Second antipsychotic dosage 0.002 0.001 3.057 0.08 1.002 1.000 1.004
Model χ =
2
10.2 P < 0.05
Pseudo R2= 0.15
n= 92

Item 3: Have you ever cut back or stopped taking your medication without telling your doctor, because you felt worse when you took it?
Route of medication administration (Both oral and −0.72 0.56 1.63 0.20 0.49 1.63 1.47
injection)
Second antipsychotic dosage 0.001 0.001 1.36 0.24 1.001 0.999 1.004
Model χ2= 7.50 P < 0.05
Pseudo R2= 0.11
n= 92

Item 5: Did you take your schizophrenia medicine yesterday?

Insight −0.10 0.05 4.64 0.03⁎ 0.90 0.82 0.99
Number of side effects (No side effects experienced) 1.10 0.51 4.64 0.03⁎ 2.99 1.10 8.12
Model χ =
2
12.1 P < 0.05
Pseudo R2= 0.18
n= 92

Item 8: How often do you have difficulty remembering to take all your medications?
Insight 0.11 0.41 7.06 0.01⁎ 1.11 1.03 1.21
Types of antipsychotics (Both typical and atypical) 1.05 0.51 4.29 0.04⁎ 2.87 1.06 7.76
Nurse-client relationship 0.09 0.04 5.74 0.02⁎ 1.09 1.02 1.17
Model χ =
2
19.9 P < 0.05
Pseudo R2= 0.26
n= 92

⁎
p < 0.05.

their medications in inpatient settings (p < 0.05) (Item 5). In addition,
those with better insights were 1.11 times more likely to remember to
take all their medications (p < 0.05) (Item 8).
These findings support the hypothesis that patient-related factors,
including religion and insight, had significant predictive effects on
medication adherence among inpatients with schizophrenia in
Singapore.
Medication-related factors
The number of side effects and types of antipsychotics (both typical
and atypical antipsychotics) had shown to be significant predictive
factors of medication adherence. Participants who experienced no side
effects from their medications, were 2.99 times more likely to take their
medication the previous day (p < 0.05) (Item 5). Participants who
were prescribed with both types of antipsychotics were 4.29 times more
likely to remember to take their medications (p < 0.05) (Item 8).
These findings support the hypothesis that medication-related fac-
tors, including side effects and types of antipsychotics, had significant
predictive effects on medication adherence among inpatients with
schizophrenia in Singapore.
Illness-related factors
There were no predictive factors of medication adherence. This
finding does not support the hypothesis that illness-related factors had
significant predictive effects on medication adherence among inpatients
with schizophrenia in Singapore.
External/environmental-related factors
Support from significant others and nurse-client relationship had
shown to be significant predictive factors of medication adherence.
Those with poor support from their significant others, were 0.95 times
less likely to take their medicine (p < 0.05) (Item 2). Participants who
had good therapeutic relationships with their nurses, were 1.09 times
more likely to remember to take all their medications (p < 0.05) (Item
8).
These findings support the hypothesis that external/environmental-
related factors, including social support and nurse-client relationship,
had significant predictive effects on medication adherence among in-
patients with schizophrenia in Singapore.
Discussion
The hypotheses as introduced earlier in the manuscript, sought to
examine if the patient [hypothesis (a)], medication [hypothesis (b)],
illness [hypothesis (c)], external/environmental-related factors [hy-
pothesis (d)] had any predictive effects on medication adherence.
Moreover, the relationship between nurse-client relationship and in-
sight was also explored [hypothesis (e)]. The findings supported the
hypotheses (a), (b), and (d). However, hypotheses (c) and (e) were not
supported by the findings.
With reference to the SEM, the illness-related factors do not con-
tribute to any behavioural change. This leads to the researcher to query
if the SEM construct, intrapersonal factors (IRPs), have any relationship
with illness-related factors given that illness-related factors do not even
exist in the model. This also means that clinicians can just focus on
patient, medication and external/environmental- related factors to en-
sure patient's medication adherence rather than all the four factors.
Moreover, based on the findings, nurse-client relationship does not
have any significant relationship with insight. This contributes to the
possibility that IRPs and IPs have no significant relationship with each
other. This finding also reinforces the model by proving that each
construct remains as a unique individual which contributes to beha-
vioural change.
According to SEM, if all five constructs were addressed, the effect of
behavioural change will be more significant than that if one construct is
addressed. Three out of the five constructs (CFs, IFs and PPs) are not
explored in this study. Although this study is comprehensive in

8
X.C. Tham et al.
Intrapersonal Community Institutional
factors (IRPs) Factors (CFs) factors (IFs)
+ +
Hypothesis (e)
?
Patient-related factors Illness-related factors
Examples: religion,
insight
+ +
Hypothesis (e)
Fig. 3. Socio-ecological Model on factors of medication adherence based on study findings.
exploring predictive factors, the net is still not wide enough to cover
factors derived from the three constructs. Hence, it is possible that
behavioural change in the patients may fail even if the predictive fac-
tors from the findings are addressed by clinicians. Fig. 3 is created for
better visualisation of the above contents.
Relationship between patient-related factors and medication adherence
Insight is one of the predictive factors in this study. The participants
had low insight ( X =28.7, SD = 6.32). This finding is consistent with
the reason to why these participants were inpatients in the first place.
This finding is also congruent to Rittmannsberger, Pachinger,
Keppelmüller, and Wancata's (2004) study whereby nearly half of the
patients (n = 44, 46.3%) had poor insight upon admission. Further-
more, literature had shown that insight is significantly related to
medication adherence (Tham et al., 2016). Insight represents a certain
amount of knowledge that the patients possess regarding their illness
(Rüsch & Corrigan, 2002). If that knowledge falls short, patients may
not able to discern the importance of adhering to their medication re-
gimen and thus, become non-adherent.
