Anda di halaman 1dari 14

Title: How satisfied are citizens with public hospital services?

Author 1: Jaime R. S. Fonseca


Department: Centre for Public Administration and Policies; Higher Institute of Social and
Political Sciences
Institution: Technical University of Lisbon
Town/City: Lisbon
Country: Portugal
Corresponding author: Jaime R. S. Fonseca
Corresponding Authors Email: jaimefonseca@i scsp.utl.pt
Biographical Details:
Author 1: Jaime Ral Seixas Fonseca is currently Professor of Statistics/Data Analysis and
Segmentation Techniques, Technical University of Lisbon, Institute of Social and Political
Sciences-ISCSP. He received a PhD in Quantitative Methods, Statistics and Data Analysis from
ISCTE Business School. He is a researcher in the Centre for Public Administration and Policies
(CAPP), research group coordinator, mixed methods for social and health sciences. His scientific
interests include mixed methods in social and health sciences, market segmentation, latent
segment models, product/service quality and customer satisfaction.
Structured Abstract:
Purpose: This study empirically tests emergency department patients service quality
perceptions by measuring their satisfaction.
Design/Methodology/Approach: Mixed method research involved collecting, analyzing and
interpreting qualitative and quantitative data in a single study, thereby providing a bridge
between qualitative and quantitative paradigms. The goal was to merge knowledge by using
qualitative conclusions to understand quantitative conclusions.
Findings: Segment 1 (62% of the patients) represents the unsatisfied group and a segment 2
(39% of the patients) represents the satisfied cohort. Dissatisfied patients classified almost all
items as very bad, bad or satisfactory; satisfied patients classified almost all items as very good
or good. This study identifies all items that contribute to low patient-satisfaction.
Practical implications: Our findings are important because administrators and policy-makers
know that patients constitute two different segments and have different characteristics, which can
help administrators and managers to develop new strategies and policies to improve
performance.
Originality/value: This mixed-method research provides valuable information for administrators
and managers. The proposed framework can be an excellent way to evaluate patient satisfaction.
Keywords: Mixed research methods, Public hospitals, Healthcare quality, patient satisfaction,
Latent segment model, Portugal.
Article Classification: Research

Introduction
Utilising health research for policy-making is important and the need to understand its
mechanisms is increasingly recognised. Recent reports calling for more resources to improve
health in developing countries and global pressures for accountability draw greater attention to
research-informed policy-making (Hanney et al., 2003). Research concerning processes in which
knowledge about policies, administrative arrangements, institutions and ideas in one political
setting is used to develop policies, administrative arrangements, institutions and ideas in another
political setting (Dolowitz and Marsh, 2000). There has been an increasingly important shift in
public policy analysis in recent years where approaches that emphasizes ideas, general precepts
and representations more than social evolution and state action are being developed (Surel,
2000). The major advance in recent times has been market researchers using economic and
behavioural theories and sophisticated analytical techniques to identify market segments.
Dickson and Ginter (1987) and Wells et al., (2010) provide a theoretical and practical
segmentation approach to behavioral psychology and wider marketing segmentation literature.
Market segmentations purpose is to identify consumption patterns by dividing
(segmenting) the market into several homogeneous units. Marketers can formulate service/
product strategies tailored specifically to homogeneous segment demands (Lin, 2002). Most
traditional healthcare-quality reports use univariate and bivariate statistics, such as mean,
standard deviation and correlation coefficients, but generally they do not influence policyprocesses; mostly because they do not show an overall picture. We can use multivariate based
models that enable us to measure/model complex phenomena to synthesize and extract
knowledge, giving us an overall picture by uncovering patterns that influence policy processes
and outcomes. Managers want to know how satisfied their customers are before translating
information into marketing strategies and organizational development. To guess diagnostic
classifications, health scientists look at models; in criminology, for example, they want to know
criminal profiles so data sets characterizing criminals help them reach adequate research
strategies.
National governance issues and corruption are particularly daunting in poorer countries,
but there are governance manifestations that afflict other countries and it is important to know
what citizens think about governance and parliamentary members, security, human rights, social
policies, crime and economic issues by finding hidden patterns. This study, therefore, aims to
empirically test hospital patient service-quality views and use them to develop new policies and
improve performance. We use latent segment models (LSM) to synthesize and extract
knowledge from patient data to: (i) assess public-hospital patient satisfaction; (ii) evaluate public
health policies using LSM to find out what causes dissatisfaction; and (iii) correct public-policy
defects.

