Академический Документы
Профессиональный Документы
Культура Документы
www.emeraldinsight.com/0952-6862.htm
IJHCQA
26,5 Measuring patients’ satisfaction
with pharmaceutical services at a
public hospital in Qatar
398
Imran Fahmi Khudair
Pharmacy Department, Hamad General Hospital, Doha, Qatar, and
Received 23 March 2011
Revised 19 July 2011 Syed Asif Raza
Accepted 15 August 2011
College of Business and Economics, Qatar University, Doha, Qatar
Abstract
Purpose – The aim of this paper is to study pharmacy service impact on patient satisfaction and to
determine what factors saliently link with pharmaceutical service performance at Hamad General
Hospital.
Design/methodology/approach – A patient satisfaction questionnaire was designed using the
literature and consultation with Hamad General Hospital medical experts. The questionnaire
contained 22 items that focused on five influencing factors: promptness; attitude; supply; location;
medication education; and respondent demographic aspects. A total of 220 respondents completed the
questionnaire. An exploratory factor analysis was used to group items and a structural equation model
was developed to test causality between five factors along with their influence on patient satisfaction.
Findings – The study establishes statistical evidence that patient satisfaction is positively
influenced by service promptness, pharmacist attitude, medication counseling, pharmacy location and
waiting area. Several socio-demographic characteristics have statistically different effect on
satisfaction, notably: gender; marital status; health status; age; educational level; and ethnicity.
However, medication supply did not influence patient satisfaction.
Practical implications – Pharmaceutical services are recognized as an essential healthcare-system
component. Their impact on customer satisfaction has been investigated in many countries; however,
there is no such study in Qatar.
Originality/value – The findings identify pharmaceutical service performance indicators and
provide guidelines to improve Qatari pharmaceutical services.
Keywords Patient satisfaction, Psychometric properties, Pharmaceutical services, Healthcare,
Structural equation modelling, Qatar, Customer satisfaction, Health services
Paper type Research paper
Introduction
Pharmacy services have been increasingly expanded beyond simple medication supply
to become a more patient-centered and caring service. Pharmacists work in harmony
with other healthcare providers to optimize patients’ quality of life and to achieve the
best clinical outcome. Good professional relationship and communication must be
established and maintained between the pharmacist and the patient to attain this goal.
International Journal of Health Care The pharmacist also should maintain an appropriate caring attitude and use his/her
Quality Assurance pharmacotherapy knowledge and skill as the medication expert to improve the
Vol. 26 No. 5, 2013
pp. 398-419 patients’ health and wellbeing.
q Emerald Group Publishing Limited Patient satisfaction is an integral healthcare-quality component. Improved
0952-6862
DOI 10.1108/IJHCQA-03-2011-0025 communication, convenience and courtesy can lead to better health service use and
ultimately better outcomes. High satisfaction promotes positive health behaviors, such Measuring
as compliance and continuity with providers. Patients who are satisfied with their patients’
overall care are likely to take medications properly and less likely to change from one
healthcare professional to another (Briesacher and Corey, 1997). According to satisfaction
Schommer and Kucukarslan (1997), patient satisfaction is about personally evaluating
or appraising a service or product perceived to be valuable and beneficial. Patient
satisfaction is becoming a popular healthcare quality indicator in which 399
pharmaceutical services are an essential part. Measuring patient satisfaction is an
approach to identifying and meeting patient needs (Lang and Fullerton, 1992).
Providing better access to quality pharmacies is a way to improve patient satisfaction
with healthcare. In a competitive healthcare market, pharmacists should provide
competent services in a satisfactory manner to ensure service continuity (Gourley et al.,
2001). As healthcare moves toward an outcomes-based (e.g., patient satisfaction) model
outcomes are important components to patients who may wish to see healthcare staff
who produce better outcomes or greater patient satisfaction. As a result, pharmacists
demonstrating greater patient satisfaction may be at a competitive advantage (Larson
et al., 2002). Patient satisfaction questionnaires have grown substantially over the last
decades as tools to measure healthcare from the patients’ view point (Quintana et al.,
2006). However, patient satisfaction had little attention in public or government
sponsored settings and in developing countries (Abd al Kareem et al., 1996).
