Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Medicines

Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Ameh et al.

BMC Health Services Research (2017) 17:229


DOI 10.1186/s12913-017-2177-4

RESEARCH ARTICLE Open Access

Relationships between structure, process


and outcome to assess quality of
integrated chronic disease management in
a rural South African setting: applying a
structural equation model
Soter Ameh1,2* , Francesc Xavier Gómez-Olivé1,3, Kathleen Kahn1,3,4, Stephen M. Tollman1,3,4 and
Kerstin Klipstein-Grobusch5,6

Abstract
Background: South Africa faces a complex dual burden of chronic communicable and non-communicable diseases
(NCDs). In response, the Integrated Chronic Disease Management (ICDM) model was initiated in primary health care
(PHC) facilities in 2011 to leverage the HIV/ART programme to scale-up services for NCDs, achieve optimal patient
health outcomes and improve the quality of medical care. However, little is known about the quality of care in the
ICDM model. The objectives of this study were to: i) assess patients’ and operational managers’ satisfaction with the
dimensions of ICDM services; and ii) evaluate the quality of care in the ICDM model using Avedis Donabedian’s
theory of relationships between structure (resources), process (clinical activities) and outcome (desired result of
healthcare) constructs as a measure of quality of care.
Methods: A cross-sectional study was conducted in 2013 in seven PHC facilities in the Bushbuckridge municipality
of Mpumalanga Province, north-east South Africa - an area underpinned by a robust Health and Demographic
Surveillance System (HDSS). The patient satisfaction questionnaire (PSQ-18), with measures reflecting structure/
process/outcome (SPO) constructs, was adapted and administered to 435 chronic disease patients and the
operational managers of all seven PHC facilities. The adapted questionnaire contained 17 dimensions of care,
including eight dimensions identified as priority areas in the ICDM model - critical drugs, equipment, referral,
defaulter tracing, prepacking of medicines, clinic appointments, waiting time, and coherence. A structural equation
model was fit to operationalise Donabedian’s theory, using unidirectional, mediation, and reciprocal pathways.
Results: The mediation pathway showed that the relationships between structure, process and outcome
represented quality systems in the ICDM model. Structure correlated with process (0.40) and outcome (0.75). Given
structure, process correlated with outcome (0.88). Of the 17 dimensions of care in the ICDM model, three structure
(equipment, critical drugs, accessibility), three process (professionalism, friendliness and attendance to patients) and
three outcome (competence, confidence and coherence) dimensions reflected their intended constructs.
(Continued on next page)

* Correspondence: sote_ameh@yahoo.com
1
Medical Research Council/Wits University Rural Public Health and Health
Transitions Research Unit (Agincourt), School of Public Health, Faculty of
Health Sciences, University of the Witwatersrand, Johannesburg, South Africa
2
Department of Community Medicine, Faculty of Medicine, College of
Medical Sciences, University of Calabar, Calabar, Cross River State, Nigeria
Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Ameh et al. BMC Health Services Research (2017) 17:229 Page 2 of 15

(Continued from previous page)


Conclusion: Of the priority dimensions, referrals, defaulter tracing, prepacking of medicines, appointments, and
patient waiting time did not reflect their intended constructs. Donabedian’s theoretical framework can be used to
provide evidence of quality systems in the ICDM model.
Keywords: Integrated Chronic Disease Management (ICDM) Model, Avedis donabedian, Constructs, Quality of care,
Satisfaction, Chronic communicable diseases, Non-communicable chronic diseases, Structural equation model,
Primary Health Care (PHC), Mpumalanga province, South Africa

Background chronic diseases; human resource audit; capacity build-


South Africa faces a complex dual burden of chronic ing; supply of critical medicines; prepacking of medica-
communicable (HIV and TB) and chronic non- tion; and appropriate referral. To prepare the
communicable diseases (NCDs - e.g. cardiovascular dis- community for chronic disease care, each clinic has a
eases, diabetes, cancer and chronic respiratory diseases), PHC outreach team operating within the community
with the prevalence of HIV estimated at 10% in 2014 [1] that the clinic serves, and consists of one professional
and mortality due to NCDs estimated at 43% in 2012 nurse, three staff nurses, and six Community Health
[2]. Effectively responding to this dual burden of chronic Workers (CHWs). With the outreach team responsible
diseases requires an integrated approach to the delivery for 6000 individuals in 1500 households (250 households
of care at the primary health care (PHC) level. per 1 CHW), it is anticipated that at least 80% of defined
The Joint United Nations Programme on HIV/AIDS health problems of the catchment population would be
(UNAIDS) recommends a globally comprehensive and managed [9]. This study focuses on the facility compo-
integrated approach to the delivery of chronic disease nent of the ICDM model.
care. This approach requires leveraging HIV pro- Multiple meanings of “Integrated health care” exist
grammes to support or scale-up services for NCDs [3, in the literature. These include the provision of health
4]. There is evidence that the integrated management of care for multiple diseases at one service delivery point
chronic diseases leads to improvement in patient health (e.g. integrated management of childhood illness);
outcomes (e.g., CD4 count, glycosylated haemoglobin, continuity of care over time across different levels of
and blood pressure) and patient satisfaction with the de- health care (e.g. an appropriate referral system); inte-
livery of chronic disease care [5]. Beyond the UNAIDS grating vertical programmes (programmes that are
mandate for the implementation of an integrated separately funded and administratively managed in a
chronic care model, integrating services for HIV and ‘silo’) with the general health care system; multi-
NCDs could also minimise fragmented chronic disease sectoral collaboration; or a combination of two or
care arising from the management of the HIV more of these meanings [10]. The World Health
pogramme in a ‘silo’ within the general healthcare sys- Organization (WHO) defines “integrated health care”
tem, leverage resources and more efficiently meet pa- as “the organisation and management of health ser-
tients’ healthcare needs [6–8]. vices so that people get the care they need, when they
In response to UNAIDS recommendation to integrate need it, in ways that are user-friendly, achieve the de-
HIV and NCD services, the National Department of sired results and provide value for money.” [10]. In
Health (NDoH) in South Africa initiated the Integrated this study, the ICDM model refers to the ‘one-stop-
Chronic Disease Management (ICDM) model [9]. The shop’ for the management of chronic diseases in PHC
pilot of the ICDM model commenced in 2011 in se- facilities as well as continuity of care in the form of
lected PHC facilities in three of South Africa’s nine prov- referral of patients.
inces (Gauteng, Mpumalanga and North West), [9] with
the expectation of enhancing the quality of chronic dis-
ease services and improving patient health outcomes. Theoretical framework for evaluating quality of care in
At the crux of the ICDM operational framework are the ICDM model
facility reorganisation to improve operational efficiency “Quality of medical care” is highly contextual and a diffi-
and quality of care in the health facilities; “assisted” self- cult concept to define. Although it is a reflection of
management to promote individual responsibility in the values and goals in the medical care system and in the
communities; and health promotion and population larger society which it is a part of, quality can be almost
screening in the population [9]. The facility component anything anyone wishes it to be [11]. Klein et al. con-
entails many areas of focus such as: designation of clude that patient care, like morale, cannot be defined by
chronic care area; use of guidelines for management of a unitary concept and that it seems unlikely that there
Ameh et al. BMC Health Services Research (2017) 17:229 Page 3 of 15