Literature review
Few service management strategies have been explicitly labeled or thought as service
management strategies (Chung, 2001). Managers frequently want to know customer satisfaction
levels, translate findings into marketing strategies and organizational development. Knowing that
service quality and customer satisfaction are distinct constructs (Dabholkar, 2000), Spreng and
Macoy (1996) and Dabholkar (2000) found empirical support for considering customer
satisfaction as a service-quality satisfaction consequence. Service quality is more difficult for
consumers to evaluate than product quality (Ueltschy et al., 2007) owing to the intangible

evidence associated with services (Hong and Goo, 2004), particularly for professional services
because they are people-based, which increases service quality variability (Ueltschy et al., 2007).
Measuring customer expectations is difficult; i.e., can they have expectations about unknown
services? So, quality is about conformance to service design or service specification (Fonseca,
2009). Once the design is set, quality is about ensuring that services meet this specification or
design and as a consequence, from a service view point, customer satisfaction is about
monitoring service-quality delivery, thus measuring how well staff are delivering services
(Fonseca, 2009). If customers recognize service quality then they become satisfied; i.e., if
customers evaluate service quality then they are evaluating customer satisfaction. Fonseca's
(2009) simplified conceptual model can be used to measure customer satisfaction (Figure 1).

Figure 1: Conceptual model


Functional
Quality

Corporate
Image

Customer
satisfaction

Technical
Quality

Policy-making, broadly, includes national health policies made by: government ministers and
officials; local health-service managers and clinical guidelines from professional bodies (Hanney
et al., 2003). Public administrators and scholars must re-engage the public in governance,
recognize the special duty they have to citizens and move research and teaching agendas in a
direction that supports these new governance processes to address the fundamental democracyimperatives (Bingham et al., 2005).
Research evidence is one input into decision-making by the public and civil society,
patients, clinicians, healthcare managers and healthcare policy-makers (Lavis et al., 2005). The
movement to include patient evaluations are burgeoning as health services have discovered what
commercial services have long known that quality methods, skillfully applied, can reduce costs
and increase patient satisfaction and clinical outcomes (Ovretveit, 2000). Indeed patient
satisfaction is a cost effective, non-invasive quality indicator. Citizens play an important role in
public policy and decision making and they have the right to decide what is important to them
and how they can best achieve their objectives concerning their healthcare rights.
Healthcare customer orientation is gaining increasing attention (Oja et al., 2006). The
need for innovation is undeniable in most sectors, but seems to be even more so in the hospital
sector (Caccia-Bava et al., 2009). Patient opinions are an important service quality issue and an
endpoint in quality evaluation (Wilde-Larsson and Larsson 2009), and consequently in
3

satisfaction studies. Contemporary healthcare faces broad challenges, including how to define
and assess quality and quality of life, plan and evaluate services, define and implement safety
procedures and measures, and use evolving technologies effectively (Trochim and Kane, 2005).
The patient perspective is becoming more integrated in the improving health-care system process
(Rahmqvist and Bara, 2010). Traditional. mostly univariate indicators, which give us a partial
view, are not useful for influencing policy-makers and policy-processes, because they are
inadequate. Nowadays, considering service quality, policy-makers need an overall indicator
drawn from several factors that influence service quality more than an indicator for each factor.
As this research studys context is overall hospital performance, it is natural to focus on customer
satisfaction as an overall hospital service indicator (Anderson et al., 1997). Thus, our goal is to
create a hospital customer satisfaction typology and the results used for marketing strategy and
organizational development purposes.