Hamad Medical Corporation’s (HMC) Hamad General Hospital (HGH) is a
community teaching hospital with 600 beds and Doha’s main non-profit healthcare
provider in Qatar State (HMC web site, www.hmc.org.qa). Qatar, located on the
Arabian Gulf’s western coast, is on a 160 km, 11,493 km2 peninsula. In April 2010, the
Qatar Statistics Authority estimated Qatar’s population at 1,670,389 (76.1 percent
males) including around 70 percent expatriates (www.qsa.gov.qa). The country’s
healthcare has advanced greatly, which resulted in remarkable growth in peoples”
health and well-being. According to WHO, the per capita healthcare expenditure in
2005 was the highest among the Eastern Mediterranean regions (www.who.int/topics/
statistics/en/). In 2009, healthcare expenditure, excluding the private sector, was 3.1
percent of gross domestic product (www.qsa.gov.qa).
Literature review
In pharmaceutical context, many patient-satisfaction definitions are ambiguous.
Schommer and Kucukarslan’s (1997) early work contributed greatly to conceptualizing
patient satisfaction within a pharmaceutical services framework. They classified
pharmaceutical services based on four patient-satisfaction conceptualizations:
performance evaluation (a service’s salient characteristics); disconfirmation of
expectations (gap between expectation and actual experience); an individual’s
emotional response to a service and resulting actions (affect-based assessment); an
individual’s evaluation of what is gained compared to what it cost (equity-based
assessment). According to Schommer (2003), satisfaction can be viewed as an extent to
which an individual is able to execute a desired behavior based on a service.
Furthermore, research defines satisfaction as “an individual’s judgment about the
extent to which a product or service provides a pleasurable level of
consumption-related fulfillment” (Schommer, 2003, p. 317). It is an emotional
reaction resulting from a product or service evaluation, which is followed by a
IJHCQA judgment that is made by an individual on how perfectly the service was provided,
26,5 which results in either pleasure or displeasure (Schommer, 2003). Similarly,
satisfaction is assumed to include a cognitive evaluation of an emotional response to
the system’s structure, process and outcome (Gourley et al., 2001). In Larson et al. (2002,
p. 44), patient satisfaction is viewed as his/her “personal evaluation of healthcare
services and providers”. Satisfaction determinants can be considered as the patient’s
400 preferences and expectations, whereas technical and interpersonal care elements can
be its components (Larson et al., 2002).
No single patient-satisfaction measure is applicable to all pharmacy situations
(Schommer and Kucukarslan, 1997). A theoretical base is required for satisfaction
evaluation in which the measure’s validity can be assessed. The measure must fit an
overall research process framework and the researcher must have an apparent scheme of
what is to be measured (Schommer and Kucukarslan, 1997). Patient satisfaction is the
function underlying satisfaction constructs, which include satisfaction with the primary
provider’s staff and waiting time (Aragon and Edwards, 2004). By understanding these
components, pharmacy managers can work on improving those areas, produce more
satisfied patients and develop the pharmacy’s viability (Lang and Fullerton, 1992).
Patient satisfaction is a valuable humanistic outcome that should be measured. It is
important for determining healthcare service sustainability (Panvelkar et al., 2009). It
also reflects how pharmacy services influence patients’ lives. However, owing to its
subjective nature, it is difficult to assess and it is not a structure or process measure
(Gourley et al., 2001). Additionally, satisfaction is more subjective than reports of care,
which provide objective evidence about what occur in an encounter (Larson et al., 2002).
Self-reporting questionnaires include: mail; face-to-face interviewing; phone interviewing
or most recently by internet response. Although the internet has become an efficient way
to distribute and complete questionnaires, it is not feasible in many countries. Further, it
may be used only by people who are familiar with the internet (Quintana et al., 2006).