will be a single criterion by which to measure the quality To the authors’ knowledge, this is the first study to apply
of patient care [12]. Donabedian’s theory in evaluating the quality of care in
Avedis Donabedian described seven elements of qual- the ICDM model in sub-Saharan Africa (SSA).
ity of medical care: Efficacy, Effectiveness, Efficiency, A systematic review to examine the effectiveness of in-
Equity, Optimality, Acceptability and Legitimacy. Al- tegrating primary health services in Low- and Middle-
though Efficacy is hard to measure, it refers to care pro- Income Countries (LMICs) showed the main focus to be
vided under optimal conditions and is the basis against on the provider side of service provision, with virtually
which measurements should be made. Effectiveness de- no considerations for lay or demand side perspective
scribes the outcome of interventions; Efficiency refers to [18]. For South Africa, little is known about satisfaction
cost reductions without compromising effects; Equity re- with the quality of care in the ICDM model. With sup-
fers to the fairness in the distribution of healthcare in porting evidence that satisfaction is a major component
populations; Optimality is about balancing the costs and and key determinant of quality of healthcare [15], this
benefits of healthcare; Acceptability encompasses acces- study examined satisfaction of both service providers
sibility of healthcare and interpersonal patient-provider and users with the quality of care in the ICDM model.
interaction; and Legitimacy refers to the social accept- The objectives of this study were to: i) assess patients’
ability of the healthcare institution regarding the manner and operational managers’ (nurses-in-charge of health
in which healthcare is delivered. The choice of which of facilities) satisfaction with the dimensions of ICDM ser-
these elements, as well as their relative prioritisation, vices; and ii) evaluate the quality of care in the ICDM
should be guided by the contexts in which quality of model, based on the satisfaction scores of patients, using
care is being assessed [13]. Donabedian’s SPO theoretical framework.
Donabedian’s definition of quality of care can be
assessed as a triad of structure, process and outcome Methods
(SPO) constructs. He postulated that there are relation- Study setting and sites
ships between SPO constructs based on the idea that This study was conducted in PHC facilities in the rural
good structure should promote good process and good Agincourt sub-district situated in the Bushbuckridge
process should in turn promote good outcome (unidir- municipality, Mpumalanga province, northeast South
ectional pathway). The SPO framework, often repre- Africa. At the time this study was conducted, the ICDM
sented by a chain of three boxes containing SPO model was being implemented in 17 of the 38 PHC facil-
constructs connected by arrows [13], can be used to ities in the sub-district. Seven of these 17 health facilities
draw inferences about the quality of health care [14]. implementing the ICDM model are situated in Agin-
Donabedian defines Structure as the professional and or- court sub-district which covers an area of about
ganisational resources associated with the provision of 420 km2. The sub-district underpinned by a robust
health care (e.g. availability of medicines/equipment and Health and Demographic Surveillance System (HDSS)
staff training); Process as the things done to and for the which has been monitoring the population in these vil-
patient (e.g. defaulter tracing and hospital referrals) and lages for two decades. The population under surveillance
Outcome as the desired result of care provided by the in the HDSS as at 1st July 2011 was 115,000 people in
health practitioner (e.g. patient satisfaction with quality 20,000 households in 27 villages [19]. Three referral hos-
of care). Donabedian distinguished between two types of pitals are situated 25 km to 45 km from the study set-
outcomes: i) technical outcomes, which are the physical ting. The pilot of the ICDM model was commenced in
and functional aspects of care, such as absence of com- these facilities in June 2011 (field diary of interviews
plications and reduction in disease, disability and death; with the operational managers and the sub-district
and ii) interpersonal outcomes which include patients’ health manager in July 2013), but preceded by two
satisfaction with care and influence of care on patient’s months of pre-implementation preparedness which
quality of life as perceived by the patient [15]. started in April 2011 [9]. Tsonga is the most widely
Avedis Donabedian’s SPO framework was used to spoken language in the study area. Having immigrated
evaluate the quality of care in the ICDM model not only into South Africa mainly as war refugees in the early-
because it is the dominant framework for evaluating the and mid-1980s, one-third of the population in the study
quality of medical care [16], but because the SPO frame- site are Mozambicans [19].
work is used by South Africa’s National Department of In the South African PHC model, the professional nurse
Health for implementing the ICDM model [9]. A study is the service provider at the PHC facilities, which is the
of quality systems conducted among department man- first point of entry into the public health system. Services
agers and quality coordinators in 386 hospitals in provided by the nurses include: maternal and child care,
Sweden showed statistically significant relationships be- immunization, family planning, treatment of sexually
tween SPO constructs, using Donabedian’s theory [17]. transmitted infections, minor trauma, care for chronic
Ameh et al. BMC Health Services Research (2017) 17:229 Page 4 of 15

diseases and referrals. Medical doctors visit the PHC facil- [21, 22], a sample size of 390 patient respondents was
ities at intervals to offer support to the nurses [20]. calculated (17 subjects per each of the 23 variables in
the study instrument). The minimum sample size of ap-
Study design and study population proximately 435 (390/0.9) patients was reached after
This was a cross-sectional survey conducted between adjusting for 10% non-response. All the seven oper-
August and November 2013. It was part of a broader ational managers of the PHC facilities, the maximum
four-year longitudinal study (January 2011 and Decem- number possible, were selected because they offered
ber 2014), with qualitative and quantitative components, clinical services to the patients and the authors per-
designed to contribute to understanding the effective- ceived their role as managers of the health facilities crit-
ness of the ICDM model in improving the quality of ically important to understanding the quality of the
healthcare and technical health outcomes of chronic dis- ICDM model more than other professional nurses.
ease patients. The study population consisted of patients The study participants were identified through a
18 years and above receiving treatment for chronic dis- three-step process (Additional file 1). First, the number
eases in the sub-district health facilities. Other study of patients recruited at each of the seven health facilities
participants included the operational managers (profes- was determined by proportionate sampling. The sam-
sional nurses-in-charge) of the selected seven PHC facil- pling fraction of 435/3602 (435 represents the desired
ities in the sub-district. sample size out of a total of 3602 HIV, hypertension, and
diabetes registered patients) was multiplied by the num-
Inclusion and exclusion criteria for the patients ber of these chronic disease patients in each health facil-
The ICDM model addresses the following chronic dis- ity to determine the number of patients to be recruited
eases: HIV/AIDS, tuberculosis, hypertension, diabetes, per facility. Secondly, the patients in each health facility
chronic obstructive pulmonary disease, asthma, epilepsy were stratified by HIV, hypertension, and diabetes status
and mental health illnesses that are to be managed at the in order to get a representative sample of the patients
PHC level [9]. Considering the burden of chronic diseases with markers of chronic diseases using a health facility-
in the study area, patients with markers of chronic dis- specific sampling frame. Finally, the numbers of patients
eases for HIV, hypertension and diabetes in the health fa- specified in step two were recruited for a daily interview
cilities were included in the study, while those with other until the desired sample size in each clinic was achieved.
chronic diseases were excluded. Patients who had their
chronic condition(s) managed five months before the initi- Study tool and variables
ation of the ICDM model until the time the study com- In this study, we used the multi-scale patient satisfaction
menced in August 2013 were identified for recruitment. questionnaire (PSQ-18) which was developed by Ware
The reason for including patients receiving treatment five et al. [23]. The PSQ-18 comprises 18 items derived from
months before the ICDM model was implemented was to the full-length version (50-item) PSQ-III counterpart
assess the levels of satisfaction of patients who had re- [23]. The PSQ-18 assesses multiple dimensions of pa-
ceived treatment before the implementation of the ICDM tient satisfaction and includes general satisfaction; tech-
model and continued to receive treatment during its im- nical quality; interpersonal relations; communication;
plementation in efforts to gauge possible changes in the financial aspect; time spent with health provider; and ac-
quality of chronic disease care attributable to the ICDM cessibility and convenience (Additional file 2). The PSQ-
model. Minors less than 18 years were excluded from the 18 sub-scales show acceptable reliability and correlate
study because they were below the age of autonomy with the sub-scales in the PSQ-III [24]. Furthermore,
(≥18 years) for judging satisfaction with the quality of ser- PSQ-18 is appropriate for use in situations where there
vices provided in the health facilities. The elderly with re- is need for brevity [24], as was the case in this study
duced capacity for comprehension during informed where it was administered to patients leaving the health
consent were also excluded from the study. Diminished facility after consultations with the nurses (patient exit
capacity for comprehension was determined by the inabil- interviews). The PSQ-18 instrument is reflective of
ity of prospective patients to comprehend or respond to Donabedian’s SPO constructs and succinctly measures
the information verbally provided by the interviewer dur- patient satisfaction with dimensions of care for which
ing informed consent. SPO constructs are intended. The authors are not aware
of any study that has used the PSQ-18 as a study instru-
Sample size determination and sampling of study ment to operationalise Avedis Donabedian’s SPO theor-
participants etical framework in SSA.
Using the subjects-to-variables ratio (minimum of 10 Mahomed et al. described the innovative approaches
subjects per variable in the study instrument) for esti- in the HIV programme leveraged for NCDs by the
mating sample size for studies utilising factor analysis NDoH [25]. From these, the study team consulted with
Ameh et al. BMC Health Services Research (2017) 17:229 Page 5 of 15