Methods
As we move into the 21st century, better methods are needed to address the healthcare systems
complex issues. These methods should be systematic and flexible to accommodate complex
issues (Trochim and Kane, 2005). Concerning the optimal way to present research evidence for
service management and policy making, we argue that a simplified conceptual model is given by
Latent Class Model multivariate statistical methods. We use multi data-collection methods
(questionnaire and semi-structured interviews) and consequently multi method data analysis
(content analysis and Latent Class Analysis). Latent class models and content analysis (based on
survey, n = 471 and semi-structured interviews, nI = 17, respectively) are practicable and
important tools for multi-method processes that have rigor and scientific credibility. These
segmentation base variables (Table I) can be divided into three main areas: (i) technical quality;
(ii) functional quality; and (iii) corporate image. We use sex and age as covariates to achieve a
better typology.
We used a mixed-methods framework to evaluate public healthcare policies and we
analyze our main research question what are homogeneous units concerning satisfaction? by
performing in-depth interviews with 17 hospital patients who gave detailed perceptions. In our
research, the mixed methods used a questionnaire survey method, after an in-depth qualitative
enquiry (Olsen 2004). All questions were based upon an in-depth qualitative enquiry known as
the pilot survey (Blaikie, 2000). All questions were set up after examining the relevant literature.
We used a questionnaire to collect data to understand if patients hospital-quality views are
homogeneous. All patients visiting one Portuguese public hospital in March 2009 were invited to
participate. A questionnaire concerning perceived healthcare quality was administrated to
patients in the emergency department. We argue that to estimate global customer satisfaction, we
must use methods recognizing that customer satisfaction must be understood as a latent variable,
quantified through multiple indicators (items). To synthesize and extract knowledge from our
dataset, we propose measuring customer satisfaction indirectly as a latent variable by estimating
LSM, assuming that there is heterogeneity, which is natural because different staff provide the
same services (Fonseca, 2009). These methods accommodate multiple attributes (including
mixed case), provide parsimonious models that account for the relationships between multiple
attributes and derive latent segments about the customers overall satisfaction, based on these
attributes or indicators on technical and functional quality and corporate image (Fonseca, 2011).

Table I: Segmentation base variables


Secretaryship and Support Personnel 1
Secretaryship and Support Personnel 2
Secretaryship and Support Personnel 3
Technical Quality
Organization Functioning 1
Organization Functioning 2
Organization Functioning 3
Waiting Room Conditions 1
Waiting Room Conditions 2
Functional Quality
Waiting Room Conditions 3
Medical Room Conditions
Doctor Teams 1
Doctor Teams 2
Doctor Teams 3
Nursing 1
Corporate Image
Nursing 2
Nursing 3
Other Professionals 1
Other Professionals 2
Auxiliary Staff
Sex
Socio-demographic
Age

Latent segment modeling has become increasingly popular in the marketing literature (Wedel
and Kamakura, 1998; Dillon and Kumar, 1994; Bhatnagar and Ghose, 2004). This approach to
clustering has advantages compared to other clustering techniques: (i) it identifies segments and
provides unbiased estimates (Dillon and Kumar, 1994); (ii) it provides a means to selecting the
best segments (McLachlan and Peel, 2000); (iii) it is able to deal with different measurement
levels (Vermunt and Magidson, 2002); (iv) demographic and other covariates can be used for
cluster description (Magidson and Vermunt, 2003); and (v) it allocates cases into segments based
upon membership probabilities estimated directly from the model rather than using ad-hoc
distance definitions; e.g., Euclidian distance (Bonilla and Huntington, 2005). Our dataset is a
data matrix (n x p) that describes n patients by p attributes (questionnaire items or segmentation
base variables). Let y i ( yip ) denote the vector representing the scores of the ith patient for the
pth segmentation base variable (i = 1,,n ; p = 1,,P). We consider that cases on which the
attributes are measured arise from a population that we assume to mix S segments, in proportions
s (mixing proportions or relative segment sizes), s = 1,,S. The statistical probability density
function of vector y i , given that y i comes from segment s is represented by f s ( yi | s ) with s
representing the vector of unknown parameters associated with the specific chosen probability
density function. Population density can be represented as a finite mixture of densities f s ( yi | s )
of S distinct segments, i.e.:
f ( y | )
i

f (y
P

s 1

p 1

| s )

