Content uploaded by Christine J. Fenenga
Author content
All content in this area was uploaded by Christine J. Fenenga on Dec 22, 2015
Content may be subject to copyright.
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
Action Research
0(0) 1–19
!The Author(s) 2015
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1476750314568209
arj.sagepub.com
Article
Participatory action
approach for client-
centered health insurance
Christine J Fenenga
Population Research Centre, Faculty of Spatial Sciences, University of Groningen,
The Netherlands; Amsterdam Institute for Global Health and Development,
Pietersbergweg, Amsterdam, The Netherlands
Edward Nketiah-Amponsah
Department of Economics, University of Ghana Accra, Ghana
Ajay Bailey and Inge Hutter
Population Research Centre, Faculty of Spatial Science, University of Groningen, The
Netherlands
Abstract
This methodological paper systematically describes the steps towards exploring client-
perceived barriers to enroll in the National Health Insurance Scheme in Ghana. To
respond to the stagnating enrollment rate and to provide evidence-based information
for developing policy and practice, the request for this study was made by the insurance
authority. We adopted a Participatory Action Approach (PAA) and employed a linked
trajectory of qualitative and quantitative methods to inquire deeply into clients’ motives
and assumptions, and the root causes of these barriers, while transferring learning
among multiple stakeholders in the system: the client, the healthcare provider, and
the insurer. Data were collected in 2011 and 2012 among clients in their local commu-
nities in two regions in Ghana. Clients were participants in the research, expressing and
sharing their emic perspectives and commonly shared opinions with the other stake-
holders. This paper explores the contribution of the PAA to healthcare and health
insurance studies. We claim that the PAA is a necessary approach for the study of a
complex situation such as the health insurance context in Ghana as well as for the
facilitation of learning and facilitate democratic decision-making. Making multidisciplin-
ary stakeholders, including clients, part of the research contributes to finding solutions
that are acceptable to different stakeholders. Our study contributes to PAA method-
ology knowledge and provides new insights on combining qualitative and quantitative
Corresponding author:
Christine J Fenenga, Amsterdam Institute for Global Health and Development, Pietersbergweg, 17 Trinity
Building C, Amsterdam 1105BM, The Netherlands.
Email: c.fenenga@aighd.org
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
research methods. We propose this as a model to be used within action research that
can also be applied in other contexts outside of Africa.
Keywords
Participatory Action Approach, clients, perceptions, health care, health insurance,
policy development, Ghana
Introduction
Data that capture the experiences of people, their perceptions of illness, health
seeking, healthcare and health insurance services, their concerns and their expect-
ations, are extremely useful for managers and policymakers in order to design
policies and programs and navigate them towards successful implementation.
Making optimal use of this information is likely to contribute to building trust
and participation (Asenso-Boadi, 2010; Jehu-Appiah et al., 2011a, 2011b). In the
rapidly expanding field of health insurance in Africa, where scientific evidence of
the effect of such programs is extremely important, the vast majority of research
currently presented in the literature focuses on willingness to join and pay for
insurance, mainly based on socioeconomic surveys (Blanchet et al., 2012;
Brugiavini & Pace, 2010; Jehu-Appiah, et al., 2010; McIntyre et al., 2005;
Sarpong et al., 2010).
The context of this study is Ghana, a country located on the west coast of Africa
with a population of over 25 million people. Health care is provided by public and
private healthcare providers (including church-based). The system has five levels of
providers: health posts, which are the first level primary care for rural areas; health
centers and clinics; district hospitals; regional hospitals; and tertiary hospitals.
Health care is variable throughout the country, with urban areas being served
well with an increasing choice of private healthcare facilities, while rural areas
have a more limited choice. Services here are mainly offered by public and/or
church-based health facilities. Patients in these areas often have to travel longer
distances to access health care or else rely on the traditional healers. In total, there
are over 5500 healthcare facilities in Ghana. Many qualified health professionals
prefer practicing in urban areas.
Ghana is one of the first countries in Africa to have introduced national health
insurance. The National Health Insurance Scheme (NHIS) was established by an
Act of Parliament (Act 650) in 2003 and aims to secure financial risk protection
against the cost of healthcare services for all residents in Ghana. Massive infor-
mation campaigns were organized for the population and for people working in the
health sector. The NHIS contracted close to 4000 accredited healthcare facilities to
deliver the healthcare services. Despite major progress in the scheme’s implemen-
tation and attainment of a cumulative enrollment rate of over 70% (17 million
people), the percentage of people actively using the scheme reached a mere 34% of
2Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
the population in 2011 (http://www.nhis.gov.gh). This reveals that a large portion
of the population is not enrolling in the scheme or is dropping out, despite exemp-
tion mechanisms for those below 18 and above 70 years of age, pregnant women
and indigenous people. Earlier studies revealed a variety of implementation prob-
lems, such as insufficient information and communication, low perceived quality of
health services delivered by the healthcare facilities, and delays in card production
and distribution (Jehu-Appiah et al., 2011a; Ministry of Health and Ghana Health
Service, 2010; NHIA, 2010; Sarpong et al., 2010).
This study has been commissioned by the National Health Insurance Authority
(NHIA) with the intention to obtain a better understanding of the clients’ perceived
barriers to enroll in the scheme so that it can improve its services. This study aims
to gain a better understanding of clients’ perceptions regarding illness, health-
seeking, health care and health insurance services, in order to contribute practical
solutions to facilitate client-centered health care and health insurance services and
increase participation in the scheme. We do not focus on barriers but rather on
what motivates people to use health care and health insurance services. The result-
ing scientific evidence is intended to inform policy and practice.