the health facility managers and officers of the Mpuma- this study because it is not the responsibility of South
langa Province Department of Health in selecting eight Africa’ Department of Health to provide transport for
dimensions of care that patients are able to respond to the implementation of the ICDM model. The 17 dimen-
as a result of their lived experiences with healthcare ser- sions of care in the adapted questionnaire are shown in
vices in the PHC health facilities. The rationale for this Fig. 1, and details of the adapted PSQ tool used in the
selection was because some aspects of these innovative current study for patients and operational managers are
approaches were functions performed by nurses, labora- shown in Additional files 3 and 4, respectively.
tory staff and health policy implementers which patients Eight dimensions of care were identified by experts on
were not privy to. quality of care in the study team as priority areas for en-
This study compared self-reported satisfaction of the pa- hancing service efficiency and quality of care: supply of
tients and self-reported satisfaction of the operational man- critical medicines, equipment, hospital referral, defaulter
agers with the dimensions of care listed in the ICDM tracing, prepacking of medicines, clinic appointments,
model using the multi-scale PSQ-18. This is in view of lit- patient waiting time, and coherence of integrated
erature depicting views of health care providers differing chronic disease care (Additional files 5 and 6) [9]. This
from users regarding the quality of health care [26]. Re- is because these priority areas are components of the
sponses to statements were scored on a five-point Likert tools and systems used in the successful HIV
scale ranging from 4 (strongly agree) to 0 (strongly dis- programme which is being leveraged to support or
agree) for positively-phrased statements, and from 4 scale-up services for improving the quality of care for
(strongly disagree) to 0 (strongly agree) for negatively- NCDs and patients interfaced directly with these areas
phrased statements for the purpose of undertaking con- in the health facilities (Fig. 1).
firmatory factor analysis and structural equation modeling.
Similar to another study in which the PSQ tool was Quality assurance
adapted to measure patient satisfaction with pharmacy The adapted PSQ tool for the patients was forward
services [27], this study adapted the PSQ-18 by altering translated to Tsonga (the local language) and back-
a number of statements to fit the ICDM model. For ex- translated to English by two experienced field workers
ample, the structure-related statement, “I have easy ac- who were blinded to each other. An experienced quanti-
cess to the medical specialists I need,” was changed to tative field worker was trained on how to administer the
the ICDM-process-related dimension, “Health care pro- adapted PSQ tool. A pilot study was conducted in Cork
viders usually refer me to the doctor/hospital when there clinic, a PHC facility situated outside the study site, to
is need for the doctor to review me - P5” (Additional files assure understanding and correct use of the PSQ tool.
2, 3 and 4). One structure-related (supply of critical Only a few statements had to be rephrased after the
medicines) and two process-related (defaulter tracing of pilot study.
patients and prepacking of medicines) variables were in- An important characteristic of the original PSQ-18,
cluded in the adapted questionnaire. One process- which was considered in the adaptation of the study in-
related statement in the PSQ-18 was changed from strument, is the control for Acquiescent Response Set
“health care providers act too business-like and imper- (ARS) - a tendency to agree with statements of opinion re-
sonal toward me” to “Health care providers are profes- gardless of their content [29]. Acquiescent response set is
sional in the conduct of their clinical duties”. Regarding a measurement error, specifically information bias, inher-
the types of outcome constructs (technical and interper- ent in surveys assessing satisfaction with medical care. Ac-
sonal) specified by Donabedian, the focus of this study cording to Ware et al. [29], there is a need to minimise
was on the subjective interpersonal outcome. Two out- information bias by assessing ARS in satisfaction surveys.
come statements on “satisfaction with perfect health Six variables were phrased in opposite directions, bringing
care” and “dissatisfaction with some care” in the PSQ-18 to 23 the total number of variables in the adapted ques-
were changed to the dimension on “satisfaction with co- tionnaire (Additional files 3 and 4). These measures are
herent integrated chronic disease care” and “dissatisfac- beneficial in detecting skewness toward satisfaction [29]
tion with coherent integrated chronic disease care”, and identifying specific programme areas that respondents
respectively. are satisfied or dissatisfied with.
Two statements around the financial costs of health
care (D1 and D2) were dropped during the adaptation of Data collection
the PSQ-18 (Additional file 3). This is because the gov- Having consulted with the professional nurses and re-
ernment of the Republic of South Africa implements a ceived their medicines, the prospective study partici-
pro-equity policy, a component of free health care for pants were invited to a (consultation) room designated
everyone using the public primary health system [28]. for patient interviews. Only the interviewer had access
However, transport-related costs were not considered in to this consultation room. Patients were invited to take
Ameh et al. BMC Health Services Research (2017) 17:229 Page 6 of 15

Fig. 1 The 17 dimensions of care for which the structure, process and outcome constructs were intended. *The dimensions in red colour indicate
the priority areas in the ICDM model

part in the satisfaction survey after explaining the pur- the study. Determining the quality of care in the ICDM
pose of the study. They were assured that there will be model was the second objective of this study which was
no penalty or loss of benefits to which they were entitled measured by conducting structural equation modelling
to if they chose to not participate in this study or decide (SEM) using the data on patients’ (dis)satisfaction with the
to discontinue participation in this study. Written in- dimensions of quality of care in the ICDM model. However,
formed consent was obtained from the patients who SEM could not be performed with the data collected from
were willing to participate in the study and interviews the operational managers because of the very small sample
were conducted with the patients. size (seven operational managers).
The following linear pathways were specified in the
The operationalisation of Donabedian’s theoretical SEM: (1) the unidirectional pathway which states that
framework good structure promotes good process and good process
The adapted PSQ contained measures reflective of SPO in turn promotes good outcome, (2) the mediation
constructs and was used to assess satisfaction of patients pathway which posits states that good structure directly
and operational managers with the dimensions of inte- promotes good outcome, good structure promotes good
grated chronic disease services. There was no clear div- process and good process in turn promotes good out-
ision of the statements in the adapted PSQ tool into the come; and (3) the reciprocal pathway which hypothesises
respective constructs. However, these statements have that good structure promotes good process, good
been categorised under these constructs in Additional process promotes good outcome and good outcome in
files 3 and 4 for clarification. In order to minimise bias turn promotes good process. The last two pathways were
that may result from assessing acquiescent response set, examined in this study to explore other linear relation-
the positive and negative statements did not follow each ships between SPO constructs other than the unidirec-
other in the questionnaire as shown in Additional files 3 tional pathway originally postulated by Donabedian
and 4. The respondents were judged to be satisfied with (Fig. 2).
the dimensions of care if the total relative frequency Fitting of the proposed pathways involved a four-step
was ≥ 50% for “strongly agree” and “agree” responses to systematic process using patient data. First, a priori iden-
positively-phrased statements. Similarly, the respondents tification of the variables for which the SPO constructs
were judged to be satisfied with the dimensions of qual- were intended was performed by the experts on quality
ity of care if the total relative frequency was ≥ 50% for of care on the study team in order to assess the validity
“strongly disagree” and “disagree” responses to of the adapted questionnaire (Additional files 3 and 4).
negatively-phrased statements. A satisfaction score of at This method was adopted by Kunkel et al. in which a
least 50% was considered an average score using a scale panel of experts categorised variables in a questionnaire
of 0% to 100%. into SPO constructs [17].
The patients and operational managers were scored com- Secondly, Cronbach’s alpha (range: 0–1), which is a
paratively on their (dis)satisfaction with the dimensions of measure of internal consistency, was used to quantify
care in the ICDM model to measure the first objective of the reliability of the multi-item variables in the adapted
Ameh et al. BMC Health Services Research (2017) 17:229 Page 7 of 15