There are some benefits from having a satisfied patient. Aharony and Strasser
(1993), for example, found that satisfied patients are more likely to continue using
healthcare services, maintain a relationship with a specific healthcare provider, comply
with treatments (including medications), participate in their own treatment and
cooperate with their healthcare providers. Increasingly, patient satisfaction is
becoming a popular health-service quality indicator (Schommer and Kucukarslan,
1997). There has been a considerable increase in the attention paid to measuring
patient satisfaction. Searching “patient satisfaction” in the MEDLINE database
resulted in 51,395 publications at December 2010; rising from 152 in 1980. Measuring
patient satisfaction in the Arabic countries was further explored by searching
MEDLINE, EBSCO, ProQuest and Google Scholar databases using “patient
satisfaction” and each Middle East Arabic country name as key words. We found 67
articles: 30 Saudi Arabian; 11 Kuwaiti; eight Egyptian; eight Jordanian; four Qatari;
two United Arab Emirates and Lebanese; one from Oman and Yemen. Some included
pharmacy service factors that affect patient satisfaction. However, no study focused on
measuring patient satisfaction with pharmaceutical services. There were four studies
measuring patient satisfaction with the healthcare services conducted in Qatar. The
Abd al Kareem et al. (1996) work used an established self-administered questionnaire
developed by Ware et al. (1983) to measure patient satisfaction with medical care using
seven constructs in two governmental out-patient healthcare facilities. Constructs
included: general satisfaction; service availability and convenience; facilities; Measuring
humaneness; service-quality and continuity. Similarly, McGivern (1999) evaluated patients’
patient satisfaction with healthcare quality in 18 patient areas. The remaining two
studies assessed patient satisfaction in primary healthcare facilities. Almujali et al. satisfaction
(2009) focused on physician consultation effect on patient satisfaction in a public health
center. Al-Emadi et al. (2009)explored all 21 primary healthcare centers in Qatar using
a questionnaire developed by Makhdoom et al. (1997) in Saudi Arabia, which covered 401
six primary healthcare service aspects. Extending the search to other Gulf Cooperation
Council (GCC) countries, we found 45 articles that discussed patient satisfaction. Most
(40 percent) dealt with patient satisfaction in primary healthcare services; 11 percent
with specialized medical services, 11 percent surgical procedures, 7 percent dental
services, 18 percent hospital services, 4 percent hospital process improvements and 9
percent miscellaneous services.
Recently, there has been an increasing trend to calibrate patient satisfaction
questionnaires in Middle Eastern regions. Many researchers used up-to-date healthcare
research methods to measure service quality perceptions. Mostafa (2005) worked on
patient satisfaction and expectation in eight Egyptian public and private hospitals
involving 332 patients. The author concludes that quantitative methods are valuable
for establishing relationships between variables, but are considered weak when
attempting to identify the reasons for those relationships. Additionally, patients may
have multifaceted beliefs that cannot be acquired in the questionnaire and combining
qualitative and quantitative methods may enhance the findings. Al-Eisa et al. (2005)
measured patient satisfaction with primary healthcare services in Kuwait using an
Arabic language questionnaire that included socio-demographic characteristics,
overall and specific satisfaction questions with different services aspects in the facility,
such as: buildings; reception; physician; nurse; pharmacy; laboratory and X-ray
services. Convenience samples were used and 1,250 patients were included. Pharmacy
generated the highest satisfaction. The Alaloola and Albedaiwi (2008) Riyadh study
highlighted healthcare quality’s importance as a patient satisfaction predictor by
capturing other effects such as patient status. It involved 1983 inpatient, outpatient and
emergency care patients using a locally developed patient satisfaction questionnaire.
The authors recommended that service standards in the areas in which patients were
significantly dissatisfied should be raised by involving senior leaders. Areas for which
patients were significantly satisfied, on the other hand, will need to be sustained or
even improved. Badri et al. (2009) used structural equation modeling to test causality
relationships between healthcare quality and patient satisfaction Their questionnaire
was completed by 244 adult patients discharged from the United Arab Emirates public
hospitals. The proposed model included five main constructs. Three represented
service quality (care, administration and information). The fourth was patient status
and the fifth patient satisfaction with care. The model captured attributes that
characterize healthcare quality in developing countries, which can be used to evaluate
healthcare practices from patient viewpoints. Anbori et al. (2010) performed a similar
study in Sana’a, Yemen. It included 819 respondents and measured their satisfaction
and loyalty to private hospitals. The authors concluded that improvements were
required to achieve high quality healthcare services in the private hospitals and to
increase loyalty among patients.