Religion, specifically, Christianity, is a significant predictor of
medication adherence. Compared to other religions, Christians have
different cultural experiences and beliefs (Furnham & Wong, 2007).
They tend to be less stigmatized towards their own religious members
who are diagnosed with schizophrenia (Richards, Hori, Sartorius, &
Kunugi, 2014), and focused on treatments based on sciences rather than
superstitions (Furnham & Wong, 2007). Hence, it is likely that those
with Christianity have higher recovery chances, compared to other
religions, as they are likely to take their medications.
Relationship between medication-related factors and medication adherence
The majority of the participants did not report any side effects.
Hence, those who reported no side effects, were significantly likely to
be adherent to medication. This means that those who experienced side
effects from medications, have a higher risk of medication non-
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
Interpersonal Public Policies
processes and (PPs)
primary groups
+ (IPs) +
Behaviour
change=Medication
adherence
Medication-related External/ Environmental-
factors related factors
Examples: side effects, IPs: Social Support,
types of antipsychotics nurse-client relationship
(both atypical and +
typical)
adherence. Side effects can bring bodily discomfort which can be in-
tolerable to an individual (Dolder, 2008). All of the reported side effects
in this study are common side effects derived from antipsychotics
(Dolder, 2008). Hence, absence of side effects can better promote
medication adherence.
Types of antipsychotics (both typical and atypical) is a significant
predictive factor of medication adherence. It is likely that those who
were prescribed with one type of antipsychotics (typical only or aty-
pical only) is not enough to improve the mental state of individuals. If
psychotic symptoms improve, patients are likely to see the benefits of
being adherent to medications. A meta-analysis revealed that patients
who were prescribed with antipsychotics augmentation, showed im-
provement in the psychiatric symptoms compared to those who were on
monotherapy (Correll, Rummel-Kluge, Corves, Kane, & Leucht, 2009).
In the sensitivity analysis, combination therapy with atypical and ty-
pical antipsychotics cotreatment was significantly superior than
monotherapy (Correll et al., 2009). Hence, those who were prescribed
with both typical and atypical antipsychotics, were likely to be ad-
herent to medications.
Among the antipsychotics prescribed, risperidone was prescribed
the most for first antipsychotics, followed by flupentixol decanoate and
chlorpromazine HCl. The majority of Yang et al.'s (2012) participants
were also prescribed with risperidone (n = 17, 33.4%). Combination of
antipsychotics is also necessary as it can be effective for some patients
who display different psychotic symptoms (Ito, Koyama, & Higuchi,
2005). However, it is interesting to note that depot antipsychotics,
specifically fluphenazine was not prescribed regularly. Fluphenazine
decanoate is not only inexpensive, but reaches serum concentration
peak within 10 h and typically lasts up to four to six weeks (McEvoy,
2006). The side effects derived from fluphenazine is comparable to
many oral atypical antipsychotics (McEvoy, 2006).
Relationship between illness-related factors and medication adherence
The factors explored in this study included non-psychiatric co-
morbidities and duration of illness. However, none of the predictors had
9
X.C. Tham et al.
significant relationships with medication adherence. Possible reasons
might be due to the limitations of this study, which will be elaborated
later in the article.
Relationship between external/environment factors-related factors and
medication adherence
Compared to social support from coworkers, friends outside of the
workplace, and relatives, support from significant others is a significant
factor of medication adherence. There are two possible reasons for this
phenomenon. Firstly, mental health patients are constantly subjected to
stigmatisation, hence, they may not want to divulge about their con-
ditions to others except for their loved ones whom they considered
close (significant others) (Link & Phelan, 2006). Secondly, the public
may have poor mental health literacy and do not understand the ec-
centric behaviours displayed by some of the mental health patients.
Hence, the public may try to distance themselves from these patients
(Jorm, 2000). Patients' significant others, on the other hand, may em-
pathise with patient (Goubert et al., 2005) and provided adequate as-
sistance (tangible and non-tangible) to patients such that they will be
adherent to medications.
Nurse-client relationship is a significant predictor of medication
adherence. In addition, it had a high score of 36.53 (SD = 7.27) out of
44. This result is expected in the mental health settings as nurses have
more time to spend with patients compared to general medical-surgical
ward settings. It is verified that mental health nurses are able to make
an individual feel unique, and they know about the patients' private
lives better than other healthcare professionals (Shattel, Starr, &
Thomas, 2007), hence, nurses are crucial to promote the importance of
medication adherence to schizophrenia patients.
Relationship between nurse-client relationship and insight
Both correlation and linear regression analyses revealed that nurse-
client relationship had no significant relationship with insight. Nurse-
client relationship is crucial in patient care as it can influence patient's
willingness to communicate to the nurse (Shattel, 2004). Nurses are in a
better position to impart illness knowledge to their patients than other
healthcare professionals (Turkington et al., 2006). Hence, it is sur-
prising that no significant relationship was detected between the two
variables.
Application of results
Nursing science
There are no similar studies conducted in Singapore. Hence, this
study contributed to the current nursing knowledge of factors affecting
medication adherence in Singapore among schizophrenia inpatients.
Moreover, this study is the first to conduct in a heterogeneous society
where there are different races, religions and cultures co-exist in a lo-
cation. In addition, there are no studies in the existing literature which
explore nurse-client relationship. Therefore, this study will add to the
knowledge that nurse-client relationship predicts medication ad-
herence, but not insight. Additionally, the use of SEM as a theoretical
framework is useful to guide similar researches. Moreover, the re-
searcher used power analysis to determine expected sample size, and
determined if the data is normal or not by using normality tests. The
researcher not only minimised methodological limitations created by
previous researches, by doing so, the researcher's findings produce
more reliable results compared to previous research studies.