Fig. 2 Pathways for operationalising Donabedian’s theory in the ICDM model of care in South Africa. a Unidirectional path: Good structure
should promote good process and good process in turn should promote good outcome. b Mediation path: Good structure directly promotes
good outcome, good structure promotes good process and good process in turn promotes good outcome. c Non-recursive (reciprocal) path:
Good stucture promotes good process, good process promotes good outcome and good outcome in turn promotes good process

PSQ in measuring the SPO constructs. Cronbach’s alpha as close as possible to the true ones by replacing
coefficient of reliability was categorised as excellent missing data with probable values based on other
(α ≥ 0.9), good (0.7 ≤ α < 0.9), acceptable (0.6 ≤ α < 0.7), available information [34].
poor (0.5 ≤ α < 0.6) and unacceptable (α < 0.5) [30]. Assessment of the fit of the pathways using MLMV
Next, the negative statements in the pair of statements approach was based on two or more of the following fit
phrased in opposite directions were dropped if there was indices [35]: (i) Relative/normed Chi-squared test statis-
no evidence of ARS. The fit of each construct and its in- tic is an absolute fit index that assesses the discrepancy
dividual items were assessed to remove any of the between observed and expected covariance matrices. It
remaining variables with low coefficient of determin- minimizes the impact of sample size on the model and
ation (CD < 0.2). Variables with low CD contribute high is derived by dividing the Chi square value by the de-
levels of error in the structural equation modelling [31]. grees of freedom (χ2/df ). Although there is no concensus
Thereafter, Confirmatory Factor Analysis (CFA) was regarding the acceptable ratio for this statistic, values
conducted to identify and remove the variables that did ranging from 2 to 5 are recommended as good fit indi-
not load significantly (factor loading < 0.300) onto their ces. [31]; (ii) Root Mean Squared Error of Approxima-
intended constructs. tion (RMSEA) is another absolute fit index that
The following step used structural equation model- measures how well a model with optimally chosen par-
ling (SEM) to assess the specified pathways, as used ameter estimates fit the population’s covariance matrix -
elsewhere [32], in order to determine the relationships RMSEA value ≤ 0.06 is a good fit; (iii) Comparative Fit
between the SPO constructs (Fig. 2). Selection of the Index (CFI) is an incremental fit index that assesses the
final path model was based on the variables that improvement in fit of the hypothesised model compared
reflected their intended factors (factor loading ≥ with a baseline (null) model, when population covari-
0.300). The Maximum Likelihood for Missing Values ance is assumed to be zero - (CFI ≥ 0.90 is a good fit);
(MLMV) technique was used to impute for S5, P1 (iv) Tucker-Lewis Index (TLI) is also an incremental fit
and P11 variables with 0.5%, 0.25% and 0.25% missing index that corrects for model complexity by favouring
observations, respectively. The MLMV is a technique parsimonious models over more complex ones - (TLI ≥
that handles missing data by estimating a set of pa- 0.90 is a good fit); and (v) Coefficient of determination
rameters that maximise the probability of getting the (CD) indicates how well data fit a statistical model. We
data that was observed. It is a more superior and used CD to decide the model that explained the most
preferable method for handling missing data than the variability. CD value of 1.00 is a perfect fit. The higher
more popular multiple imputation [33], which is a the number of criteria used, the better the fit of the
simulation-based method that predicts missing values model with the data [31].
Ameh et al. BMC Health Services Research (2017) 17:229 Page 8 of 15

Statistical analysis Table 1 Socio-demographic characteristics of the patients


Data were entered into Access 2010 and imported into attending health facilities in Agincourt sub-district in 2013 (n =
Stata 12.0 (College Station, TX, USA) for statistical ana- 435)
lysis. Relative frequencies were used to quantify satisfac- Variable Frequency (%)
tion of the patient and operational managers with the Age (years)
dimensions of integrated chronic disease services. At p- 18–29 23 (5.3)
value ≤ 0.05, CFA and SEM were used to fit the specified 30–39 69 (15.8)
structural path models in order to determine the quality
40–49 68 (15.6)
of care in the ICDM model from the patient perspective.
50–59 88 (20.3)
Results 60–79 187 (43.0)
Socio-demographic characteristics of the patients Mean ± SD (55 ± 16.5); Median = 56
Table 1 shows the mean age of the 435 chronic disease Gender
patients to be 55 ± 16 years. Forty-eight percent of the Female 354 (81.4)
patients were hypertensive; 81% females; 96% South Af-
Male 81 (18.6)
ricans; 99% unemployed; and 90% were not looking for a
paid job. Most of the patients received an old age grant Education (years)
(69%) and 88% of them had no formal or less than six No formal education 164 (37.6)
years of education. The response rate for the patient in- ≤6 217 (49.9)
terviews was 97%” >6 54 (12.5)
Type of grant
Satisfaction with structure-, process- and outcome-related
None 91 (20.9)
dimensions of care in the ICDM model
Figure 3a shows that the patients (P) and operational Old agea 299 (68.7)
managers (OM) reported being satisfied (scores ≥ 50%) Disability 44 (10.1)
with all the structure-related dimensions of care in the HIV 1 (0.3)
ICDM model. There were no statistically significant dif- Labour status
ferences (p > 0.05) between the satisfaction scores of the Not presently working 431 (99.0)
patients and operational managers with structure-related
Presently working 4 (1.0)
dimensions of care, except for availability of equipment
(S1): P-97% vs. OM-52%, p < 0.001. Nationality
Figure 3b shows that the operational managers reported South African 415 (95.5)
being satisfied (scores ≥ 50%) with all process-related di- Mozambican 20 (4.5)
mensions of care in the ICDM model. However, the pa- Chronic disease status b

tients were not satisfied (scores < 50%) with defaulter Hypertension 292 (67.0)
tracing of patients (P7-29%) and appointment systems
HIV 141 (32.4)
(P14-20%). Of all the process-related dimensions of care,
there were statistically significant differences in the scores Diabetes 2 (0.5)
of the patients and operational managers in appointment
a
Old age grant is a social security grant given to South Africans ≥ 60 years
of age
system (P14): P-20% vs. OM-100%, p < 0.001; physical b
Diagnoses of chronic diseases were retrieved from the patients’ clinic records
examination of patients (P11): P-96% vs. OM-57%, p <
0.001; defaulter tracing of patient (P7): P-29% vs. OM-
86%, p = 0.001; hospital referral of patients (P5): P-62% vs.
OM-100%, p = 0.039; and friendliness of the nurses to pa- Acquiescent response set
tients (P4): P-92% vs. OM-71%, p = 0.041; . Figure 3d shows patients’ satisfaction scores for the posi-
Figure 3c shows that the patients and operational man- tively- and negatively-phrased statements: supply of crit-
agers reported being satisfied (scores ≥ 50%) with three of ical drugs (93% vs. 92%), hospital referrals (62% vs. 62%),
the four outcome-related dimensions of care in the ICDM defaulter tracing (29% vs. 30%), prepacking of drugs be-
model. On the other hand, the patients and operational fore clinic visits (50% vs. 50%), time nurses spent with
managers were not satisfied (scores < 50%) with patient patients during consultation (70% vs. 70%) and coher-
waiting time (O4): P-17% vs. OMs-43%. A comparison of ence of integrated chronic disease care (97% vs. 96%).
the satisfaction scores of the patients and operational There were no statistically significant differences (p >
managers with all the outcome-related dimensions of care 0.05) in the responses of the patients to the pair of posi-
showed no statistically significant differences (p > 0.05). tively- and negatively-phrased statements.
Ameh et al. BMC Health Services Research (2017) 17:229 Page 9 of 15