Understanding clients’ perceptions implies studying clients in their own sociocul-
tural context and taking account of other stakeholders who are part of that context.
The initial stakeholder meeting revealed not only the urgency of the issue but also the
great need for solutions to improve the situation. The aforementioned features of
multi-stakeholder involvement in a complex environment and need for social change
required a Participatory Action Approach (PAA) combined with the application of a
linked trajectory of qualitative and quantitative methods. In this paper, we aim to
demonstrate the value of our research design. We argue that the PAA and used
methods are instrumental in the inquiry process to generate rich data grounded in
practice, which helps to obtain a better understanding of clients’ perceptions, behav-
ior and decision-making. Moreover, the process that allows stakeholders to engage
in the research as partners is key in facilitating ownership and learning, both import-
ant conditions for social change and for the improvement of services. This study,
therefore, also contributes to methodology knowledge. We propose this as a model
within action research that can be applied in contexts in Africa and beyond.
This paper is organized as follows: The next section describes the study design,
including the underlying theories, research approach, and data management. We
illustrate this with some examples of data generated through our approach and
methods. The last section, Conclusion and recommendations, concludes with a
discussion on the scientific contribution of this research.
Research design
Underlying theories
Studying perceptions of clients and identifying what motivates people to use health
care and health insurance services requires positioning the client in a broader
Fenenga et al. 3
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
sociocultural system, while acknowledging the political dynamics and complex
interrelationship between the clients, the healthcare providers and the health insur-
ance provider. A review of the existing literature in this field resulted in a selection
of social theories used to develop the deductive conceptual framework and research
questions. These theories are the Socio-Anthropological Explanatory model on
healthcare systems (Kleinman, 1978); the Social Capital theory (Bourdieu, 1986;
Coleman, 1988; Fukuyama, 2000; Grootaert & Bastelaer, 2002; Putnam, 2001;
Woolcock & Harper, 2001); the New Institutional Economic theory (North,
1990); and Mechanics Three-Dimensional Trust model (Mechanic, 1998).
Anticipating the broader sociocultural system, key concepts from these theories
that we expect to influence clients’ perceptions and motives to use healthcare and
health insurance services are explanatory models
1
and social capital.
2
Differences in
explanatory models of ‘illness’, ‘the need for care’, ‘quality of care’ and ‘role in
care’ between clients, healthcare providers and the NHIS are believed to lead to
miscommunication, resulting in decreased trust in relationships. Social relations
within and between (community) groups and the social support that people derive
from their social networks do influence perceptions and choice of healthcare. We
expect that diverse channels of social capital offer potential opportunities to clients,
healthcare providers and health insurance schemes for mobilization, information
sharing, collective decision-making and reducing opportunistic behavior. These
theories are explored in Fenenga et al. (2013). Based on a review of these theories,
trust emerges as a pivotal component in clients’ perceptions and health-seeking
behavior. We are referring to interpersonal trust between clients and their doctor
or the insurance officer, but also to the trust clients have in the health facility and
insurance organization as institutions (institutional trust). Finally, their behavior
and decision-making is influenced by rooted cultural beliefs as well as the views and
opinions of family, friends and other people in their community whom they trust.
People’s need for trust is related to the vulnerability associated with being ill, but
also to the asymmetries in information (Calnan & Row, 2006). The complex inter-
relation between clients, healthcare providers and health insurer, and the NHIS
requires not only looking at the perspectives of clients, but also of the other two
stakeholders in the tripod (Figure 1). This provides insights from three perspectives
and comparison of data. The more specific conceptual framework that guides the
design, interviews and analysis of clients’ perceptions is presented and explained in
detail in Fenenga et al. (2013).
The stakeholders
This methodological paper focuses on the client perspective. Our data were col-
lected from NHIS clients and the potential clients of 64 primary healthcare facilities
in the Western Region and the Greater Accra Region in Ghana, characterized by a
mixed urban and rural population. However, the larger study consists of three sub-
studies that were conducted simultaneously. Each of the sub-studies explores the
perspectives of one of the stakeholder groups in the tripod (Figure 1). These are: (1)
4Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
the clients (clients and potential clients) of the NHIS; (2) the healthcare providers
(facility managers, nurses, doctors, and laboratory staff) of the 64 selected primary
healthcare facilities (public, church-based, and private) in Western Region and
Greater Accra Region, with health authorities from district, regional and national
levels participating in the stakeholder meetings (SM); and (3) the NHIS staff from
the district, regional, and national levels. At various moments in the research, these
three stakeholder groups are brought together as part of the PAA. This will be
explained in the following paragraphs and is illustrated in Figure 2. This paper
should be seen in the context of this larger study. Papers of the other studies are
forthcoming (Duku, Kaba-Alhassan, forthcoming).
The research team
The researcher is a PhD student, who works in a team with two other PhD stu-
dents, (two Ghanaians, one Dutch) each responsible for one sub-study as described
earlier. The PhD students are supported each by two supervisors. Besides overall
responsibility of a sub-study, other tasks were training the research team, managing
the data, liaise with the other researchers in the two other sub-projects, moderation
of focus group discussions (FGD), and SM. While the Ghanaian researchers had
the advantage to be more familiar with the specific sociocultural context and
Ghanian health sector than the Dutch researcher, the diversity of local languages
required moderators with specific linguistic skills to support the three researchers.