Nursing practice
Moreover, it is important to review essential medicines for the pa-
tients. A study was shown that psychiatrists usually prescribe certain
medicines based on the nurses' recommendations (Ito et al., 2005).
Since the presence of side effects can lead to nonadherence to
10
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
medication, it is imperative that nurses observe for any side effects
displayed by patients who are prescribed with antipsychotics. Nurses
should report any side effects and offer alternative suggestions to an-
tipsychotics or medications to reduce side effects such as benzhexol, to
the attending psychiatrists. Hence, mental health nurses play an im-
portant role in making critical decision if any of the medicines is ne-
cessarily prescribed for their patients.
Psychosocial interventions to promote medication adherence and
insight include motivational interviewing, psychoeducation, and family
therapy (Zygmunt, Olfson, Boyer, & Mechanic, 2002). Nursing leaders
can review which of these interventions are effective for the local po-
pulation.
The majority of the participants had lipid disorders as their first co-
morbidity, followed by hypertension, diabetes. These three diseases
have the most prevalence rates in Singapore population (Ministry of
Health, 2010b). Most schizophrenia patients tend to develop these three
chronic conditions compared to the general population. This is likely
because of metabolic abnormalities as a result of taking antipsychotics,
and unhealthy and sedentary lifestyle adopted by these patients (Liao
et al., 2011). Exercise and diet programmes can be set up to encourage
healthy living for these patients. Examples of exercise programmes are
resistance and aerobic exercises (Marzolini, Jensen, & Melville, 2009),
and cardio-workout exercise programmes (Scheewe et al., 2013). The
benefit of exercising is not only to maintain a healthy lifestyle, but also
reduce schizophrenia symptomatology and affective symptoms and
improve cognitive performances (Malchow et al., 2013). Diet man-
agement will require the involvement of dieticians to formulate diets
specially to reduce weight or maintain a healthy diet (Kwon et al.,
2006).
Furthermore, based on the descriptive results, patients have good
therapeutic relationships with their nurses. Hence, nurses should con-
tinue to be the key pillars of support for them as these patients receive
poor social support overall according to the results. By doing so, it can
also further enhance nurse-client relationships. Support from significant
others is predictive of medication adherence. Interventions to enhance
social support should include patients' significant others involving in
the treatment, and informing and addressing concerns about schizo-
phrenia (McFarlane, Dixon, Lukens, & Lucksted, 2003).
Hospital policies
Nurses do not dictate or emphasize any religion. Hence, the pro-
motion of any religion, in this case, Christianity, is beyond the scope of
nursing practice. A better alternative will be the introduction of peer
support services in inpatient settings. To the researcher's knowledge,
this service is not available in the acute inpatient settings. In this ser-
vice, a healthy and recovering individual who is diagnosed with schi-
zophrenia, is assigned to a patient who has schizophrenia as well. This
healthy individual, acts as a role model, and provides support and
personal experiences in this mental illness (Dixon et al., 2010). There
are studies which had shown the efficacy of such services, however,
strong evidences are still required to prove improved recovery and
medication adherence in patients who have used this service (Dixon
et al., 2010).
Poor insight is a significant predictive factor of medication non-
adherence which can lead to relapse of illness. Improvement of insight
does not occur only during inpatient settings, but also outpatient set-
tings. Local hospitals can introduce Assertive Community Treatment
(ACT). ACT is a program whereby outpatients are assigned to multi-
disciplinary teams and these teams provide services round the clock
including weekends. These teams will visit their allocated patients in
their homes to ensure that they have taken their medications and
provide various interventions which include psychoeducation about
illness and medications. ACT has also proven to be a successful program
to promote medication adherence among psychiatric patients
(Appelbaum & Le Melle, 2008).
Local hospitals should also have their own policies to promote the
X.C. Tham et al.
use of depot antipsychotics. As mentioned earlier in the article, flu-
phenazine decanoate has several advantages but it is underused in the
local settings. Fluphenazine decanoate, a typical antipsychotic, can be
used additionally with atypical antipsychotics to enhance medication
adherence. Hence, hospitals can introduce policies to alter prescribing
patterns of psychiatrists.
National policies
Given the positive significance in nurse-client relationships, the
government can provide resources so that nurses are well-trained to
provide interventions to patients. Although not explored in this study, it
is possible that stigmatisation may lead to medication nonadherence
(Link & Phelan, 2006). Nurses can also collaborate with local hospitals
to start anti-stigmatisation campaigns. These campaigns involve edu-
cating the public through various means such as the schools, work
places, media, and inviting mental health experts to address any
stigmas. Another strategy is to prevent discrimination by ensuring that
the media do not propagate stigmatisation sentiments, and have
workplace ethics in the nation by employing people based on qualifi-
cations rather than assessing their mental illness (Rüsch, Angermeyer, &
Corrigan, 2005). These forms of campaigns have been proven successful
in USA and Germany (Rüsch et al., 2005). Moreover, just like in Austria,
the government can also encourage school students to volunteer in
mental health settings (Rüsch et al., 2005).