IJHCQA Motivation and problem statement
26,5 As patient satisfaction is a health service quality indicator, assessing patient
satisfaction with pharmacy services is essential for improving Qatari public hospital
facilities and performance. No study focused on measuring patient satisfaction with
pharmaceutical services in Qatar or other GCC countries. Additionally, assessment
tools for measuring patient satisfaction with pharmacy services have not been
402 developed so far in this region. Regarding these shortcomings and a clear gap in the
literature, our purpose is to develop a reliable patient satisfaction instrument for
Qatar’s pharmaceutical services. We expect this study to help increase pharmacy
service viability and to help manages achieve internationally accredited public
healthcare service status and continuous quality improvement in Qatar’s public
healthcare system. The Joint Commission International (JCI) accreditation process
began in 2005 at HMC. This study is vital, therefore, for continuously improving HMC
facilities and for maintaining its accreditation standing. Among the major challenges
in our study is developing a patient satisfaction questionnaire, which should be
practical, not too lengthy or complex to suit the limited time patients have in the
out-patient pharmacy area. Lengthy questionnaires are likely to be partially completed
by respondents (Schommer and Kucukarslan, 1997) and a questionnaire with missing
information owing to difficult or improper question placement could severely impact
data collection (Aragon and Edwards, 2004). Thus, the questionnaire must include the
most relevant pharmacy service aspects that managers seek to understand. We focus,
therefore, on developing a survey-based on patient satisfaction methods; identifying
the dominant factors affecting patient satisfaction; developing conceptual models that
investigate and validate the relationship among these dominant factors and their
impact on patient satisfaction; determine patient satisfaction among different
socio-demographic groups and suggest ways to improve patient satisfaction with
pharmaceutical services in the Qatari public healthcare system.
Research methodology
Survey design and data collection
The survey design process was started by searching the literature and the pooled results
went through a review process. They were examined and filtered to measure and
improve our pharmacy practice. Seven constructs and sub items were obtained from the
literature: pharmacist communication skills (Lang and Fullerton, 1992; Panvelkar et al.,
2009); service location (MacKeigan and Larson, 1989; Gourley et al., 2001; Kamei et al.,
2001; Larson et al., 2002; Panvelkar et al., 2009); patient education (Larson et al., 2002;
Panvelkar et al., 2009); service promptness (Gourley et al., 2001; Larson et al., 2002;
Panvelkar et al., 2009); staff empathy and attitude (Kamei et al., 2001; Larson et al., 2002);
medication supply (MacKeigan and Larson, 1989); and staff competence (Larson et al.,
2002). An expert panel (senior pharmacists and HGH quality assurance department
experts) reviewed the questionnaire’s appropriateness, simplicity and clarity. The panel
suggested modifications and recommended using the pilot study outcomes. Most
pharmacy users are Arabic speaking so the English-language questions were translated
into Arabic. Any discrepancies between the two versions were resolved by the expert
panel using the approach suggested in Al-Emadi et al. (2009). The final questionnaire
was based on a pilot study involving 30 patients visiting HGH pharmacies to determine
the most salient constructs (see the Appendix, Figure A1). They were asked to select
three constructs likely to be the most important for patient satisfaction. Five constructs Measuring
from seven constructs were finally selected (Table I), so we designed the questionnaire patients’
by taking these five constructs into consideration.
The items under each construct were positively phrased using a five-point satisfaction
Likert-scale ranging from strongly disagree (1) to strongly agree (5). There were 22
items/questions: service promptness (two); pharmacist attitude and empathy (five);
medication supply (five); pharmacy location (four); and teaching (six). An open-ended 403
question asked the respondent to suggest ways to improve HGH pharmacy services.
The questionnaire’s second part was about the respondent’s socio-demographic
characteristics, including nationality, gender, marital status, age, educational level,
income, medication intake and average visits to HGH pharmacies in the last year. The
questionnaire took approximately five minutes to complete. From 195 patient
satisfaction studies that we located, 80 percent used a new satisfaction assessment
instrument and 10 percent modified existing questionnaires. Most studies are
context-specific and authors felt that existing instruments included irrelevant and
excluded relevant or an unsuitable item and construct mix (Sitzia, 1999). The survey
was conducted between October and November, 2010, during the HGH out-patient
pharmacy working week (Sunday-Thursday).
Sampling
We included respondents 18 years or older, able to read and write either Arabic or
English, who had obtained at least one prescription from HGH out-patient clinic
pharmacies in the previous year, and willing to complete the questionnaire. Individuals
unable to answer the questionnaire owing to language barriers were excluded.