(1)

where i = 1,,n, s 0, s 1, { , }, with {1 , , s 1 } , { 1 , , s } . is the


s 1
vector of all unknown parameters.
S

The LSM estimation simultaneously addresses the distributional parameters and case
classification into segments, yielding mixing probabilities. The estimation process is typically
directed to maximum likelihood using the expectation-maximization (EM) algorithm
(McLachlan and Peel, 2000; Dempster et al., 1977). Latent segment modeling naturally provides
means for constituting a partition by assigning each case to the segment with the highest
segment-membership probability; i.e., with Max is , where:
s 1,...,S

is

(k)
s

f s ( y i |

(k)
s

S (k)
(k)
f j ( y i | )
j
j 1 j

(2)

Deriving meaningful results from clustering requires that the mixture model must be identifiable;
i.e., a unique maximum likelihood solution should exist (Bozdogan, 1994). Traditional LSM
estimation determines the smallest latent segment S total, sufficient to explain the relationships
observed among the segmentation base variables. If the baseline model (S = 1) provides a good
fit to the data then no LSM is needed since there is no relationship among the variables to be
explained otherwise a model with S = 2 segments is then fitted to the data. This process
continues by fitting successive LSM to the data, each time adding another dimension by
incrementing segments by 1 until a parsimonious model is found that provides an adequate fit.
We used the information criterion AIC3 (Akaiks Information Criterion Family) because the
segmentation base variable only contains categorical variables (Fonseca, 2010). Thus we select
the best model that presents the minimum value for AIC3 or an elbow, criterion value with slowly
decreasing.
Data analysis and results
The Latent Segment Model estimation, using information criterion AIC3, selects a two-segment
latent solution (Table II to V), because AIC3 values present an elbow for S = 2. In this solution,
we have cluster one with 62% and cluster two with 38% of public hospital patients. Table II
displays customer corporate image model estimates. The conditional probabilities displayed in
Tables II to IV allow us name two segments and profile customers. We have segment one with
62% of patients representing the unsatisfied group and a segment two with 38% representing
satisfied patients. Tables II - IV display two probability sets (model 1), which are estimated by
LSM to characterize customer typology; the probabilities s (s = 1,...,S) of belonging to segment
s, and probabilities f s ( y i | s ) , of belonging to a certain variable category, conditional on
belonging to a segment s: for instance, 0.043 and 0.001 are the probabilities of answering 1 =
very bad quality to secretaryship and support personnel 1 variable, given that s/he belongs to
segment one or two, respectively.

Table II: Corporate image model parameter estimates


Segment Size Segment one (62%) Segment two (38%)
Indicators
Secretaryship and Support Personnel 1
Very bad
0.0430
0.0010
Bad
0.0778
0.0074
Satisfactory
0.5147
0.2086
Good
0.3365
0.5792
Very good
0.0279
0.2038
Secretaryship and Support Personnel 2
Very bad
0.0336
0.0006
Bad
0.1111
0.0085
Satisfactory
0.4519
0.1617
Good
0.3777
0.6297
Very good
0.0257
0.1995
Secretaryship and Support Personnel 3
Very bad
0.0715
0.0020
Bad
0.0590
0.0065
Satisfactory
0.5395
0.2389
Good
0.2989
0.5308
Very good
0.0311
0.2218
Organization Functioning 1
Very bad
0.0466
0.0031
Bad
0.1208
0.0240
Satisfactory
0.5376
0.3196
Good
0.2593
0.4622
Very good
0.0358
0.1912
Organization Functioning 2
Very bad
0.2684
0.0595
Bad
0.2948
0.1493
Satisfactory
0.2956
0.3420
Good
0.1189
0.3143
Very good
0.0224
0.1350
Organization Functioning 3
Very bad
0.0286
0.0008
Bad
0.1318
0.0142
Satisfactory
0.4839
0.2037
Good
0.3139
0.5145
Very good
0.0418
0.2668

From Table II parameter estimates, we conclude that for public hospital patients, quality is weak.
After segmentation, we can illustrate the two segments graphically and we choose nursing, other
professional, secretaryship and support personnel, organization functioning and medical room
conditions. We now understand customers so we have a better understanding about customer
satisfaction. We now extract knowledge about customer typology and consequently about public