The disadvantage of being familiar with the local context is that the researcher too
easily understands what the participant is talking about, hence this knowledge and
experience can block recognizing new or unexpected issues (Hennink, Hutter, &
Bailey, 2011). Interviewees who know that the person interviewing has had similar
experience might be less explicit in their answers because they suppose that the
researchers knows already what they mean. The researchers were aware of this.
Figure 1. Perspectives of the three stakeholder groups.
Fenenga et al. 5
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
They were trained in qualitative and quantitative research methods. The
researcher conducting the client sub-study was trained in participatory action
research and was responsible for this approach in the study. All researchers and
moderators were made aware of their position in the study and the importance to
make the clients, healthcare providers and health insurer active participants in the
research.
Participatory Action Approach
To explore clients’ perceived barriers and to search for solutions to reduce these
barriers and improve the clients’ position, it is important to understand clients’
health seeking behavior (Verstehen), their underlying sociocultural schemas, and
their experiences enrolling in the NHIS (Snape & Spencer, 2003). We studied NHIS
clients and potential clients in their own communities, where we obtained inside
perspectives and captured the decision-making processes. This approach features
the PAA to research, which is context-specific, cyclical and reflective in design and
focuses on problem solving (Hardon et al. 1994; Hennink et al., 2011; Koning &
Martin, 1996; Maxwell, 2005; McIntyre, 2008). Another important feature of the
PAA is the emancipatory aspect of participation of clients, healthcare providers
and health insurance staff
3
as partners throughout the research process: from
Figure 2. PAA and linked trajectory of methods.
6Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
initiating the study and identifying the problem to testing the interventions. At
various stages in the research process, we created space for dialogue among stake-
holders, pooling knowledge and facilitating awareness, interest and commitment to
social change. Democratic processes are expected to ultimately influence policy and
practice in health care and health insurance services.
Linked trajectory of methods
To collect the data, we followed a linked trajectory of methods, which can be seen as
a sequential procedure of methods (Creswell, 2003). However, as described by
Bailey and Hutter (2008), the linked trajectory of methods is more than a mixed
method. In most studies that use mixed methods, the reasoning on how the
researchers mixed the methods is not clear. We claim that applying the methods
in a specific sequence (Figure 2), interpreting and validating information from the
previously used method into the next one, will result in strongly grounded and
generalizable evidence. In the following paragraph, we elaborate on the PAA and
linked trajectory of methods and illustrate these with examples from our data.
PAA and linked trajectory of methods together
Figure 2 shows the framework of our study approach and data collection methods.
The left column indicates the stakeholders involved in the different steps of the
research, while the right column depicts the steps taken that involve strong par-
ticipatory action in the form of dialogue and awareness raising. The middle col-
umns show the trajectory of the methods.
Initiation Meeting. The first step of the study was an Initiation Meeting (IM,
Figure 2) with the initiators of the research, the NHIS. Here the problem was
presented and the need for exploring clients’ perceptions on illness, health-seeking,
health care and health insurance services stated. Insights in clients’ views and per-
ceptions should help understand their views, behavior and decision-making, and
this was expected to contribute to finding solutions for the problems experienced in
uptake and retention in the scheme. Scientific evidence would support development
of policy and practice.
Stakeholder Design Meeting. A subsequent Stakeholder Design Meeting (SDM,
Figure 2), including clients, healthcare providers and NHIS, was held to analyze
the current situation and explore and specify problems and priorities. The design
meeting allowed the researchers to capture information about other relevant stu-
dies in the field of health insurance that had been or were being conducted.
Stakeholders jointly emphasized that this client-centered study should focus pri-
marily on the client/provider and client/insurer relations, which would make this
study quite unique in Ghana. This consensus created an important basis for build-
ing local ownership. The central position of the client was emphasized repeatedly at
Fenenga et al. 7
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
various SM during the research process. This SDM enabled a relatively quick
survey of the situation, resulting in an outline of research questions and study
design. While recognizing the limitation of this setting to identify individual opin-
ions, the joint situation analysis facilitated determining the relevance and appropri-
ateness of the research questions.
The four moderators who were to carry out the qualitative data collection were
trained in a four-day workshop focusing on participative action approach and
moderation techniques. Topic guides were pre-tested and refined together with
the moderators. For the selection of the participants, we liaised with the regional
health authorities and chose facilities in different areas so that we would collect a
rich variety of responses. Participants were selected in consultation with the local
community worker, using a snowball approach. The interviews were conducted at a
neutral location, such as at clients’ homes or community buildings. All qualitative
data were recorded, transcribed, coded and analyzed using N-Vivo 9.
Individual Health Histories. To acquire insight into the communities’ perspectives, we
employed Individual Health Histories (IHH, Figure 2) to gather in-depth indi-
vidual views and interpretations from literate and (semi) illiterate individuals. In
individual interviews, clients shared their experiences and views on an illness
episode that had taken place within the last six months. Recalling the experiences
step by step gave insight into their beliefs and interpretations of their health
problem, the support structures and subsequent perceived barriers they encoun-
tered when accessing healthcare and insurance services. The following quote from
a participant in the Western Region reflects a view commonly found among
clients:
...It seems that here in the western region we really face problems. When I was in
Kumasi, my wife was sick, she was bleeding and admitted at the hospital. She was
holding her insurance card and they took care of her very well without asking ‘‘kaple’’
(extra money). They gave her medicines until she got well. ......But when we came
here, they told me the insurance did not cover this [health problem]. So next time she is
ill we will not go here. (Male, IHH, Western Region)
The quote provides insight into the participant’s beliefs about illness, which led to
the decision that his wife required professional care from a hospital. It also reveals
perceived experience with the healthcare services as an NHIS enrollee, which
apparently varies per healthcare provider. The finding points to discrepancies
between what the NHIS promises to clients and what clients actually experience
in practice. We subsequently used other methods to explore and validate this
information.