Strengths of the study
This study has its strengths. Firstly, this study was conducted in the
natural settings where the participants were situated. This means that
there were no changes made to the study sites to facilitate the conduct
of this study, and hence, supported external validity of the findings
(Mann, 2003). Moreover, to further promote external validity, the re-
searcher also collected data from two hospitals (Ferguson, 2004). Sec-
ondly, to promote internal validity of this study, the researcher mini-
mised selection biases by dedicating mornings to male wards and
afternoons to female wards. This ensured that both genders had equal
chances of getting selected (Grimes & Schulz, 2002). Thirdly, unlike
other studies, the researcher adopted SEM to guide this study. A fra-
mework is important in guiding the progress of the research study, and
relating the findings to the available nursing knowledge and real life
applications (Burns & Grove, 2009). Fourthly, the researcher used
power analysis to determine expected sample size. Based on the ex-
pected sample size, the researcher collected adequate sample size. Fif-
thly, normality tests were used in this study for the purpose of testing
assumptions of the inferential analyses. Normality tests can also help to
identify non-linear data, and appropriate tools can be used to examine
these data (Steinskog et al., 2007). Lastly, the researcher tested the
assumptions of the different inferential analyses used. Testing of as-
sumptions is important as it will avoid either Type I or II error, or
under- or over-estimation of significance (Osborne & Waters, 2002).
Limitations of the study
This study contains the following limitations: (a) Although it had
exceeded expected sample size of 30 participants, this study's sample
size of 92 was still inadequate to detect any potential relationships
between the variables. It is possible that because of this small sample
size, it failed to detect significant associations or relationships to
medication adherence and insight in this study (Bartlett, Kotrlik, &
Higgans, 2001), which explains why there were no predictive factors of
medication adherence identified; (b) It is likely that due to the homo-
geneity of the sample (the selection criteria were too narrowed), the
MMAS-8 had a low internal consistency reliability in this study as a
result (Mehrens & Lehman, 1991). Because of this, this study could not
detect significant associations or relationships; (c) This study is a cross-
sectional study, hence, by conducting one time cannot confirm the
11
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
causal relationships between the variables (Song & Chung, 2010); (d)
Convenience sampling was adopted and the sample might not be re-
presentative of population (Burns & Grove, 2009); (e) This study only
included those who understood English, the results might not be ap-
plicable to those who speak other languages only; (f) Self-reported
questionnaires used in this study could attract social desirable answers,
thus not reflecting the reality (Kikkert et al., 2008); (g) There were
some participants who spoke basic English and needed further clar-
ification during the completion of the questionnaires. Explanations
given to these participants might be inferred differently for each in-
dividual. However, the researcher did not record which of the partici-
pants required explanations; (h) Despite the variables (nurse-client re-
lationship and insight) were not normal, the researcher proceeded to
conduct linear regression analysis. This would result in either Type I or
II error, or under- or over-estimation of significance (Osborne & Waters,
2002), hence, a possibility of not detecting any significant result; (i)
This study did not explore the strength of relationship between each of
the SEM constructs and the individual factors. Not exploring the
strengths of such relationships may bring about the possibility that the
SEM may not be applicable to real life settings.
Recommendations for future research
Based on the limitations as presented earlier, future quantitative
studies can focus on (a) increasing the sample size from the two sites,
(b) recruiting schizophrenia patients with psychiatric co-morbidities to
enhance heterogeneity of the data, (c) conducting a longitudinal study
to confirm causal relationship over time, (d) adopting random sampling
instead of convenience sampling, (e) recruiting non-English speaking
participants, (f) using an objective instruments or using both clinician-
rated adherence and self-reported adherence scales at the same time,
(g) using other self-reported adherence scales, (h) using other appro-
priate inferential analyses based on statistical assumptions, (i) identi-
fying which specific medicines can promote medication adherence, (j)
measuring the strength of relationship between each of the SEM con-
structs and the individual factors, (k) exploring other potential pre-
dictive factors from the SEM constructs, CFs, Ifs, and PPs, such as
(substance and alcohol use, use of emergency services, frequency of
outpatient visits, frequency of hospitalizations, patient's personality,
age of onset, functioning, fear of medicine, frequency of medications,
mood, cognitive impairment, specific schizophrenia diagnosis, physical
disability, presence of healthcare plan, living area, ability to co-pay,
presence of medical supervision, accompany for appointments, and
convenient access to psychiatrists).
Future qualitative studies can explore the reasons of medical ad-
herence/non-adherence and the feedback on the interventions provided
by the hospitals to improve medication adherence. By doing so, po-
tential researchers are able to discover deeper knowledge that mea-
suring scales cannot explore. Additionally, phenomenological studies
should be conducted to illicit the experiences of the informants when
they take their medications.
Conclusion
In conclusion, this study revealed that there were six factors (in-
sight, religion, side effects, types of antipsychotics, social support from
significant others, and nurse-client relationship) which are significant
predictive factors of medication adherence in Singapore. Nurse-client
relationship does not have any significant relationship with insight.
SEM was used as a theoretical framework to guide the study. This study
contributed to the existing nursing knowledge of medication adherence
among schizophrenia patients in Singapore. The researcher managed to
reduce several methodological limitations and addressed issues derived
from knowledge gaps through this study. Descriptive and inferential
statistics from this study had proven useful. Interventions, based on the
results, can be carried out in the clinical, hospital and national level to
X.C. Tham et al.
promote medication adherence (such as improving insight, im-
plementation of peer support services in inpatient settings, reducing
side effects, introducing both typical and atypical antipsychotics aug-
mentation medication regime, and emphasis of social support from
significant others), enhance healthy lifestyle, and reduce public stig-
matisation towards schizophrenia patients. Moreover, further research
can be conducted to ascertain factors of medication adherence among
schizophrenia patients.
Acknowledgment
The authors have received no financial support from any organisa-
tion for the research, authorship, and/or publication of this paper.
References
Abdel-Baki, A., Ouellet-Plamondon, C., & Malla, A. (2012). Pharmacotherapy challenges
in patients with first-episode psychosis. Journal of Affective Disorders, 138, S3–S14.
http://dx.doi.org/10.1016/j.jad.2012.02.029.