Fig. 3 Satisfaction of respondents with the ICDM model and assessment of acquiescent response set for patients. *Priority areas in the ICDM
model †p-value < 0.05. a Satisfaction with structure-related dimensions of quality of care. b Satisfaction with process-related dimensions of quality
of care. c Satisfaction with outcome-related dimensions of quality of care. d Patient satisfaction with statements phrased in opposite directions

Fitting of the proposed structural pathways did not load significantly (factor loadings < 0.3) onto their
Figure 4 shows that the Cronbach’s alpha coefficients of intended constructs in the CFA (Table 2), and were
reliability of the variables intended for their respective dropped after CFA.
SPO constructs ranged from acceptable to good: struc-
ture (0.790), process (0.702), and outcome (0.600), an in- Assessment of fit indices of the specified path models
dication that the variables were a reliable measure of Figure 4 also shows the remaining nine variables that
their intended constructs [30]. reflected their intended SPO constructs (factor loading >
Before running the factor analysis, six negatively 0.300) in the structural equation model. These were
phrased statements (S3, P6, P8, P10, P13 and O2) in the three structure-related dimensions: availability of equip-
adapted questionnaire were dropped because there was no ment (S1), supply of critical medicines (S2) and accessi-
evidence of ARS in the pair of statements phrased in op- bility of chronic disease care (S4); three process-related
posite directions. In assessing the fit of the constructs and dimensions: attending to patients’ health needs (P2),
the remaining 17 variables, three process-related variables: professional conduct of the nurses (P3) and friendliness
communication with patients (P1), hospital referral (P5) of the nurses (P4); and three outcome-related dimen-
and physical examination of patients (P11) with coefficient sions: coherence of integrated chronic disease care (O1),
of determination values < 0.20 were dropped [31]. Of the patient confidence in the nurses (O3), and competence
remaining 14 variables, four process-related variables: de- of the nurses (O5).
faulter tracing of patients (P7), prepacking of drugs before The fit indices of the three specified pathways are as
clinic visit (P9), time patients spent with nurses during follows: (a) unidirectional pathway – [χ2/df = 2.44;
consultation (P12) and appointment system (P14); and RMSEA = 0.064 (90% CI - 0.052–0.077); CFI = 0.915;
one outcome-related variable: patient waiting time (O4) TLI = 0.892; CD = 0.911]; (b) mediation pathway – [χ2/
Ameh et al. BMC Health Services Research (2017) 17:229 Page 10 of 15

Fig. 4 Goodness-of-fit, reliability and correlation assessment of the relationships between structure, process and outcome. *Relationships between
the constructs represented by the Pearson correlation values. NB: The dimensions in red colour are the priority areas in the ICDM model. RMSEA -
Root Mean Squared Error of Approximation (≤0.06 is a good fit). CFI - Comparative Fit Index (CFI ≥ 0.90 is a good fit). TLI - Tucker-Lewis Index
(TLI ≥ 0.90 is a good fit). CD - Coefficient of determination (range 0–1. There is a perfect fit of the data with the model if CD = 1). Cronbach’s alpha
coefficient of reliability (≥0.6 is acceptable)

Table 2 The result of the confirmatory factor analysis


Constructs Variables Loading Standard error
Structure
Availability of equipment (S1) 0.462a 0.038
Supply of critical medicines (S2) 0.994a 0.012
a
Accessibility of services (S4) 0.383 0.041
Process
Attendance to patients’ needs (P2) 0.664a 0.035
Professionalism (P3) 0.758a 0.032
a
Friendliness (P4) 0.669 0.035
Defaulter tracing (P7) 0.200 0.056
Prepacking of drugs (P9) 0.268 0.055
Time spent with nurses (P12) 0.074 0.056
Appointment system (P14) 0.163 0.053
Outcome
Coherence (O1) 0.310a 0.057
Competence (O3) 0.485a 0.053
Waiting time (O4) 0.229 0.058
Confidence (O5) 0.651a 0.054
Variables with factor loading ≥ 0.300
a
Ameh et al. BMC Health Services Research (2017) 17:229 Page 11 of 15

df = 3.15; RMSEA = 0.058 (90% CI - 0.045–0.070); CFI = Findings from the mediation path model (Fig. 4)
0.931; TLI = 0.913; CD = 1.00; and (c) reciprocal pathway showed that three structure-related dimensions of care
– [χ2/df = 2.78; RMSEA = 0.059 (90% CI - 0.047–0.070); (availability of equipment; supply of critical medicines;
CFI = 0.919; TLI = 0.910; CD = 0.632]. and accessibility of chronic disease care) correlated dir-
Table 3 showed that when using at least two criteria, ectly with three outcome-related dimensions of care (co-
all the specified path models fit the data, but only the herence of integrated chronic disease care; and
mediation pathway fulfilled all the criteria used. competence of the nurses and patient confidence in the
nurses) and three process-related dimensions of care
Summary of the main findings (nurses’ friendliness with patients; professional conduct
The patients and operational managers’ were satisfied of the nurses; and attendance of the nurses to patients’
(scores ≥ 50%) with the following SPO related dimen- health needs). Independent of structure, good process
sions of care: correlated with good outcome, an indication that good
process mediated the relationship between good struc-
i) structure-related construct: availability of ture and good outcome.
equipment; supply of critical medicines; and
accessibility of chronic disease care. Discussion
ii) process-related construct: communication of the In view of the increasing emphasis on health system
nurses with patients; attendance of the nurses to strengthening and integration, this study contributes to
patients’ health needs; professional conduct of the the national and global debates on the feasibility of inte-
nurses; nurses’ friendliness with patients; hospital grating HIV services with those of NCDs. More specific-
referral of patients, pre-packing of medicines; phys- ally, we examined the satisfaction of patients and
ical examination of patients; and time nurses spent operational managers with the dimensions of integrated
with patients during consultation chronic disease services and evaluated the quality of care
iii) outcome-related: coherence of integrated chronic in the ICDM model from patient perspectives using
disease care; and competence of the nurses, and Donabedian’s theory of the relationships between SPO
patients’ confidence in the nurses. constructs as a measure of the quality of care.
Similar to a Togolese study in which the majority of
The patients and operational managers’ were less satis- service providers positively viewed the impact of inte-
fied (scores < 50%) with patient waiting time (an out- grating family planning services to the routine expanded
come construct). The patients recorded satisfaction programme on immunisation [36], the operational man-
scores < 50% for two process-related dimensions of care, agers in this study reported being satisfied with 16 of the
defaulter tracing of patients and appointment systems. 17 dimensions of quality of care in the ICDM model.
There were statistically significant differences (p < 0.05) However, this was less so for the patients who reported
in the satisfaction scores of the patients and operational satisfaction with 14 of these dimensions of care. The sig-
managers with regard to availability of equipment; nificant differences in the satisfaction scores of the pa-
friendliness of the nurses; hospital referral of patients; tients and operational managers in this study supports
defaulter tracing of patients; physical examination of pa- evidence-based literature that suggests assessing the sat-
tients; and appointment systems. isfaction of the quality of care from the perspectives of