policy concerning healthcare, so that public administrators and policy makers can implement new
public policies. Table III displays technical quality using two-segment latent model parameters
estimates and we see that hospital technical quality also is weak. Patients are dissatisfied with
technical quality. From Figure 2, we can see that the two clusters are well separated, throughout
all indicators.
Table III: Technical quality model parameter estimates
Segment Size Segment one (62%) Segment two (38%)
Waiting Room Conditions 1
Very bad
0.0758
0.0029
Bad
0.2162
0.0343
Satisfactory
0.5198
0.3480
Good
0.1789
0.5054
Very good
0.0092
0.1095
Waiting Room Conditions 2
Very bad
0.1720
0.0275
Bad
0.2873
0.1224
Satisfactory
0.4338
0.4933
Good
0.1005
0.3049
Very good
0.0064
0.0518
Waiting Room Conditions 3
Very bad
0.0675
0.0007
Bad
0.1867
0.0117
Satisfactory
0.4796
0.1873

Table IV displays two segment latent model parameter estimates for functional quality. Patients
are dissatisfied with both functional and corporate image.
Figure 2: Cluster separation by clustering base variables (Cluster one +, Cluster two )
1,0

0,8

0,6

0,4

0,2

PSA1
0-1 Mean
PSA2
0-1 Mean
PSA3
0-1 Mean
OF1
0-1 Mean
OF2
0-1 Mean
OF3
0-1 Mean
CSE1
0-1 Mean
CSE2
0-1 Mean
CGM1
0-1 Mean
CGM2
0-1 Mean
EM1
0-1 Mean
EM2
0-1 Mean
EM3
0-1 Mean
ENFERMAG
0-1 Mean
V17_A
0-1 Mean
V18_A
0-1 Mean
OP1
0-1 Mean
OP2
0-1 Mean

0,0

Table IV: Functional quality model parameter estimates


Segment Size
Segment one (62%) Segment two (38%)
Doctor Teams 1
Very bad
0.0485
0
Bad
0.1638
0.0017
Satisfactory
0.6035
0.1132
Good
0.1791
0.5904
Very good
0.0051
0.2946
Doctor Teams 2
Very bad
0.0339
0.0001
Bad
0.1763
0.0051
Satisfactory
0.5815
0.1639
Good
0.1973
0.5381
Very good
0.0111
0.2927
Doctor Teams 3
Very bad
0.0677
0.0003
Bad
0.1756
0.0062
Satisfactory
0.5455
0.1672
Good
0.1978
0.5218
Very good
0.0134
0.3045
Nursing 1
Very bad
0.0339
0
Bad
0.1494
0.0015
Satisfactory
0.6418
0.1218
Good
0.1713
0.6263
Very good
0.0036
0.2504
Nursing 2
Very bad
0.0339
0
Bad
0.1640
0.0015
Satisfactory
0.6119
0.1075
Good
0.1866
0.6483
Very good
0.0035
0.2427
Nursing 3
Very bad
0.0485
0
Bad
0.1688
0.0015
Satisfactory
0.6159
0.1167
Good
0.1640
0.6458
Very good
0.0029
0.2360
Other Professionals 1
Very bad
0.0432
0.0008
Bad
0.1283
0.0120
Satisfactory
0.5985
0.2920
Good
0.2196
0.5568
Very good
0.0105
0.1384
Other Professionals 2
Very bad
0.0576
0.0010
Bad
0.1466
0.0138
Satisfactory
0.6120
0.3246
Good
0.1758
0.5260
Very good
0.0080
0.1347
Auxiliary Staff
Very bad
0.0530
0.0005
Bad
0.1909
0.0128
Satisfactory
0.5689
0.2617
Good
0.1798
0.5660
Very good
0.0074
0.1589

We know from latent class analysis that only 38% of patients consider Doctor Teams 1, Doctor
Teams 2 and Doctor Teams 3 quality is good. The same findings emerge regarding Nursing 1,
Nursing 2 and Nursing 3. For Other Professionals 1, Other Professionals 2 and Auxiliary Staff,
only 38% of patients considered quality good or very good. From Table V, using covariate
information, in segment one, we have mainly male and middle age (36 to 40 yrs) patients; in
segment two, we have mainly female and younger (29 or less) and older (47 to 64) patients. It
implies that unsatisfied patients are mainly male and middle age.
Summarized parameter estimates in Table II to VI allow us to display a profile concerning both
clustering variables and covariates. From the typology displayed in Table VI, we see that public
hospital customers are divided into two segments: one with 62% customers unsatisfied with all
three dimensions and a segment with 38% satisfied customers on all the three dimensions.
Table V: Covariate model parameter estimates
Segment Size Segment one (62 percent) Segment two (38 percent)
Covariates
Age
Less than 12
0.2123
0.1657
13 18
0.1693
0.1992
19 29
0.1824
0.2805
30 46
0.2500
0.1327
47 64
0.1863
0.2200
Sex
Male
0.6042
0.5933
Female
0.3958
0.4067