Key Informant Interviews. Key Informant Interviews (KII, Figure 2) were held with
specific knowledgeable community members such as a chief, a fetish priest or spir-
itual healer. These key informants helped to comprehend information and
8Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
contextualize data. In certain areas, for example, we found traditional beliefs to be
strong. Here, herbalists or fetish priests were common or sometimes the preferred
providers to seek care from.
Sometimes one has stomach problems and has visited the hospital three or four times
without improvement. When this happens, I use some herbal medicines to cure them.
Sometimes they bring children with convulsions, unconscious. If I find that he/she ‘‘has
breath in him’’ I take some herbal medicine. I squeeze juice out of it and put this in their
nose, making them sneeze or go to toilet. After finishing the herbal treatment I refer them
to the hospital. (Spiritual healer, KII, Western Region)
This quote demonstrates the common practice of using informal and formal ser-
vices when patients do not find immediate relief. It also indicates the widely used
herbal therapy. The fact that traditional practitioners are not affiliated with the
NHIS may influence the demand for health insurance. Traditional practitioners,
however, do refer patients to western hospitals. Collaboration between traditional
and healthcare authorities varied from area to area and was generally considered
important by the key informants.
Focus group discussions. Focus Group Discussions (FGD, Figure 2) were used to
pursue the question of how individual beliefs and emic concepts identified in the
IHH and KII were shared between community members as common opinions. The
groups of on average nine participants were of various compositions: female
insured; male insured, female uninsured; male uninsured; and mixed groups.
Topics included views and interpretations on health and illness; common support
structures; and experiences with health care and health insurance services, with
specific attention to accessibility, quality, and trust. The FGD generated a broad
range of opinions on the issues mentioned above, facilitated through interaction in
the group.
R1 ‘.... there are different schemes in [name district] and this brings a lot of variations;
even in [name primary hospital], services such as scans and other operations are all
covered by the scheme but here in [other primary hospital], the situation is very
different.’
R2 ‘ah .... Yes, and if you do not have insurance, they give you better and more
medicines than someone who has the insurance card. With the insurance, you’ll get
some para and maybe some other medicine. Then they will say for this and that you
should pay 2 cedis; you have to pay, in addition to your insurance.’ (FGD, female
insured)
The first quote suggests that observed variations in treatment are rooted in the
differences in decision-making between different district offices of the NHIS. The
second quote suggests differences in treatment are based on the insurance status of
Fenenga et al. 9
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
the client: those not insured are treated better than those insured. In another FGD
in the same area, we obtained yet more new information:
R1 ‘On the TV they announced health insurance and we heard that it is good for our
health. But the health insurance does not cover a lot of drugs nowadays, unlike at first.
Some blood tonics and other drugs, which used to be under the scheme, are no longer
covered. These days you have to pay for those.’ (FGD female and male insured)
This quote suggests not only differences per location, but also changes in services
over time by the same healthcare provider: insurance benefits are gradually redu-
cing. This was supported by the next quotes:
R1 ‘There are so many differences. I went to hospital and instead of two drugs I was
given one. I asked them and they told me to go and buy the other. I brought my money
and paid. Just imagine. You can compare ‘‘health’’ [health insurance] to cash and carry.’
R2 ‘Yes ...you see, because there is no money attached to the insurance card, doctors
don’t want to attend to you fast ....... After you have shown your card they will tell you
the medicine you want they don’t have. ...But that same medicine will be given to the one
that has the money. So those having money will be attended faster and better than those
having the insurance card.’ (FGD male insured)
During the same discussion, another man added information regarding the type of
facility and the behavior of certain providers.
R3 ‘Mostly these are the government hospitals. .... when you go there and you are
holding your ‘‘health insurance’’, it is like they would not get any benefit from you to
put in their own pocket ...With the health insurance they are not interested to attend to
you at all.’
The specific sequence and complementarities of the qualitative methods were
intended to show the convergence and dissonances of information. For example,
the FGD in our linked trajectory confirmed and validated the perspectives found in
the IHH on services in the healthcare facilities. We determined common aspects
in the opinions of clients. Informing each subsequent method by the previous step
allows for validation and optimizes the quality of the collected data.
Regional Validation Meeting. Two Regional Validation Meetings (RVM, Figure 2)
were organized to validate and assess the generalizability of findings to a regional
level. Here the three stakeholder groups were invited for a full day. During the first
part of the day, stakeholders were organized into separate groups to validate the
data gathered through qualitative methods. In each RVM, 20 purposively selected
clients, who had taken part in the earlier methods, were invited. By discussing the
findings, we allowed for interaction and trust building within the group of clients
10 Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
(see Appendix 1, Case description). This prepared the clients for the plenary
Stakeholder Meeting (SM) in the afternoon.
Stakeholder Meeting. In the subsequent SM (see Appendix 1, Case description), in
which on average 55 people participated, stakeholder-specific interests and needs
became more apparent, such as the eminent differences in perceptions between
clients and providers on what constitutes quality health care. The sociocultural
dimension of power relations confirmed the assumptions expressed by the generally
weak voice of clients. When clients have queries about the healthcare services and
benefits of the NHIS, they are rarely passed on to the managers of the healthcare
facility or insurance office. Likewise, feedback from clients that could help pro-
viders and insurer to improve procedures or correct any misconduct by staff was
only occasionally given. This was either due to lack of information about where to
file complaints or, more commonly, due to lack of trust and to fear of negative
repercussions.