Altman, D. G., Machin, D., Bryant, T. N., & Gardner, M. J. (2000). Statistics with confidence
(2nd ed.). UK: BMJ Books.
Amador, X. F., Strauss, D. H., Yale, S. A., & Gorman, J. M. (1991). Awareness of illness in
schizophrenia. Schizophrenia Bulletin, 17(1), 113–132. Retrieved from http://
schizophreniabulletin.oxfordjournals.org/content/17/1/113.full.pdf.
American Psychiatric Association (2013). Diagnostic and statistical manual of mental dis-
orders (5th ed.). Arlington, VA: American Psychiatric Association.
Appelbaum, P. S., & Le Melle, S. (2008). Techniques used by assertive community
treatment (ACT) teams to encourage adherence: Patient and staff perceptions.
Community Mental Health Journal, 44(6), 459–464. http://dx.doi.org/10.1007/
s10597-008-9149-4.
Bartlett, J. E., Kotrlik, J. W., & Higgans, C. C. (2001). Organizational research:
Determining appropriate sample size in survey research. Information technology,
learning, and performance journey. 19(1). Information technology, learning, and perfor-
mance journey (pp. 43–50). . Retrieved from ftp://149.222-62-69.ftth.swbr.surewest.
net/TreePDF/Determing%20Appropriate%20Sample%20Size%20in%20Survey
%20Research.pdf.
Bishara, D. (2010). Once-monthly paliperidone injection for the treatment of schizo-
phrenia. Neuropsychiatric Disease and Treatment, 6, 561–572. http://dx.doi.org/10.
2147/NDT.S8505.
Burns, N., & Grove, S. K. (2009). The practice of nursing research. Appraisal, synthesis, and
generation of evidence(6th ed.). MO, Missouri: Elservier Inc.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed). New York:
Academic Press.
Correll, C. U., Rummel-Kluge, C., Corves, C., Kane, J. M., & Leucht, S. (2009).
Antipsychotic combinations vs monotherapy in schizophrenia: A meta-analysis of
randomized controlled trials. Schizophrenia Bulletin, 35(2), 443–457. http://dx.doi.
org/10.1093/schbul/sbn018.
de las Cuevas, C., & Peñate, W. (2015). Psychometric properties of the eight-item Morisky
Medication Adherence Scale (MMAS-8) in a psychiatric outpatient setting. International
Journal of Clinical and Health Psychology. Advance online publication https://doi.
org/10.1016/j.ijchp.2014.11.003.
Dixon, L. B., Dickerson, F., Bellack, A. S., Bennett, M., Dickinson, D., Goldberg, R. W., ...
Kreyenbuhl, J. (2010). The 2009 schizophrenia PORT psychosocial treatment re-
commendations and summary statements. Schizophrenia Bulletin, 36(1), 48–70.
http://dx.doi.org/10.1093/schbul/sbp115.
Dolder, C. R. (2008). Side effects of antipsychotics. In K. M. Mueser, & D. V. Jeste (Eds.).
Clinical handbook of schizophrenia (pp. 168–177). New York, NY: The Guilford Press.
Farley, J. F., Hansen, R. A., Kristina, S., & Maciejewski, M. L. (2012). Antipsychotic ad-
herence and its correlation to health outcomes for chronic comorbid conditions. The
primary care companion to CNS disorders. 14(3). PCC.11m01324 https://doi.org/10.
4088/PCC.11m01324.
Ferguson, L. (2004). External validity, generalizability, and knowledge utilization.
Journal of Nursing Scholarship, 36(1), 16–22. http://dx.doi.org/10.1111/j.1547-5069.
2004.04006.x.
Flaherty, J. A., Gaviria, F. M., & Pathak, D. S. (1983). The measurement of social support:
The social support network inventory. Comprehensive Psychiatry, 24(6), 521–529.
http://dx.doi.org/10.1016/0010-440X(83)90019-6.
Forchuk, C., Westwell, J., Martin, M. L., Bamber-Azzapardi, W., Kosterewa-Tolman, D., &
Hux, M. (2000). The developing nurse-client relationship: Nurses' perspectives.
Journal of the American Psychiatric Nurses Association, 6(1), 3–10. http://dx.doi.org/
10.1177/107839030000600102.
Furnham, A., & Wong, L. (2007). A cross-cultural comparison of British and Chinese
beliefs about the causes, behaviour manifestations and treatment of schizophrenia.
Psychiatry Research, 151, 123–138. http://dx.doi.org/10.1016/j.psychres.2006.03.
023.
Goubert, L., Craig, K. D., Vervoort, T., Morley, S., Sullivan, M. J. L., de CAC, W., ...
Crombez, G. (2005). Facing others in pain: The effects of empathy. Pain, 118(3),
285–288. http://dx.doi.org/10.1016/j.pain.2005.10.025.
Grimes, D. A., & Schulz, K. F. (2002). Bias and casual associations in observational re-
search. The Lancet, 359(9302), 248–252. http://dx.doi.org/10.1016/S0140-6736(02)
07451-2.
12
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
IBM Corp. (2015). IBM SPSS Statistics for Windows. Armonk, NY: IBM CorpVersion 23.0.
Ito, H., Koyama, A., & Higuchi, T. (2005). Polypharmacy and excessive dosing: psy-
chiatrists' perceptions of antipsychotic drug prescription. British Journal of Psychiatry,
187, 243–247. http://dx.doi.org/10.1192/bjp.187.3.243.
Jorm, A. F. (2000). Mental health literacy public knowledge and beliefs about mental
disorders. The British Journal of Psychiatry, 177(5), 396–401. http://dx.doi.org/10.
1192/bjp.177.5.396.