Table VII summarizes 17 interviews and we can see from a qualitative viewpoint, a structure
characterized by two patient classes: Unsatisfied and Satisfied.
Table VI: Customer typology
Variables
Secretaryship and Support Personnel 1
Secretaryship and Support Personnel 2
Secretaryship and Support Personnel 3
Organization Functioning 1
Organization Functioning 2
Organization Functioning 3
Waiting Room Conditions 1
Waiting Room Conditions 2
Waiting Room Conditions 3
Medical Room Conditions
Doctor Teams 1
Doctor Teams 2
Doctor Teams 3
Nursing 1
Nursing 2
Nursing 3
Other Professionals 1
Other Professionals 2
Auxiliary Staff
Covariates
Age
Genre

UNSATISFIED (62%)
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Very bad; Bad; Satisfactory
Between 30-46
Mainly Male

10

MUCH SATISFIED (38%)


Good; Very good
Good; Very good
Good; Very good
Good; Very good
Satisfactory; Good; Very good
Good; Very good
Good; Very good
Satisfactory; Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Good; Very good
Less than 30 and between 47 and 64
Mainly Female

These results allow us to update our quantitative findings using qualitative data; thus we gain
knowledge from mixed-research methods.
Table VII: Most frequent considerations concerning quality
Eleven interviews (eight male)
Unsatisfied Patients

Six interviews (four female)


Satisfied Patients

Interview 1: () Secretaryship and Support Personnel quality is bad


()
Interview 3 () Organization Functioning quality is very bad ()
Interview 4 () Waiting Room Conditions quality is very bad ()
Interview 7 () Doctor Teams quality is bad ()
Interview 8 () Nursing quality is very bad ()
Interview 10 () Other Professionals quality is bad ()
Interview 12 () Auxiliary Staff quality is very bad ()
Interview 13() Doctor Teams quality is bad ()
Interview 15 () Nursing quality is very bad ()
Interview 16 () Other Professionals quality is bad ()
Interview 17 () Auxiliary Staff quality is very bad ()

Interview 2 () Secretaryship and Support


Personnel quality is very good ()
Interview 5 () Organization Functioning
quality is good ()
Interview 6 () Waiting Room Conditions
quality is good ()
Interview 9 () Doctor Teams quality is very
good ()
Interview 11 () Nursing quality is very good
()
Interview 14() Other Professionals quality is
good ()

Conclusions
Managing healthcare takes place in complex environments involving different professionals,
extensive materials, equipment and services (Peltier et al., 2009). Moreover, global (because we
analyzed it from a multivariate perspective, with several dimensions) patient satisfaction is
influenced by several quality dimensions. Our study focuses on technical and functional quality,
and corporate image, suggested in the interviews and after controlling age and gender. By using
theses dimensions, administrators and managers have an overall hospital service-quality picture
by uncovering data patterns that influence policy makers and policy-processes. For instance,
corporate image and redressing service failures mean more than smiles; it means delivering core
hospital services so that employees can improve customer situations (Smith and Bolton, 2002).
To uncover patterns, we apply LSM, a classification scheme based on segmenting public hospital
customers into clusters that share certain relevant marketing characteristics. We evaluate health
service performance by studying customer satisfaction in one public hospital; uncovering the
satisfaction factors to help providers understand what patients need.
Previous patient satisfaction studies show that 80% or more express overall satisfaction
with their care, with few respondents answering negatively to any item (Williams and Calnan,
1991). We found a two-segment latent structure from the quantitative analysis: segment one, the
dissatisfied customers (62%), mainly male and middle-age patients and segment two, satisfied
customers (38%), mainly female and younger/older patients. Usually, older patients are more
satisfied while gender effects on global satisfaction are mostly inconsistent (Rahmqvist and Bara,
2010). We conclude that youngest and oldest female patients are more satisfied. The qualitative
analysis allowed us to complete the knowledge obtained from our quantitative analysis.
Our research findings can be used theoretically and practically, and we show that they
can influence public policy. Regarding patient attitudes, policy makers can develop new policies
to influence positively public attitudes, thus affecting quality of life positively. We intend to
study a larger sample, by extending the study to other public hospitals for a more consistent and
effective healthcare public policy evaluation.