The openness of discussions (or lack of openness) demonstrated the contextual
differences between the two regions. In Greater Accra, the capital region, the flow
of communication between the stakeholders was much easier than in the predom-
inantly rural Western Region. This reflected the limited level of empowerment
found in the latter. The timely sharing of the preliminary findings of our study
(and the other two sub-studies), among managers and policymakers who were also
participating, facilitated important first steps in awareness as well as action for
change. Table 1 shows the areas defined in the SM to achieve improved, client-
centered services.
The three researchers who moderated the meeting and also made observations
and notes. These were compared, analyzed, and compiled in a report. The reports
were shared with participants and used for development of the intervention.
The SM also demonstrated the complex interdependency in the client-healthcare
provider-health insurance tripod: an insurance membership card has no value if the
NHIS-affiliated healthcare provider offers poor services. The healthcare provider
cannot run his services if claims are not reimbursed in time.
The more open dialogue in the Greater Accra stakeholder meeting resulted in a
response to a client’s question from one of the NHIS directors. The client’s dis-
satisfaction and limited confidence could likely be the result of insufficient infor-
mation provision regarding the policy changes introduced by the NHIS. The
authority’s communications had focused on healthcare providers rather than
their clients, while these changes directly influenced services and benefits for
clients.
Household Survey Figure 2. To obtain knowledge that is generalizable for a larger
population and relevant for policy advice, the qualitative data were subsequently
used to inform the household survey (HHS, Figure 2): the rich qualitative data
ensured survey questions were well grounded in the local context. To illustrate, in
Ghana assembly men play an important role in community activities
Fenenga et al. 11
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
and communication. This information was integrated into the survey section on
community trust and solidarity. For the data collection, 40 surveyors were trained
in four-day workshops. Topic guides and the questionnaire were pretested and
refined together with the research team. For the quantitative data, respondents
were sampled through a multi-stage sampling strategy. First, 64 clusters of health-
care facilities were selected on the basis of their ownership (public/private), location
(rural/urban), and NHIS accreditation quality scores. Subsequently, 30 households
were randomly sampled from within a radius of 10 km around each selected health-
care facility. Using a semi-structured questionnaire, we collected information on
sociodemographics, sociocultural beliefs regarding health and illness, support
structures, employment status, health status, and healthcare utilization,
NHIS enrollment, perceived quality of healthcare and health insurance services
and consumption expenditure patterns and dwelling characteristics. Quantitative
data were cleaned and analyzed using SPSS-20. All data were validated, triangu-
lated and compared with data of the other sub-studies of respectively healthcare
provider and health insurer. This quantitative data enhanced further insight into
the extent of perceived barriers, while surveys among the healthcare providers (a
different sub-study) provided insight into the technical quality levels of these facil-
ities. As an example, we take the information from the SM, which reveals that
many clients are not happy with the quality of services provided by the healthcare
providers.
Table 2 illustrates that clients’ satisfaction with the quality of services of the
healthcare provider is significantly positively associated with enrollment.
Table 1. Areas for improvement.
Stakeholder level Identified areas for improvement in Stakeholder Meeting
Clients Involve community groups, clients, and community leaders in com-
munication, dissemination, and monitoring of health services and
health insurance services
Clients to respect healthcare providers and insurance and under-
stand limitations
The NHIA to consider accreditation of qualifying traditional
practitioners
Healthcare
Provider
Health staff to improve attitude
No discrimination in treatment and waiting time (insured versus non-
insured)
Fair and transparent queue system
Reliable availability of drugs
NHIS Reliable enrollment and renewal system and timely provision of
NHIS-ID cards
Information service package
Regular enforcement NHIS services/regulations and clinical audits
12 Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
Stakeholder Meeting. The findings of the survey were presented and discussed in two
regional SM (Figure 2), in which stakeholders were invited to respond.
Stakeholders elaborated on the convergences and divergences between the quali-
tative and quantitative findings. The qualitative findings generally revealed a more
critical, negative perception of clients regarding service quality than the quantita-
tive data showed. Discussion generated possible answers to these questions and
facilitated learning.
Intervention MyCare. The suggestions for improvement in Table 1 form the basis of
the Intervention MyCare (IMC, Figure 2), which is part of the PAA. By involving
clients in the monitoring services of the healthcare provider and the NHIS, we
anticipate a number of effects such as enhanced inter-stakeholder communication,
empowerment of clients, improved client-centered services, and trust-building. The
development and results of the intervention are presented in two forthcoming
papers (‘Engaging clients in monitoring healthcare and health insurance services’
and ‘Health service quality improvement through systematic client engagement –
evidence from Ghana’).
Conclusion and recommendations
The rapidly growing interest in health insurance in Africa, where Ghana has pion-
eered the introduction of National Health Insurance for the entire population, has
attracted interest in the scientific world. Most studies, however, are socioeconomic
studies, using quantitative methods, sometimes complemented with few individual
interviews. This study introduced a different model. We aimed to obtain a better
understanding of clients’ perceptions regarding illness, health seeking, health care,
Table 2. Results from household survey.