Table 3 The result of the goodness of fit of the specified path models
Criteria Specified path models
Unidirectional Mediation Reciprocal
Relative Chi square statistic (χ2/df) 127/52 = 2.44✓ 164/52 = 3.15✓ 145/52 = 2.78✓
RMSEA value ≤ 0.06 0.064 0.058 ✓ 0.059 ✓
(90% CI - 0.052–0.077) (90% CI - 0.045–0.070) (90% CI - 0.047–0.070)
CFI ≥ 0.90 0.915 ✓ 0.931 ✓ 0.919 ✓
TLI ≥ 0.90 0.892 0.913 ✓ 0.910 ✓
CD close to 1.00 (perfect fit is preferred if CD value = 1.00) 0.911 ✓ 1.00 ✓ 0.632
Ranking** 3rd 1st 2nd
✓Indices with goodness of fit
**The mediation model ranked first because it fulfilled all five criteria (Relative/normed Chi square statistic, RMSEA, CFI, TLI and CD). In addition, it showed a
perfect fit based on CD value of 1.00
**The reciprocal model ranked second because it fulfilled four criteria (Relative/normed Chi square statistic, RMSEA, CFI and TLI)
**The unidirectional model ranked third because it fulfilled three criteria (Relative/normed Chi square statistic, CFI and CD)
Ameh et al. BMC Health Services Research (2017) 17:229 Page 12 of 15

both health providers and users [18] because of differing reported by patients as contributing to long waiting
views [26]. The patients rated satisfaction with availabil- time in the qualitative study were late arrival of filing
ity of equipment higher than the operational managers clerks and nurses; long morning prayer sessions be-
because the patients may not be aware of the lack of fore commencement of clinical duties; staff meetings;
equipment. The patients’ satisfaction scores for friendli- prolonged tea or lunch breaks; and nurses giving
ness of the nurses and physical examination of patients preferential treatment to friends or relatives who skip
was higher than those of the operational managers. The the queues [43].
operational managers who responded to the interviews The lack of an Acquiescent Response Set (ARS) found
were professional nurses who often performed a dual in this study does not support literature evidence that
role of providing routine care to the patients and man- suggests patient satisfaction surveys are almost always
aging the facilities. In the course of performing their ad- skewed toward satisfaction with positively worded state-
ministrative duties in the office, these managers may not ments [23]. The reasonable explanation for the absence
have the opportunity to see other professional nurses be- of ARS in this study can be attributed to two factors: (1)
ing friendly to patients in the consultation rooms. This the fieldworker received training on how to read the
may have accounted for the managers’ lower satisfaction statements in the interviewer-administered question-
scores compared with the patients’ scores. naire very slowly and carefully to the patients in a way
An earlier household survey conducted in the that the statements were understood, and (2) the ques-
study site reported health system weakness as one of tionnaire was pre-tested to provide feedback to the study
the barriers to chronic disease care. At the time of team. The purpose of testing for ARS in this study was
the survey in 2004, community members attended to minimise information bias [29] by checking to see if
public hospitals for diagnosis and treatment of the patients understood the statements in the adapted
chronic illness due to the lack of capacity and ser- PSQ-18. As implementation of ARS does not eliminate
vices in the PHC facilities [37]. A decade after the coercion, we addressed the possibility of coercion, which
2004 survey and two years after the initiation of the is more likely to occur in people of low socioeconomic
ICDM model in South Africa, community members status by assuring patients that there would no penalty
now have access to chronic disease services in PHC or loss of patient benefits if they chose not to participate
facilities in their local areas. These facilities have a or decided to discontinue participation at any point in
more regular supply of critical drugs and trained the study.
professional nurses who are better able to provide A Swedish study used Donabedian SPO theoretical
integrated services for the diagnosis and treatment framework to show a statistically significant relationship
of chronic diseases. This may be an indication some between SPO constructs through the mediation pathway
progress that has been made in chronic disease care [17]. This research corroborates the Swedish study and
in the study setting. further reinforces the usefulness of Donabedian’s theory in
In this study, patient waiting time was the only di- evaluating the quality of healthcare generally, and more
mension of care in the ICDM model in which the pa- specifically in the context of the ICDM model. The per-
tients and operational managers reported low ception of the patients about the quality of care in the
satisfaction scores. Similar studies assessing the qual- ICDM model can be interpreted to mean that the
ity of service in public clinics in South Africa showed provision of good structure directly promotes good out-
that the clinics were easily accessible and services come; and that the relationship between good structure
were of acceptable quality [38], but the time spent by and good outcome is mediated by good process. More
patients at the clinic to complete the services was specifically, the patients thought that the provision of
very long [38, 39]. These findings suggest that public equipment, drugs and accessibility of chronic disease ser-
health services in many resource-constrained LMICs vices contributed to the nurses’ ability to be professional
are characterised by long waiting periods [40–42], in their duties, become friendly to patients and attend to
which could be a consequence of operational chal- patients health needs. If the nurses performed these du-
lenges such as performance of multiple tasks and ties, the patients had confidence in the nurses, thought
work overload of health workers [18]. In addition to that the nurses were competent, and perceived there was
staff shortage which was reported by operational coherence in the services provided by the nurses.
managers and patients in the qualitative component Although Donabedian’s framework continues to be the
[43] of the broader mixed methods study, patients dominant touchstone paradigm for assessing the quality
who missed previous clinic appointments were being of health care, it has been described as too linear to rec-
made to wait in the queues during subsequent visits ognise complex interactions between SPO constructs
until nurses had attended to patients who were on [16]. Donebedian’s critics argue that his theory fails to
the appointment list for that day [43]. Other factors incorporate patient characteristics which are important
Ameh et al. BMC Health Services Research (2017) 17:229 Page 13 of 15