Respondents were asked to evaluate their experiences because studies show that the
person completing the questionnaire have an important effect on the results; that is,
negative satisfaction levels were recorded in those surveys answered by someone other
than the patient (Quintana et al., 2006). All respondents were HGH pharmacy users
including Qatari citizens and multi-ethnic expatriates.
Ethics
The research questionnaire was approved by the hospital medical research center
(Research proposal no. 10154/10, www.hmc.org.qa/mrc) and informed consent was
obtained from each respondent.
Constructs Frequencya %
Factor analysis
Factor analysis (FA) is different from other statistical techniques since it is not a
confirmatory technique that tests hypotheses or determines whether one group is
Measuring
Characteristics Number %
patients’
Gender satisfaction
Female 130 59
Marital status
Married 157 71 405
Single 59 27
No response 4 2
Age group (years)
18-25 24 11
26-35 98 45
36-45 40 18
46-55 32 15
Above 56 24 11
No response 2 1
Education
Primary/secondary 2 1
High school 40 18
University 122 56
Higher education 50 23
No response 18 8
Language
Arabic 165 75
English 55 25
Ethnicity
Middle Easterners 132 60
South Asian 36 16
North African 28 13
Southeast Asian 8 4
American/Canadian 4 2
European 2 1
No response 10 5
Medication usage
Daily 121 55
When needed 93 42
No response 6 3
Average HGH Pharmacies visits last yeara 6 ^ 4.6
Average of service waiting time in the last visit (minutes)b 34 ^ 30
Average medication number in the last visit (items)b 3 ^ 1.7 Table II.
Demographics and
Notes: n ¼ 220; a1 US Dollar ¼ 3:65 Qatari Riyals; bMean^ ¼ SD characteristics
IJHCQA
Characteristics Number Mean SD p-value
26,5
Gender (n ¼ 220) , 0.001
Male 90 4.09 0.816
Female 130 3.49 1.086
Marital status (n ¼ 216) 0.023
406 Married 157 3.82 0.990
Single 59 3.46 1.088
Language (n ¼ 220) 0.763
Arabic 165 3.73 1.043
English 55 3.74 0.984
National (n ¼ 210) 0.086
Citizen 36 3.89 1.260
Noncitizen 174 3.71 0.984
Medication use (n ¼ 214) , 0.001
Table III. Daily 121 3.95 0.995
Satisfaction levels When needed 93 3.39 0.978
significantly different from another (Tabachnick and Fidell, 2007). It is a data reduction
procedure, which takes large variable-sets and looks to reduce or summarize them
using smaller factors or components. It does this by looking for clumps or groups
among the inter-correlations amid variables/items. This is an almost impossible task to
do by eye with anything more than a few variables. Factor analysis techniques have Measuring
several uses, which are used extensively by researchers who are developing and patients’
evaluating tests and scales. The scale (factor) developer starts with several individual
scales (items) and questions. By using FA techniques, they can refine and reduce these satisfaction
items to form smaller, coherent subscales, more commonly referred as factors. Factor
analysis can also be used to reduce several related variables to a more manageable
number prior to using them in other analyses such as multiple regression, multivariate 407
analysis of variance, structural equation modeling, etc. There are two main FA
approaches – exploratory and confirmatory. Exploratory FA is often used in the early
stages to gather information about exploring the interrelationships among variables.
Confirmatory FA, on the other hand, is a more complex and sophisticated technique
used later in the research process to test hypotheses concerning the structure
underlying variables. There are three main steps when conducting FA: assessing data
suitability for factor analysis; extracting factors; factor rotation and interpretation. All
22 items in our study were included in the FA. Two performance measures,
Kaiser-Meyer-Olkin (KMO) sampling adequacy and Bartlett’s sphericity test
(Tabachnick and Fidell (2007) were used to check data suitability. The KMO result
should range from 0 to 1 with 0.6 suggested as the minimum value for a good FA, and
Bartlett’s test should be significant. In our FA, the KMO value is 0.875 and also
Bartlett’s p value was , 0.001. Factor extraction, which involves determining the least
factors that can be used to best represent the interrelationships among the variables, is
done next. Although there are several approaches to factor extraction, the most
commonly used approach is principal component analysis (Tabachnick and Fidell,
2007). Kaiser’s criterion is used to determine the right factors. Using this rule, only
factors with an Eigenvalue of 1.0 or more are retained for investigation. Thus, five
components/factors were retained, which represent our construct number. The scree
plot in Figure 1 shows that Eigenvalues above 1 generate five factors. The variation
explained by five factors is about 73 percent of the total.