Perceptions of quality of services at nearest accredited health facility
Perceptions Proportion of household heads (agree/are satisfied)
Insured
(%)
Uninsured
(%)
P-value
Respectful treatment from doctor,
medical assistant, nurse
85.5 78.7 0.000
Organized and fair queuing system
at health facility
84.7 74.6 0.000
Availability of all prescribed drugs at
health facility
68.5 57.4 0.000
Equal treatment for insured and
uninsured patients
62.1 51.0 0.000
Note: N¼1803 households.
Source: Project Survey, 2012.
Fenenga et al. 13
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
and health insurance services in their own sociocultural context. The features of
multi-stakeholder involvement in a complex environment and need for social change
required the use of a PAA combined with the application of a linked trajectory of
qualitative and quantitative methods. In this paper, we aim to demonstrate the
value of our research design. We argue that the PAA with the linked trajectory of
methods is an essential approach for researching a complex situation such as the
health insurance context in Ghana, where various stakeholder groups are involved,
with different views, experiences and expectations. Providing evidence for policy
and practice that would contribute to client-centered services requires a deep under-
standing of the perspectives of the community regarding illness, health care, and
health insurance services. Using qualitative methods enables this research process.
The PAA in our study is instrumental to the facilitation of learning and support for
democratic decision-making in the process of developing client-oriented care.
Making clients, together with other stakeholders, part of the research has helped
to foster relationships and find practical solutions that are acceptable and supported
by clients and all stakeholders. Such an approach is not only suitable for the support
of policy changes in Africa, but also in other parts of the world.
In Ghana, as in many other African countries, clients are at the bottom of the
pyramid. They are in a receiving position, with little attention being paid to their
voice. The used approach shows how participatory action can contribute to social
change, empowering clients by letting them be part of the discussions and solu-
tions. As such, they obtain a role in making healthcare providers and the NHIS
accountable for their services. This interaction can contribute to an improved
power balance and more trust within the client–provider–insurer tripod.
The choice of approach and of the specific methods used in the linked trajectory
has suitably informed the research to design the well-founded questions in the
survey. A summary of the value of the methods is provided in Table 3. The left
column shows the participatory methods in which the three stakeholder groups
participated. The second column indicates the methods used to obtain data from
the separate stakeholder groups, while the third and fourth columns present the
findings and learning outcomes.
When looking at our position and actions as researchers, the participatory
approach has continuously required reflection and reaction, in particular during
the SM. This to anticipate on the stakeholders, but also on our own functioning.
While streamlining the three sub-studies was sometimes challenging, also because
the researchers had different approaches of planning activities, the interaction
between the researchers and moderators provided a valuable way to compare and
verify information and provide feedback to each other. The Dutch researcher had a
wide experience in health care in Africa. She did not have the closeness with the
Ghanaian system as the Ghanaian researchers; however, this facilitated an alertness
for new information which the Ghanaian research found sometimes difficult.
While in the client sub-study most interviews were conducted by the moderator,
the researcher drew discussion diagrams, made notes and reflected upon these
together with the moderator after the interview. This was a useful learning process
14 Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
Table 3. Value and conclusion of approach and linked methods.
PAA Method Main findings Learning outcomes
Stakeholder meetings
(IM, SDM, SM 7 & 9):
Platforms for dialogue and
sharing of opinions and
experiences.
Awareness raising and learn-
ing.
Empowerment of clients
Co-creation of new know-
ledge and that lead to social
action for change (changes in
practice and potentially in
policy)
Qualitative methods
Individual health histories and
key informant interviews:
(individual)
Experiences, views
Emic perspectives
Beliefs
High response, incl. from
illiterate
Focus Group Discussions,
Regional Validation
Meetings:
Sharing common views,
beliefs and experiences in
community by larger group
Detection trends
Generalizing findings
(views/opinions) for regio-
nal level
Observed power relations and
strong position of healthcare
providers; regional differ-
ences; difference in percep-
tion (e.g. what does quality of
care mean?), information gaps,
sensitivities
Strong appreciation by stake-
holders for dialogue
Insight in social structures of
support; traditional beliefs still
strong; People often use a
combination of various health
care (traditional/professional/
popular).
Multiple barriers to access
healthcare and insurance
Differences in context
Reflection on individual findings
of IHH and KII, defining com-
monalities and incidental
issues; Action–reaction evok-
ing new insights; RVM
revealed that we captured
correct information; partici-
pation from different districts.
Defined key messages for SM
First SM created platform for
info sharing, dialogue and
stakeholder input
Created ownership, trust and
support for democratic prob-
lem solving
Safe environment crucial for
multi stakeholder dialogues
SM discussion on differences
qualitative/quantitative find-
ings informative
Generated first insight and
enabled checking of appropri-
ateness of topic guides. Emic
insight of perceptions and
processes. Review of topics
Determined common key issues.
Triangulation and validation of
findings; Insight of situation at
regional level; Learning and
awareness raising among par-
ticipants. Empowerment of
present client voice at SM
Quantitative
Survey:
Quantifying information
Generalizing findings for
larger population
Survey questions well founded.
Analysis identified trends and
revealed differences when
comparing with our qualitative
findings, which were again
discussed and clarified in SMs
Essential for generalizing findings
and conclusions for the NHIS
Fenenga et al. 15
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
and offered a rich exchange of observations. The way people employed space for
dialogue, their voice, the intensity of their discussion and expressions reveal a lot
about relationships.
The researcher sincerely appreciated the openness and enthusiasm of clients and
other stakeholders during this study, making it a rich experience. Stakeholders were
appreciative to be part of this participatory process and responded at various
occassions with positive remarks. ‘Through their research we realized the import-
ance of our client’s complaints reaching us so we can meet their expectations and
clarify issues of their concern’ (District Health Insurance Scheme manager,
Western Region). ‘We are glad we could say what really happens. This process
has given us a voice’ (Lady client, Greater Accra).