precursors in the evaluation of the quality of health ser- priority areas of the ICDM model in the study settings.
vices [44]. However, these limitations do not affect the Study findings should be interpreted in light of the fol-
validity of our study for the following reasons. First, the lowing limitations: (1) Causal relationships between SPO
linearity of Donabedian’s theory forms the basis of our constructs cannot be inferred because this study was
study which assesses the linearity of the relationships be- cross-sectional by design, (2) Data on interpersonal out-
tween SPO constructs through the specified unidirec- comes (dis)satisfaction with care do not necessarily reflect
tional, mediation and reciprocal pathways. The linearity technical outcomes (e.g. reduced diseases, disabilities and
of Donabedian’s theory would have been a limitation in deaths) in the ICDM model of care in the study settings,
our study if we sought to determine non-linear relation- (3) The perspectives of clinic defaulters were not taken
ships between SPO constructs . Regarding the limitation into account, (4) Inferences could not be made about the
of Donabedian’s theory not accounting for patients’ (dis)satisfaction of other professional nurses with services
socio-demographic characteristics, it is important to in the ICDM model, due to the small number of oper-
note that patients’ characteristics could not have been ational managers who were interviewed, and (5) It was
categorised as dimensions of care under SPO constructs not feasible to include all the priority dimensions of care
in the ICDM model. This is because patients’ character- in the questionnaire because patients were not privy to
istics do not fit into Donebedian’s definition of SPO con- some of these aspects of care in the ICDM model.
structs and therefore have no role to play in explaining
Donabedian’s theory of quality of care; hence, the ration-
Conclusion
ale for selecting his theoretical framework for evaluating
The patients and operational managers were satisfied
the quality of care in the ICDM model.
with many areas of the integrated chronic disease ser-
vices, but had divergent opinions about satisfaction with
Implications
some dimensions of care. Of the 17 dimensions of care
Of the eight priority areas in the ICDM model (supply
assessed in the ICDM model, nine refelected their
of critical medicines, equipment, hospital referral, de-
intended constructs. Five of the eight priority areas
faulter tracing, prepacking of medicines, clinic appoint-
assessed in the ICDM model (hospital referral, defaulter
ments, reducing patient waiting time, and coherence of
tracing, prepacking of medicines, clinic appointments,
integrated chronic disease care), the supply of critical
and patient waiting time) did not reflect their intended
medicines, availability of equipment and coherence of
constructs; hence the need to strengthen services in
integrated chronic disease care reflected their intended
these areas.
constructs in the final model. On the other hand, the
remaining five priority areas (hospital referral, defaulter
tracing, prepacking of medicines, clinic appointments, Additional files
and reducing patient waiting time) di not reflect their
intended constructs. Additional file 1: Sampling of study participants. (PDF 95 kb)
The authors suggest an interaction of factors responsible Additional file 2: Patient satisfaction questionnaire-18 developed by
Ware et al. (PDF 285 kb)
for why hospital referral, defaulter tracing, prepacking of
Additional file 3: Patient satisfaction questionnaire adapted for patients
medicines, clinic appointments, and reducing patient wait- in the study. (PDF 90 kb)
ing time did not reflect their intended constructs in the Additional file 4: Patient satisfaction questionnaire adapted for
study settings. For instance, patient waiting time may have operational managers in the study. (PDF 190 kb)
been unnecessarily prolonged in the study settings due to Additional file 5: Innovative approaches in the HIV programme
many factors. The purpose of prepacking medicines before leveraged for NCDs in the ICDM model by the NDoH in South Africa.
(PDF 89 kb)
patient arrival is to reduce patient waiting time; however,
Additional file 6: Table of definition of terms used in the article. (PDF 97 kb)
the high rate of patient’s missed appointments and un-
availability of prepacking bags could have served as a de-
terrent from nurses prepacking medicines [43]. Abbreviations
ARS: Acquiescent response set; ART: Antiretroviral treatment; CD: Coefficient
of determination; CFA: Confirmatory factor analysis; CFI: Comparative fit
Study strengths and limitations index; CHWs: Community health workers; HDSS: Health and demographic
The main strength of this study was the use of the pa- surveillance system; HIV/AIDS: Human immunodeficiency virus/acquired
tient satisfaction survey to evaluate the quality of care in immune deficiency syndrome; ICDM: Integrated chronic disease
management; LMICs: Low- and middle-income countries; MLMV: The
the ICDM model in PHC facilities in a rural setting in maximum likelihood for missing values; NCDs: Non-communicable diseases;
South Africa using Donabedian’s theory. In addition, we NDoH: National Department of Health; PHC: Primary health care; PSQ: Patient
assessed satisfaction with integrated chronic disease ser- satisfaction questionnaire; RMSEA: Root mean squared error of
approximation; SEM: Structural equation modelling; SPO: Structure, process,
vices, from the perspectives of healthcare providers and and outcome; SSA: Sub-Saharan Africa; TLI: Tucker-Lewis Index; UNAIDS: Joint
users. This study also provided insight on some of the United Nations Programme on HIV/AIDS; WHO: World Health Organization.
Ameh et al. BMC Health Services Research (2017) 17:229 Page 14 of 15

Acknowledgements Received: 15 April 2016 Accepted: 18 March 2017


Data collection for this study was made possible through the MRC/Wits Rural
Public Health and Health Transitions Research Unit (Agincourt) South Africa.
The authors are grateful to Caroline W. Kabiru of the ADDRF Programme for
the critical review of an earlier version of this manuscript. References
1. Statistics South Africa . Mid year population estimates. 2014. http://beta2.
statssa.gov.za/publications/P0302/P03022014.pdf. Accessed 2 Feb 2015.
Funding 2. World Health Organization. Noncommunicable diseases country profiles.
Research reported in this publication was supported by the Fogarty 2014. http://www.who.int/nmh/countries/zaf_en.pdf?ua=1. Accessed 8 Feb
International Centre of the National Institutes of Health under the Award 2015.
number D43 TW008330. The content is solely the responsibility of the 3. Chan M. (2007) Speech at the launch of the UK Department for
authors and does not necessarily represent the official view of the National International Development’s new health strategy. 2007. http://www.who.int/
Institutes of Health. This research was also co-funded by the African Doctoral dg/speeches/2007/050607_DFID/en/. Accessed 23 July 2013.
Dissertation Research Fellowship, Nairobi, Kenya. 4. The Joint United Nations Programme on HIV/AIDS. Chronic care of HIV and
non-communicable diseases: How to leverage the HIV experience. 2011.
http://www.unaids.org/en/media/unaids/contentassets/documents/
Availability of data and materials unaidspublication/2011/20110526_JC2145_Chronic_care_of_HIV-1.pdf.
The dataset on which the conclusions of this manuscript rely will not be Accessed 8 July 2014.
made available given the conditions stated in the informed consent form to 5. Janssens B, Van Damme W, Raleigh B, Gupta J, Khem S, et al. Offering
protect the identity of the patients and health facilities in which the study integrated care for HIV/AIDS, diabetes and hypertension within chronic
was conducted. However, datasets that do not identify the patients and disease clinics in Cambodia. Bull World Health Organ. 2007;85(11):880–5.
health facilities could be shared on request after approval by the Committee 6. Levitt NS, Steyn K, Dave J, Bradshaw D. Chronic noncommunicable diseases
for Research on Human Subjects (Medical) of the University of the and HIV-AIDS on a collision course: relevance for health care delivery,
Witwatersrand, Johannesburg, and the Mpumalanga Provincial Research and particularly in low-resource settings - insights from South Africa. Am J Clin
Ethics Committee. Nutr. 2011;94(6):1690S–6S.
7. Mayosi BM, Flisher AJ, Lalloo UG, Sitas F, Tollman SM, et al. The burden of
non-communicable diseases in South Africa. Lancet. 2009;374(9693):934–47.
Authors’ contributions 8. Tollman SM, Kahn K, Sartorius B, Collinson MA, Clark SJ, et al. Implications of
SA, FXG, SMT and KKG developed the study protocol. SA conducted the mortality transition for primary health care in rural South Africa: a
statistical analysis. The results were interpreted by SA, FXG, SMT, KKG, and KK. population-based surveillance study. Lancet. 2008;372(9642):893–901.
SA, FXG, SMT, KKG and KK critically reviewed the manuscript and approved 9. Department of Health, Republic of South Africa. Integrated Chronic Disease
the final version. Management Manual. 2014. http://www.kznhealth.gov.za/family/Integrated-
chronic-disease-management-manual.pdf. Accessed 23 Apr 2015.
10. World Health Organization. Integrated health services - what and why?
Competing interests Making health systems work. 2008. Technical Brief No. 1. (Page 1). http://
The authors declare that they have no competing interests. www.who.int/healthsystems/technical_brief_final.pdf. Accessed 8 July 2014.
11. Donabedian A. Evaluating the quality of medical care. Milbank Mem Fund
Q. 1966;44(3):166–206.
Consent for publication 12. Klein MW. Problems of Measuring Patient Care in the Out Patient
Not applicable. Department. J Health Hum Behav. 1961;2:138–44.
13. Donabedian A. An introduction to quality assurance in health care, vol. 1.
1st ed. New York: Oxford University Press; 2003.
Ethics approval and consent to participate
14. Donabedian A. The quality of care. How can it be assessed? JAMA. 1998;
Ethical clearance for this research was granted by the Committee for
260(12):1743–8.
Research on Human Subjects (Medical) of the University of the
15. Donabedian A. Quality assessment and assurance: unity of purpose,
Witwatersrand, Johannesburg, South Africa (Ref No. M120943) and the
diversity of means. Inquiry. 1988;25(1):173–92.
Mpumalanga Provincial Research and Ethics Committee. In addition to the
16. Mitchell PH, Ferketich S, Jennings BM, Quality health outcomes model.
efforts to ensure confidentiality earlier mentioned during data collection,
American Academy of Nursing Expert Panel on Quality Health Care. Image J
patients’ information were encrypted in a computer programme that keeps
Nurs Sch. 1998;30(1):43–6.
the information in a form that does not allow anyone to interpret it, except
17. Kunkel S, Rosenqvist U, Westerling R. The structure of quality systems is
the study team, and a study number was generated for each patient. Written
important to the process and outcome, an empirical study of 386 hospital
informed consent was obtained from the patients who were willing to
departments in Sweden. BMC Health Serv Res. 2007;7:104.
participate in the study.
18. Briggs CJ, Garner P. Strategies for integrating primary health services in
middle- and low-income countries at the point of delivery. Cochrane
Database Syst Rev. 2006;(2):CD003318
Publisher’s Note 19. Kahn K, Collinson MA, Gomez-Olive FX, Mokoena O, Twine R, et al. Profile:
Springer Nature remains neutral with regard to jurisdictional claims in Agincourt health and socio-demographic surveillance system. Int J
published maps and institutional affiliations. Epidemiol. 2012;41(4):988–1001.
20. Kerry Cullinan. Health services in South Africa: A basic introduction. Health-e
Author details News Service. 2006. http://www.health-e.org.za/2006/01/29/health-services-
1
Medical Research Council/Wits University Rural Public Health and Health in-south-africa-a-basic-introduction/. Accessed 15 June 2014.
Transitions Research Unit (Agincourt), School of Public Health, Faculty of 21. Pett MA, Lackey NR, Sullivan JJ. Making Sense of Factor Analysis: The use of
Health Sciences, University of the Witwatersrand, Johannesburg, South Africa. factor analysis for instrument development in health care research.
2
Department of Community Medicine, Faculty of Medicine, College of California: Sage Publications Inc; 2003.
Medical Sciences, University of Calabar, Calabar, Cross River State, Nigeria. 22. Garson DG. Factor Analysis: Statnotes. Retrieved March 22, 2008, from North
3
The International Network for the Demographic Evaluation of Populations Carolina State University Public Administration Program.
and Their Health in Developing Countries (INDEPTH), Accra, Ghana. 4Umeå 23. Ware JE, Snyder MK, Wright WR. Development and validation of scales to
Centre for Global Health Research, Epidemiology and Global Health, Umeå measure patient satisfaction with health care services: review of literature,
University, Umeå, Sweden. 5Division of Epidemiology and Biostatistics, School overview of methods and results regarding construction of scales.
of Public Health, University of the Witwatersrand, Johannesburg, South Africa. Springfield: National Technical Information Service; 1976.
6
Julius Global Health, Julius Center for Health Sciences and Primary Care, 24. Patient Satisfaction Questionnaire from RAND Health. http://www.rand.org/
University Medical Center Utrecht, Utrecht, The Netherlands. health/surveys_tools/psq.html. Accessed 2 Sept 2014.
Ameh et al. BMC Health Services Research (2017) 17:229 Page 15 of 15