Kay, S. R., Flszbein, A., & Opfer, L. A. (1987). The positive and negative syndrome scale
(PANSS) for schizophrenia. Schizophrenia Bulletin, 13(2), 261–276. http://dx.doi.org/
10.1093/schbul/13.2.261.
Kikkert, M. J., Barbui, C., Koeter, M. W., David, A. S., Leese, M., Tansella, M., ... Schene,
A. H. (2008). Assessment of medication adherence in patients with schizophrenia:
The Achilles heel of adherence research. The Journal of Nervous and Mental Disease,
196(4), 274–281. http://dx.doi.org/10.1097/NMD.0b013e31816a4346.
Krousel-Wood, M., Islam, T., Webber, L. S., Re, R., Morisky, D. E., & Muntner, P. (2009).
New medication adherence scale versus pharmacy fill rates in hypertensive seniors.
The American Journal of Managed Care, 15(1), 59–66. Retrieved from http://www.
ncbi.nlm.nih.gov/pmc/articles/PMC2728593/pdf/nihms-68741.pdf.
Kwon, J. S., Choi, J. S., Bahk, W. M., Yoon, K. C., Hyung, K. C., Chul, S. Y., ... Geun, O. C.
(2006). Weight management program for treatment-emergent weight gain in olan-
zapine-treated patients with schizophrenia or schizoaffective disorder: A 12-week
randomized controlled clinical trial. The Journal of Clinical Psychiatry, 67(4),
547–553. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16669719.
Lacro, J. P., Dunn, L. B., Dolder, C. R., Leckband, S. G., & Jeste, D. V. (2002). Prevalence
of and risk factors for medication nonadherence in patients with schizophrenia: A
comprehensive review of recent literature. The Journal of Clinical Psychiatry, 63(10),
892–909. Retrieved from http://europepmc.org/abstract/med/12416599.
Leucht, S., Samara, M., Heres, S., Patel, M. X., Woods, S. W., & Davis, J. M. (2014). Dose
equivalents for second-generation antipsychotics: The minimum effective dose
method. Schizophrenia Bulletin, 40(2), 314–326. http://dx.doi.org/10.1093/schbul/
sbu001.
Liao, C. H., Chang, C. S., Wei, W. C., Chang, S. N., Liao, C. C., Lane, H. Y., & Sung, F. C.
(2011). Schizophrenia patients at higher risk of diabetes, hypertension and hyperli-
pidemia: A population-based study. Schizophrenia Research, 126(1), 110–116. https://
doi.org/10.1016/j.schres.2010.12.007.
Linden, M., Scheel, T., & Eich, F. X. (2006). Improvement of patient compliance after
switching from conventional neuroleptics to the atypical neuroleptic amisulpride.
Journal of Psychopharmacology, 20(6), 815–823. http://dx.doi.org/10.1177/
0269881106061154.
Link, B. G., & Phelan, J. C. (2006). Stigma and its public health implications. The Lancet,
367(9509), 528–529. http://dx.doi.org/10.1016/S0140-6736(06)68184-1.
Machin, D., Campbell, M., Fayers, P., & Pinol, A. (1997). Sample size tables for clinical
studies (2nd ed.). UK: Blackwell Science.
Malchow, B., Reich-Erkelenz, D., Oertel-Knöchel, V., Keller, K., Hasan, A., Schmitt, A., ...
Falkai, P. (2013). The effects of physical exercise in schizophrenia and affective
disorders. European Archives of Psychiatry and Clinical Neuroscience, 263(6), 451–467.
http://dx.doi.org/10.1007/s00406-013-0423-2.
Mann, C. J. (2003). Observational research methods. Research design II: Cohort, cross
sectional, and case-control studies. Emergency Medicine Journal, 20(1), 54–60. http://
dx.doi.org/10.1136/emj.20.1.54.
Marks, K. A., Fastenau, P. S., Lysaker, P. H., & Bond, G. R. (2000). Self-appraisal of Illness
Questionnaire (SAIQ): Relationship to researcher-rated insight and neuropsycholo-
gical function in schizophrenia. Schizophrenia Research, 45(3), 203–211. http://dx.
doi.org/10.1016/S0920-9964(99)00208-X.
Marzolini, S., Jensen, B., & Melville, P. (2009). Feasibility and effects of a group-based
resistance and aerobic exercise program for individuals with severe schizophrenia: A
multidisciplinary approach. Mental Health and Physical Activity, 2(1), 29–36.
Retrieved from https://doi.org/10.1016/j.mhpa.2008.11.001.
McCann, T. V., Boardman, G., Clark, E., & Lu, S. (2008). Risk profiles for non-adherence
to antipsychotic medications. Journal of Psychiatric and Mental Health Nursing, 15(8),
622–629. http://dx.doi.org/10.1111/j.1365-2850.2008.01277.x.
McCann, T. V., Clark, E., & Lu, S. (2009). Subjective side effects of antipsychotics and
medication adherence in people with schizophrenia. Journal of Advanced Nursing,
65(3), 534–543. http://dx.doi.org/10.1111/j.1365-2648.2008.04906.x.
McCann, T. V., Deans, C., Clark, E., & Lu, S. (2008). A comparative study of antipsychotic
medication taking in people with schizophrenia. International Journal of Mental Health
Nursing, 17(6), 428–438. http://dx.doi.org/10.1111/j.1447-0349.2008.00561.x.
McEvoy, J. P. (2006). Risks versus benefits of different types of long-acting injectable
antipsychotics. The Journal of Clinical Psychiatry, 67, 15–18. Retrieved from http://
europepmc.org/abstract/med/16822092.
McFarlane, W. R., Dixon, L., Lukens, E., & Lucksted, A. (2003). Family psychoeducation
and schizophrenia: A review of the literature. Journal of Marital and Family Therapy,
29(2), 223–245. http://dx.doi.org/10.1111/j.1752-0606.2003.tb01202.x.
McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on
health promotion programs. Health Education Quarterly, 15(4), 351–377. http://dx.
doi.org/10.1177/109019818801500401.
Mehrens, W. A., & Lehman, I. (1991). Measurement and evaluation in education and psy-
chology. Chicago, IL: Holt, Rinehart, and Winston.
Meltzer, H. Y. (2013). Update on typical and atypical antipsychotic drugs. Annual Review
of Medicine, 64, 393–406. http://dx.doi.org/10.1146/annurev-med-050911-161504.
Ministry of Health (2010a). Chronic diseases. Retrieved from https://www.moh.gov.sg/
content/moh_web/home/costs_and_financing/schemes_subsidies/medisave/Chronic_
Diseases.html.
Ministry of Health (2010b). National Health Survey 2010 Singapore. Singapore:
Epidemiology & Disease Control Division. Retrieved from https://www.moh.gov.sg/
content/dam/moh_web/Publications/Reports/2011/NHS2010%20-%20low%20res.
X.C. Tham et al.
pdf.
Ministry of Health (2011). Schizophrenia. MOH clinical practice guidelines 4/
2011Singapore: Ministry of Health. Retrieved from https://www.moh.gov.sg/
content/dam/moh_web/HPP/Doctors/cpg_medical/current/2011/Schizophrenia
%20Pg1-56.pdf.
Ministry of Health (2014). Singapore burden of disease study 2010. Singapore:
Epidemiology & Disease Control Division. Retrieved from https://www.moh.gov.sg/
content/dam/moh_web/Publications/Reports/2014/Singapore%20Burden%20of
%20Disease%20Study%202010%20Report_Final.pdf.
Ministry of Health Singapore (2009). Updates on the number of Singaporeans suffering
from mental illness. Retrieved from https://www.moh.gov.sg/content/moh_web/
home/pressRoom/Parliamentary_QA/2009/Updates_on_the_number_of_
Singaporeans_suffering_from_mental_illness.html.
Misdrahi, D., Verdoux, H., Lançon, C., & Bayle, F. (2009). The 4-point ordinal alliance
self-report: A self-report questionnaire for assessing therapeutic relationships in
routine mental health. Comprehensive Psychiatry, 50(2), 181–185. http://dx.doi.org/
10.3109/08039488.2011.598556.
Misdrahi, D., Petit, M., Blanc, O., Bayle, F., & Llorca, P. M. (2012). The influence of
therapeutic alliance and insight on medication adherence in schizophrenia. Nordic
Journal of Psychiatry, 66(1), 49–54. http://dx.doi.org/10.3109/08039488.2011.
598556.
Morisky, D. E., Ang, A., Krousel-Wood, M., & Ward, H. J. (2008). Predictive validity of a
medication adherence measure in an outpatient setting. The Journal of Clinical
Hypertension, 10(5), 348–354. http://dx.doi.org/10.1111/j.1751-7176.2008.
07572.x.
Morisky, D. E., & DiMatteo, M. R. (2011). Improving the measurement of self-reported
medication nonadherence: Final response. Journal of Clinical Epidemiology, 64(3),
258–263. http://dx.doi.org/10.1016/j.jclinepi.2010.09.002.
Morisky, D. E., Green, L. W., & Levine, D. M. (1986). Concurrent and predictive validity of
a self-reported measure of medication adherence. Medical Care, 24(1), 67–74.
Retrieved from http://bcrsrc.jhmi.edu/courses/c34066001/Morisky_Validity%20of
%20self-reported%20measure%20article_1986.pdf.
Moritz, S., Favrod, J., Andreou, C., Morrison, A. P., Bohn, F., Veckenstedt, R., ... Karow, A.
(2013). Beyond the usual suspects: Positive attitudes towards positive symptoms is
associated with medication noncompliance in psychosis. Schizophrenia Bulletin, 39(4),
917–922. http://dx.doi.org/10.1093/schbul/sbs005.
Mullins, C. D., Obeidat, N. A., Cuffel, B. J., Naradzay, J., & Loebel, A. D. (2008). Risk of
discontinuation of atypical antipsychotic agents in the treatment of schizophrenia.
Schizophrenia Research, 98(1), 8–15. http://dx.doi.org/10.1016/j.schres.2007.04.
035.
National Health Services (2012). Guidelines for Antipsychotics. Retrieved from http://
www.humber.nhs.uk/Downloads/Services/Pharmacy/Guidelines/Antipsychotic
%20medication%20switches%20guidelines.pdf.
Osborne, J., & Waters, E. (2002). Four assumptions of multiple regression that researchers
should always test. Practical assessment, research & evaluation. 8(2). Practical assess-
ment, research evaluation (pp. 1–9). . Retrieved from http://www-psychology.
concordia.ca/fac/kline/495/osborne.pdf.
Priebe, S., & Gruyters, T. (1993). The role of the helping alliance in psychiatric com-
munity care: A prospective study. The Journal of Nervous and Mental Disease, 181(9),
552–557. Retrieved from http://www.researchgate.net/profile/Stefan_Priebe/
publication/14953214_The_role_of_the_helping_alliance_in_psychiatric_community_
care._A_prospective_study/links/0912f5107b1824aa16000000.pdf.
Richards, M., Hori, H., Sartorius, N., & Kunugi, H. (2014). Cross-cultural comparisons of
attitudes toward schizophrenia amongst the general population and physicians: A
series of web-based surveys in Japan and the United States. Psychiatry Research,
215(2), 300–307. http://dx.doi.org/10.1016/j.psychres.2013.12.012.