We have demonstrated that the PAA with a linked trajectory of qualitative and
quantitative research methods is an effective approach both to research a complex
situation such as the position of clients in the health insurance context in Ghana
and to facilitate learning and support for democratic decision-making. While cli-
ents have played a key role in informing and shaping this study, the meetings with
other stakeholders were important to enable collaborative learning and to prepare
practical, acceptable solutions. With the enthusiasm of clients and other stake-
holders during the research process, we co-created new knowledge and a willing-
ness for social change. We claim that our PAA with a linked trajectory of
qualitative and quantitative methods has contributed to new scientific method-
ology knowledge and practice. We propose this as a model within action research
that can be applied in contexts in Africa and around the world.
Acknowledgement
The authors would like to thank Svante Lifvergren for leading the review process of this
article. Should there be any comments/reactions you wish to share, please bring them to the
interactive portion (Reader Responses column) of the website: http://arj.sagepub.com.
Notes
1. Explanatory Models are defined as notions about an episode of sickness and its treatment
that are employed by those engaged in the clinical processes (Kleinman, 1978, p. 88).
2. Social capital is defined as ‘networks of community relationships that facilitate trust and
motivate action’ (Coleman, 1988 p.105).
3. The perspectives of Healthcare providers and NHIS staff were studied by two other
researches, while at stakeholder meetings the three groups were brought together.
References
Asenso-Boadi, F. (2010). Finding sustainable means of financing health care delivery in Ghana.
Germany: Lambert Academic Publisher.
Bailey, A., & Hutter, I. (2008). Qualitative to Quantitative: linked trajectory of method
triangulation in a study on HIV/AIDS, Goa, India. AIDS Care,20, 1119–1124.
Blanchet, N., Fink, J. G., & Osei-Akoto, I. (2012). The effect of Ghana’s National Health
Insurance Scheme on health care utilisation. Ghana Medical Journal,46, 76–84.
16 Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
Bourdieu, P. (1986). The forms of capital. In J. G. Richardson (Ed.), Handbook of theory and
research for the sociology of education. New York, NY: Greenwood.
Brugiavini, A., & Pace, N. (2010). Extending health insurance: effects of the National Health
Insurance Scheme in Ghana. Contribution paper to the European Report on Development
2010, Department of Economics, University of Venice, 1–40.
Calnan, M., & Rowe, R. (2006). Researching trust relations in health care: conceptual
and methodological challenges. Journal of Health Organization and Management,20,
349–358.
Creswell, J. W. (2003). Research design. Qualitative, quantitative and mixed methods
approaches. London, UK: Sage Publications.
Coleman, J. S. (1988). Social capital in the creation of human capital. The American Journal
of Sociology,94, 95–120.
Fenenga, C. J., Boahene, K., Arhinful, D., Rinke de Wit, T., & Hutter, I. (2013). Do pre-
vailing theories sufficiently explain perceptions and health-seeking behavior of
Ghanaians? The International Journal of Health Planning and Management. DOI:
10.1002/hpm.2159: 1–17.
Fukuyama, F. (2000). Social captital and civil society. IMF Working paper, IMF, pp. 1–18.
Grootaert, C., & Bastelaer, T. V. (2002). Understanding and measuring social capital a syn-
thesis of findings and recommendations from the social capital initiative. Forum Series on
the Role of Institutions in promoting Economic Growth (p. 39). USAID-IRIS,
Washington: IRIS.
Hardon, A., Boonmongkon, P., Streefland, P., Tan, M. L., Hongvivatana, T., Geest,
v.d. S. ... Varkevisser, C. (1994). Applied Health Research Manual- Anthropology of
Health and Health Care. Amsterdam.
Hennink, M., Hutter, I., & Bailey, A. (2011). Qualitative research methods. London: Sage.
Jehu-Appiah, C., Aryeetey, G., Agyepong, I., Spaan, E., & Baltussen, R. (2010). Efficiency,
equity and feasibility of strategies to identify the poor: An application to premium
exemptions under National health Insurance in Ghana. Health Policy,95, 166–173.
Jehu-Appiah, C., Aryeetey, G., Agyepong, I., Spaan, E., & Baltussen, R. (2011a).
Household perceptions and their implications for enrolment in the NHIS in Ghana.
Health Policy and Planning, 1–12.
Jehu-Appiah, C., Aryeetey, G., Agyepong, I., Spaan, E., & Baltussen, R. (2011b). Equity
aspects of the National Health Insurance Scheme in Ghana: Who is enrolling, who is not
and why? Social Science & Medicines,72, 157–165.
Kleinman, A. (1978). Concepts and model for the comparison of medical systems as cultural
systems. Social Science & Medicine,12, 85–93.
Koning de, K., & Martin, M. (1996). Participatory research in health; issues and experiences.
Johannesburg, SA: National Progressive Primary Health Care Network.
Maxwell, J. A. (2005). Qualitative research design: an interactive approach. Thousand Oaks,
CA: Sage Publications.
McIntyre, A. (2008). Participatory action research. Thousand Oaks, CA: Sage Publications.
McIntyre, D., Gilson, L., & Mutyambizi, V. (2005). Promoting equitable healthcare finan-
cing in the African context: current challenges and future prospects. Equinet Discussion
Paper. University of Cape Town, University of Witwatersrand and London School of
Hygiene and Tropical Medicine, 27, 1–70.