Once the factors have been determined, the next step is to interpret them by rotating
the factors. Rotation does not change the underlying solution; it presents loadings in a
manner that is easier to interpret. There are two approaches to factor rotation resulting in
either orthogonal (uncorrelated) or oblique (correlated) factor solutions. According to
Tabachnick and Fidell (2007), orthogonal rotation results in solutions that are easier to
interpret and to report; however, they require the researcher to assume that the
underlying constructs are not correlated, which may not be the case. We use the
orthogonal rotation approach. In the PASW (formerly SPSS) 18 software, the most
commonly used orthogonal approach (Varimax) is adapted. Table V presents the
Varimax rotation results. Additionally, the 22 items are listed in Table VI. All items
loaded well among factors representing the five constructs except item S5 and T6 (their
higher loading was in different factor). However, all factor loadings, ranging from
0.482-0.854 were significant (higher than 0.4, Larson et al., 2002), demonstrating the scale
items” convergent validity. All items in Table VI are distributed among the five factors.
Reliability (the scale’s internal consistency) was measured using Cronbach’s a
coefficient (Table V). Reliability coefficients above 0.9 are excellent (such as Attitude and
Teaching); those above 0.8 are very good (such as Promptness and Place) and values
above 0.7 are acceptable (such as Supply) (Heidegger et al., 2006). Cronbach’s a
coefficient ranged from 0.786-0.918 for the subscales and 0.941 for the total scale.
IJHCQA
26,5
408
Figure 1.
Scree plot
Sitzia (1999) reported 17 studies that measured patient satisfaction. The median a
coefficient value was 0.82 for the subscale and 0.93 for the total scale in another 15
studies. Content validity was assured by first involving pharmacy user views when we
selected the most relevant constructs that steer their satisfaction. Experts also selected
appropriate items that best measured each construct. Adding an open-ended question,
which asks for ways to improve the service, gave a dynamic aspect to the questionnaire
that helped to detect pharmacy user perceptions and create an on-going instrument that
tracks satisfaction with pharmacy services over time (Lang and Fullerton, 1992).
satisfaction
Table V.
409
IJHCQA
Construct Item Description
26,5
Promptness P1 Receive medications within a reasonable time
P2 Waiting time is acceptable considering the quantity of
prescription medication
Attitude A1 Pharmacist helped me to get the medications
410 A2 Pharmacist helped to solve any problem getting the medication
A3 Pharmacist answered my questions
A4 Pharmacist understood the medical case
A5 Pharmacists treat me with respect
Supply S1 Medication quantity was sufficient
S2 All my medications were available in the pharmacy
S3 Medication name was clear and easy to read
S4 Medication label/sticker instructions were clear
S5 Medication appearance and quality was good
Place C1 The pharmacy was easily found
C2 The waiting area was comfortable
C3 The pharmacy area was clean and acceptable
C4 The pharmacy working hours are suitable to me
Teaching T1 Pharmacist explained the reason for my medication
T2 Pharmacist told how to take the correct medication dose
T3 Pharmacist explained my medication’s possible side effects
T4 Pharmacist explained how to store my medication
T5 I had enough time with the pharmacist
T6 Pharmacy had a private place for teaching medication use
Table VI.
Construct descriptions Satisfaction SAT I am satisfied with HGH pharmaceutical services
that set it apart from older multivariate procedures. It takes a confirmatory rather than
an exploratory approach (Kutner et al., 2005). Moreover, by demanding that the
inter-variable relations are specified a priori, SEM lends itself well to data analysis for
inferential purposes. In contrast, most other multivariate procedures are essentially
descriptive (e.g., exploratory factor analysis), so that hypothesis testing is difficult, if
not impossible. Traditional multivariate procedures are incapable of either assessing
or correcting measurement error, while SEM explicitly estimates error variance
parameters. Alternative methods, such as linear regression (Kutner et al., 2005) assume
that error(s) is the exploratory (i.e., independent), and normally distributed with zero
mean. Applying methods, when there is error in the exploratory variables, is
tantamount to ignoring errors, which may lead to inaccuracies, especially when errors
are sizeable. Such mistakes are not encountered normally when SEM analyses are
used. Although exploratory data analyses methods based on observed measurements
only are reported, whereas those using SEM procedures can incorporate both
unobserved (latent) and observed variables. Lastly, there are no widely and easily
applied alternative methods for modeling multivariate relations, or for estimating point
and/or interval indirect effects. These important features are available using SEM.