11

References
Anderson, E.W., Fornell, C. and Rust, R.T. (1997), Customer Satisfaction, Productivity, and
Profitability: Differences between Goods and Services, Marketing Science, Vol. 16, No. 2,
129-145.
Bhatnagar, A. and Ghose, S. (2004), A latent class segmentation analysis of e-shoppers,
Journal of Business Research, Vol. 57, pp. 758-767.
Bingham, L.B., Nabatchi, T. and OLeary, R. (2005), The New Governance: Practices and
Processes for Stakeholder and Citizen Participation in the Work of Government, Public
Administration Review, Vol. 65 No. 5, 547-558.
Blaikie, P.M. (2000), Development, post- anti- and populist: a critical review, Environment and
Planning A, Vol. 32, pp. 10331050.
Bonilla, J. and Huntington, J. (2005), Leveraging Latent Class Segmentation to Optimize
Marketing Campaigns and Sales Force Implementation, PMSA Conference, Pacific
Midrange System Association, March 17, Chicago, IL.
Bozdogan, H. (1994), Mixture-Model Cluster Analysis using Model Selection Criteria and a
New Informational Measure of Complexity. Proceedings of the First US/Japan Conference
on the Frontiers of Statistical Modeling: An Approach, 69-113. H. Bozdogan, Kluwer
Academic Publishers Dordrecht, the Netherlands, pp. 69-113.
Caccia-Bava, M.C., Guimares, V. C. K., Guimares, T. (2009), "Testing some major determinants
for hospital innovation success", International Journal of Health Care Quality Assurance,
Vol. 22, No. 5, pp.454 - 470
Chung, B.G. (2001), A service market segmentation approach to strategic human resource
management, Journal of Quality Management, Vol. 6, pp. 117138.
Dabholkar, P.A. (2000), A Comprehensive Framework for Service Quality: An Investigation of
Critical Conceptual and Measurement Issues Trough a Longitudinal Study, Journal of
Retailing, Vol. 76 No 2, pp. 139-173.
Dempster, A.P., Laird, N.M. and Rubin, D.B. (1977), Maximum Likelihood from incomplete
Data via EM algorithm, Journal of the Royal Statistics Society, B (Methodological), Vol.
39, pp. 1-38.
Dickson, P. and Ginter, J.L. (1987), Market Segmentation, Product Differentiation, and
Marketing Strategy, Journal of Marketing, Vol. 51, pp. 1-10.
Dillon, W.R. and Kumar, A. (1994), Latent structure and other mixture models in marketing: An
integrative survey and overview, Chapter 9 in R.P. Bagozi (ed.), Advanced methods of
Marketing Research, pp. 352-388, Blackwell Publishers, Cambridge.
Dolowitz, D.P. and Marsh, D. (2000), Learning from Abroad: The Role of Policy Transfer in
Contemporary Policy-Making, Governance, Vol. 13 No.1, pp. 5-24.
Fonseca, J.R.S. (2009), Customer satisfaction study via a latent segment model, Journal of
Retailing and Consumer Services, Vol. 16, pp. 352-359.
Fonseca, J.R.S. (2011), Why Does Segmentation Matter? Identifying Market Segments
Through a Mixed Methodology, European Retail Research, Vol. 25 No 1, pp. 1-26.
Fonseca, J.R.S. (2010), On the Performance of Information Criteria in Latent Segment Models.
Proceedings of ICMSE, International Conference on Mathematical Science and
Engineering, World Academy of Science,
Engineering
and
Technology,
WASET, March 29-31, Rio de Janeiro, Brazil.