25. Mahomed OH, Asmall S, Freeman M. An integrated chronic disease


management model: a diagonal approach to health system strengthening
in South Africa. J Health Care Poor Underserved. 2014;25(4):1723–9.
26. The Health Foundation: Measuring patient experience. http://www.health.
org.uk/publication/measuring-patient-experience. Accessed 1 Feb 2015.
27. MacKeigan LD, Larson LN. Development and validation of an instrument to
measure patient satisfaction with pharmacy services. Med Care. 1989;27(5):
522–36.
28. African National Congress. Policy Documents. A basic guide to the
Reconstruction and Development Programme. http://www.anc.org.za/
content/basic-guide-reconstruction-and-development-programme.
Accessed 29 May 2014.
29. Ware JE. Effects of acquiescent response set on patient satisfaction ratings.
Med Care. 1978;16(4):327–36.
30. Kline P. The handbook of psychological testing. 2nd ed. London: Routledge;
2000.
31. Hooper D, Coughlan J, Mullen R. Structural Equation Modelling: Guidelines
for Determining Model Fit. Electron J Bus Res Methods. 2008;6(1):53–60.
32. Uthman OA, Moradi T, Lawoko S. Are individual and community acceptance
and witnessing of intimate partner violence related to its occurrence?
Multilevel structural equation model. PLoS One. 2011;6(12):e27738.
33. Allison PD. Handling Missing Data by Maximum Likelihood. http://www.
statisticalhorizons.com/wp-content/uploads/MissingDataByML.pdf. Accessed
12 Oct 2014.
34. Rubin DB. Multiple imputation after 18+ years. J Am Stat Assoc. 1996;91:
473–89.
35. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Struct Equ Model
Multidiscip J. 1999;6(1):1–55.
36. Huntington D, Aplogan A. The integration of family planning and childhood
immunization services in Togo. Stud Fam Plann. 1994;25(3):176–83.
37. Goudge J, Gilson L, Russell S, Gumede T, Mills A. Affordability, availability
and acceptability barriers to health care for the chronically ill: Longitudinal
case studies from South Africa. BMC Health Serv Res. 2009;9(1):75.
38. Kinkel H, Adelekan A, Marcus T, Wolvaardt G. Assessment of service quality
of public antiretroviral treatment (ART) clinics in South Africa: a cross-
sectional study. BMC Health Serv Res. 2012;12(1):228.
39. Wouters E, Heunis C, Van Rensburg D, Meulemans H. Patient satisfaction
with antiretroviral services at primary health-care facilities in the Free State,
South Africa–a two-year study using four waves of cross-sectional data.
BMC Health Serv Res. 2008;8:210.
40. Babirye JN, Engebretsen IM, Rutebemberwa E, Kiguli J, Nuwaha F. Urban
settings do not ensure access to services: findings from the immunisation
programme in Kampala Uganda. BMC Health Serv Res. 2014;14:111.
41. Fawole AO, Okunlola MA, Adekunle AO. Clients' perceptions of the quality
of antenatal care. J Natl Med Assoc. 2008;100(9):1052–8.
42. Lule GS, Tugumisirize J, Ndekha M. Quality of care and its effects on
utilisation of maternity services at health centre level. East Afr Med J. 2000;
77(5):250–5.
43. Ameh S, Klipstein-Grobusch K, D’Ambruoso L, Kahn K, Tollman SM, Gómez-
Olivé FX. Quality of integrated chronic disease care in rural South Africa:
user and provider perspectives. Health Policy Plan. 2016. doi:10.1093/
heapol/czw118. First published online September 1.
44. Coyle YM, Battles JB. Using antecedents of medical care to develop valid
quality of care measures. Int J Qual Health Care. 1999;11(1):5–12.

Submit your next manuscript to BioMed Central


and we will help you at every step:
• We accept pre-submission inquiries
• Our selector tool helps you to find the most relevant journal
• We provide round the clock customer support
• Convenient online submission
• Thorough peer review
• Inclusion in PubMed and all major indexing services
• Maximum visibility for your research

Submit your manuscript at


www.biomedcentral.com/submit

You might also like