Given, these highly desirable characteristics, SEM has become popular for
non-experimental research, when methods for testing theories are not well developed
and ethical considerations make experimental design infeasible (Bentler, 1989). Thus,
SEM can be used effectively to address numerous research problems involving Measuring
non-experimental research. patients’
In behavioral sciences, researchers are often interested in studying theoretical
constructs that cannot be observed directly. These abstract phenomena (latent variables or satisfaction
factors) are not observed directly; thus, a researcher must operationally define the latent
variable behaviorally, which s/he believe represents it. A latent or unobserved variable is
linked to one that is observable and thus measurable. Assessing behaviors constitutes 411
directly measuring an observed variable, albeit indirectly measuring an unobserved
variable. In SEM, these measurements are termed observed or manifest variables. They
indicate underlying constructs presumed to represent. Given this necessary bridging
process between observed variables and unobserved latent variables, it is important that
SEM requires researchers to be circumspect when selecting assessment measures. The
proposed SEM-based conceptual model is presented in Figure 2, which describes five
construct’s impact: service promptness; pharmacist attitude; medication supply; pharmacy
location; and teaching on patient satisfaction and their inter-relationships.
All five (exogenous) constructs were measured using observed (endogenous)
variables. Observed variables relating to each construct are measured using our
questionnaire data. Exogenous constructs and endogenous items used in the
questionnaire with their codes and description displayed in Table VI. In AMOS
(Analysis of MOment Structures) software, the variables send out arrowheads
denoting which variable it is predicting. Using AMOS, SEMs are schematically
Figure 2.
Proposed structural
equation model
IJHCQA portrayed using a four geometric symbols-a circle (or ellipse) configuration, a square
26,5 (or rectangle), a single-headed arrow and a double-headed arrow. By convention, circles
(or ellipses) represent unobserved latent factors. Squares (or rectangles) represent
observed variables, single-headed arrows represent one variable’s impact on another
and the double-headed arrows represent co-variances or correlations between variable
pairs. Our motive was to study the conceptual model presented in Figure 2.
412 Determining fitness is a basic task in SEM modeling. The data analyzed by SEM
programs include a matrix of estimated relationships between model variables.
Assessment of fit calculates how similar the predicted data are to matrices containing the
relationships in the model’s data. The AMOS model fitness was tested and the results are
summarized in Table VII. The x 2/df ratio should be less than five, which is used as an
appropriate measure of model fit. The NFI, NNFI, and CFI values should be on the higher
side of a 0-1 scale to indicate a model with good fit (Bentler, 1989). Additionally, for an
adequate model fitness, SRMR should be lower than 0.1 (Sivo et al., 2006) and the present
model fitness results show that SRMR ¼ 0:067 , 0:1, which meets another model
fitness criterion. By looking at these results, we might conclude that this model
reasonably satisfies the fitness criteria. As mentioned previously, the proposed model
presented in Figure 2 is implemented using AMOS, which tests the hypotheses listed in
Table VIII. The covariance relationships derived from the AMOS output are presented in
Table IX and the regression relationships are presented in Table X.
Hypothesis interpretations
Hypothesis test results were interpreted as follows:
H1-H4. Service promptness is positively correlated with pharmacist attitude;
medication supply, pharmacy location and teaching. Covariance
estimates for service promptness are statistically significant
(p , 0:001). Thus, H1-H4 are supported.
H5-H7. Pharmacist attitude is positively correlated with medication supply;
pharmacy location and medication teaching. The covariance estimates
for pharmacist attitude in Table IX are statistically significant
(p , 0:001). Therefore, H5-H7 are supported.
H8 and H9. Medication supply is positively correlated with pharmacy location and
teaching. Covariance estimates for medication supply are also
statistically significant for these two hypothesis (p , 0:001). H8 and
H9, therefore, are supported.
418 Appendix
Figure A1.
Patient satisfaction
questionnaire
Measuring
patients’
satisfaction
419
Figure A1.
Corresponding author
Syed Asif Raza can be contacted at: syedar@qu.edu.qa