Richman, J. A., Rospenda, K. M., & Cloninger, L. (2009). Terrorism, distress, and
drinking: Vulnerability and protective factors. The Journal of Nervous and Mental
Disease, 3(12), 909–917. http://dx.doi.org/10.1097/NMD.0b013e3181c29a39.
Rittmannsberger, H., Pachinger, T., Keppelmüller, P., & Wancata, J. (2004). Medication
adherence among psychotic patients before admission to inpatient treatment.
13
Archives of Psychiatric Nursing xxx (xxxx) xxx–xxx
Psychiatric Services, 55(2), 174–179. http://dx.doi.org/10.1176/appi.ps.55.2.174.
Rüsch, N., & Corrigan, P. W. (2002). Motivational interviewing to improve insight and
treatment adherence in schizophrenia. Psychiatric Rehabilitation Journal, 26(1), 23.
http://dx.doi.org/10.2975/26.2002.23.32.
Rüsch, N., Angermeyer, M. C., & Corrigan, P. W. (2005). Mental illness stigma: Concepts,
consequences, and initiatives to reduce stigma. European Psychiatry, 20(8), 529–539.
http://dx.doi.org/10.1016/j.eurpsy.2005.04.004.
Scheewe, T. W., Backx, F. J. G., Takken, T., Jörg, F., Strater, A. V., Kroes, A. G., ... Cahn,
W. (2013). Exercise therapy improves mental and physical health in schizophrenia: A
randomised controlled trial. Acta Psychiatrica Scandinavica, 127(6), 464–473.
Retrieved from http://www.timtakken.nl/pdf/131.pdf.
Shattell, M. (2004). Nurse-patient interaction: A review of the literature. Journal of
Clinical Nursing, 13(6), 714–722. http://dx.doi.org/10.1111/j.1365-2702.2004.
00965.x.
Shattell, M. M., Starr, S. S., & Thomas, S. P. (2007). ‘Take my hand, help me out’: Mental
health service recipients' experience of the therapeutic relationship. International
Journal of Mental Health Nursing, 16(4), 274–284. http://dx.doi.org/10.1111/j.1447-
0349.2007.00477.x.
Song, J. W., & Chung, K. C. (2010). Observational studies: Cohort and case-control stu-
dies. Plastic and Reconstructive Surgery, 126(6), 2234–2242. http://dx.doi.org/10.
1097/PRS.0b013e3181f44abc.
Steinskog, D. J., Tjøstheim, D. B., & Kvamstø, N. G. (2007). A cautionary note on the use
of the Kolmogorov-Smirnov test for normality. Monthly Weather Review, 135(3),
1151–1157. http://dx.doi.org/10.1175/MWR3326.1.
Suresh, K. P., & Chandrashekara, S. (2012). Sample Size estimation and power analysis for
clinical research studies. Journal of Human Reproductive Sciences, 5(1), 7–13. http://
dx.doi.org/10.4103/0974-1208.97779.
Sweileh, W. M., Ihbesheh, M. S., Jarar, I. S., Sawalha, A. F., Taha, A. S. A., Zyoud, H., &
Morisky, D. E. (2012). Antipsychotic medication adherence and satisfaction among
Palestinian people with schizophrenia. Current Clinical Pharmacology, 7(1), 49–55.
http://dx.doi.org/10.2174/157488412799218761.
Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach's alpha. International
Journal of Medical Education, 2, 53–55. Retrieved from http://www.ncbi.nlm.nih.
gov/pmc/articles/PMC4205511/.
Tham, X. C., Xie, H., Chng, C. M. L., Seah, X. Y., Lopez, V., & Klainin-Yobas, P. (2016).
Factors affecting medication adherence among adults with schizophrenia: A litera-
ture review. Archives of Psychiatric Nursing, 30(6), 797–809. Retrieved from https://
doi.org/10.1016/j.apnu.2016.07.007.
Turkington, D., Kingdon, D., Rathod, S., Hammond, K., Pelton, J., & Mehta, R. (2006).
Outcomes of an effectiveness trial of cognitive-behavioural intervention by mental
health nurses in schizophrenia. The British Journal of Psychiatry, 189(1), 36–40.
http://dx.doi.org/10.1192/bjp.bp.105.010884.
Velligan, D. I., Weiden, P. J., Sajatovic, M., Scott, J., Carpenter, D., Ross, R., & Docherty,
J. P. (2009). The expert consensus guideline series: Adherence problems in patients
with serious and persistent mental illness. The Journal of Clinical Psychiatry, 70(Suppl.
4), 1–46. Retrieved from http://ddi.wayne.edu/ddmi_articles/adherence_psych_
illness_j_clin_psy_2009.pdf.
Woods, S. W. (2003). Chlorpromazine equivalent doses for the newer atypical anti-
psychotics. Journal of Clinical Psychiatry, 64(6), 663–667. http://dx.doi.org/10.4088/
JCP.v64n0607.
World Health Organisation (2003). Adherence to long-term therapies. Evidence for
actionSwitzerland: World Health Organisation. Retrieved from http://www.who.int/
chp/knowledge/publications/adherence_full_report.pdf.
Yang, J., Ko, Y. H., Paik, J. W., Lee, M. S., Han, C., Joe, S. H., ... Kim, S. H. (2012).
Symptom severity and attitudes toward medication: Impacts on adherence in out-
patients with schizophrenia. Schizophrenia Research, 134(2), 226–231. http://dx.doi.
org/10.1016/j.schres.2011.11.008.
Zygmunt, A., Olfson, M., Boyer, C. A., & Mechanic, D. (2002). Interventions to improve
medication adherence in schizophrenia. The American Journal of Psychiatry, 159(10),
1653–1664. http://dx.doi.org/10.1176/appi.ajp.159.10.1653.