12

Hanney, S.R., Gonzalez-Block, M.A., Burton, M. J. and Kogan, M. (2003), The utilisation of
health research in policy-making: concepts, examples and methods of assessment, Health
Research Policy and Systems, Vol. 1 No 2, pp. 1-28.
Hong, S.C. and Goo, Y. J. (2004), A causal model of customer loyalty in professional service
firms: an empirical study, International Journal of Management, Vol. 21 No. 4, pp. 531540.
Lavis J.N., Davies H.T.O., Oxman A.D., Denis J.L., Golden-Biddle K. and Ferlie, E. (2005),
Towards Systematic Reviews That Inform Healthcare Management and PolicyMaking, Journal of Health Services Research and Policy, Vol. 10 Supplement 1, pp.35
48.
Lin, C.F. (2002), Segmenting customer brand preference: demographic or psychographic,
Journal of Product & Brand Management, Vol. 11, No 4, pp. 249-268.
Magidson, J. and Vermunt, J. K. (2003), A Nonthechnical Introduction to Latent Class Models,
Tilburg University, Nederlands, Statistical Innovations, Inc, p.15.
McLachlan, G.F. and Peel, D. (2000), Finite Mixture Models, John Wiley & Sons, New York.
Oja, P. I., Kouri, T. T. and Pakarinen, A. J. (2006), From customer satisfaction survey to
corrective actions in laboratory services in a university hospital, International Journal of
Quality Healthcare, Vol. 18 No 6, p. 422428 .
Olsen, J.P. (2004), Unity, diversity and democratic institutions: Lessons from the European
Union, The Journal of Political Philosophy, Vol. 12, No. 4, pp.461-495.
vretveit J. (2000), 'The economics of quality: a practical approach', International Journal of
Healthcare Quality Assurance, Vol. 13, No. 5, pp.2007.
Peltier, J., Dahl, A. and Mulhern, F. (2009), The Relationship Between Employee Satisfaction
and Hospital Patient Experiences, University of Wisconsin - Whitewater, Northwestern
University.
Rahmqvist, M. and Bara, A.C. (2010), Patient characteristics and quality dimensions related to
patient satisfaction, International Journal For Quality In Healthcare, Vol. 22 No 2, pp.
86-92.
Smith, A.K. and Bolton, R.N. (2002), The Effects of Customers Emotional Responses to
Service Failures on their Recovery Effort Evaluations and Satisfaction Judgments, Journal
of the Academy of Marketing Science, Vol. 30 No 1, pp. 5-23.
Spreng, R.A. and Macoy, R.D. (1996), An Empirical Examination of a Model of Perceived
Service Quality and Satisfaction, Journal of Retailing, Vol. 72 No 2, pp. 201-214.
Surel, Y. (2000), The role of cognitive and normative frames in policymaking, Journal of
European Public Policy, Vol. 7 No 4, pp. 495-512.
Trochim, W. and Kane, M. (2005), Concept mapping: an introduction to structured
conceptualization in healthcare, International Journal for Quality in Healthcare, Vol. 17
No 3, pp. 187191.
Ueltschy, L.C., Laroche, M., Eggert, A. and Bindl, A. (2007), Service quality and satisfaction:
an international comparison of professional services perceptions, Journal of Services
Marketing, Vol. 21 No 6, pp. 410423.
Vermunt, J. K. and Magidson, J. (2002), Latent class cluster analysis, in Hagenaars, J.A. and
McCutcheon, A.L. (eds.), Applied Latent Class Analysis, Cambridge University Press,
Cambridge, UK, pp. 89-106..
Wedel, M. and Kamakura, W.A. (1998), Market Segmentation: Concepts and methodological
foundations, Kluwer Academic Publishers, Boston.

13

Wells, V. K., Chang, S. W. ,Oliveira-Castro, J., Pallister, J., (2010), Market Segmentation from
a Behavioral Perspective, Journal of Organizational Behavior Management, Vol. 30 No 2,
pp. 176 - 198.
Wilde Larsson, B. and Larsson, G. (2009), Patients view on quality of care and attitudes towards
re-visiting the same provider, International Journal of Health Care Quality Assurance,
Vol. 22 No. 6, pp.600-611.
Williams, S.J. and Calnan, M. (1991), Convergence and divergence: assessing criteria of
consumer satisfaction across general practice, dental and hospital care setting, Social
Science & Medicine, Vol. 33 No 6, pp. 707-716.

14

Anda mungkin juga menyukai