Mechanic, D. (1998). The functions and limitations of trust in the provision o medical care.
Journal of Health Politics, Policy and Law,23, 661–686.
Fenenga et al. 17
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
Ministry of Health and Ghana Health Service. (2010). Annual report. Ghana: Ministry of
Health.
NHIA. (2010). National Health Insurance Authority annual report. Accra, Ghana: NHIA.
NHIA. (2011). National Health Insurance Authority annual report. Growing healthier and
stronger. Accra, Ghana: NHIA.
North, D. C. (1990). Institutions, institutional change and economic performance. Cambridge:
Cambridge University Press.
Putnam, R. (2001). Social capital. Measurement and consequences. Canadian Journal of
Policy Research,2001, 41–51.
Sarpong, N., Loag, W., Fobil, J., Adu-Sarkodoe, Y., May, J., & Schwartz, N. G. (2010).
National health insurance coverage and socio-economic status in a rural district of
Ghana. Tropical Medicine and International Health,15, 191–197.
Snape, D., & Spencer, L. (2003). The foundation of qualitative research qualitative research
practice: a guide for social science students and researchers (1–23). London: Sage.
Woolcock, M., & Harper, R. (2001). Social capital, a review of the literature. Paper of the
office of the National Statistics (1–45). London, UK: Office of National Statistics.
Appendix 1. Case description (Figure 2, 6.RVM and 7.SM)
To validate the qualitative data from clients, healthcare providers and the health
insurer, two regional meetings were organized in Takoradi and Accra. From each
interviewed stakeholder group, representatives were invited, using a purposively
selection process. This allowed a good mix of representatives (age, gender) and
people who were willingly and vocal to share their views in a larger group.
The validation process (RVM) was conducted in three groups (clients, health-
care providers and health insurance staff). Here participants reflected on the ana-
lyzed data from their respective groups. The session was facilitated by the
researcher and moderators who were known by the participants and cultural-sen-
sitive, understanding, and inviting participants’ views. The data had been anon-
ymized and was presented through oral presentation, supported by posted quotes
and pictures of villages on the wall. Findings were organized by codes defined by
the researcher and moderators during the analysis of the data. The validation
process in both Western Region as Greater Accra evoked lively interaction.
People liked to re-confirm the presented findings with their own experience. On
the question whether any views were missing, participants confirmed that their
perceptions had been covered satisfactory. The atmosphere was relaxed and posi-
tive. We argue that these validation meetings facilitated a process of familiarization
and trust building within the group. Following the validation, the researcher pre-
pared the group for the stakeholder meeting (SM) that would follow that same
afternoon. In the SM, we would combine the three stakeholder groups and present
in plenary the findings of respective groups. We argue that a strong message should
involve the voice of clients themselves, using their own words, their own language.
Hence, the researcher asked participants during the validation meeting their pre-
paredness to participate in the presentation. While the researcher presented the
18 Action Research 0(0)
by guest on January 18, 2015arj.sagepub.comDownloaded from
XML Template (2015) [12.1.2015–3:56pm] [1–19]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/ARJJ/Vol00000/140035/APPFile/SG-
ARJJ140035.3d (ARJ) [PREPRINTER stage]
main findings, clients would add and share some of their experiences during this
plenary presentation. Several clients enthusiastically expressed their willingness to
present their story. The group decided which stories to present.
During the SM (7.SM Figure 2), the three presentations reflected the key find-
ings of the three stakeholders groups. Here, respective groups could listen and learn
what transpired from respective groups on topics such as illness, health seeking,
service quality of healthcare providers and health insurance scheme offices. In the
presentation of clients, the ‘life stories’ in Greater Accra were presented without
any reluctance. In the Western Region, one client found it difficult to talk freely,
likely feeling intimidated by the size and composition of the group. Another client
anticipated on this by encouraging her to present her story as she did during the
validation meeting. Other stakeholders sympathized and patiently waited for the
client to complete her story. Most clients, however, did not feel constrained.
This reveals that we were able to create a safe environment for all people to
talk. We argue that the RVM contributed to build confidence among the group
members. In addition, the careful moderation of the meeting contributed to the
success of these interactions.
Following the presentation, participants were asked to respond and pose ques-
tions. We found people doing so without hesitation, i.e. clients questioned directors
of the NHIS, and NHIS staff questioned healthcare providers. Everyone responded
in pleasant way. In both regions, the atmosphere was good. In the Western Region,
there was a slight tension between some healthcare providers and a participant
from a District Health Insurance Scheme when discussing a certain incident, but
this was solved and the good atmosphere restored.
Thereafter, the plenary group was asked to form three mixed groups to discuss
the findings with each other and come up with suggestions to respond to the
challenges found. The researchers did not join the groups but walked around for
any question while making observations on the group dynamics and discussions.
We found generally all stakeholders participating in the discussion. We argue that
this was due to mixed composition of the groups and the trust developed during the
course of the day.
The group assignment resulted in three posters, each being presented by a group
representative. The assignment guided the participants to explore challenges and
possible solutions at various levels: i.e. client level, healthcare provider level, and
health insurance level. This triggered all participants to think how to stimulate
change in procedures or behavior in each of the stakeholder groups. Results
showed that although the composition of each groups might have differed slightly,
results were quite similar. Striking was the suggestions for more involvement of
clients in the health care services (see Table 1).
Fenenga et al. 19
by guest on January 18, 2015arj.sagepub.comDownloaded from
