What criteria guide national entrepreneurs’ policy decisions on user fee removal for maternal health care services? Use of a best–worst scaling choice experiment in West Africa

Abstract

Objective

Several countries in sub-Saharan Africa have implemented policies to remove or reduce user fees. Our aim was to identify criteria guiding such decisions among national policy entrepreneurs, those who link up problem definition, solution development and political processes.

Methods

We administered a best–worst scaling (BWS) experiment to 89 policy entrepreneurs, asking them to identify the most and the least important criteria on a series of predefined sets. Sets were compiled using a Balance Incomplete Block Design which generated random combinations of all 11 criteria included in the experiment. In turn, those had emerged from a prior set of focus group discussions organized among policy entrepreneurs. Ordered logit models were used to investigate the value of single criteria as well as heterogeneity of preferences.

Results

Political commitment was identified as the most important criterion guiding policy decisions on user fee abolition or reduction to the overall sample, but particularly so for more experienced respondents aged over 50 years. International pressure and donor money were identified as least important while equity and institutional capacity were deemed of relatively little importance. Respondents more involved in advising on policy than on formulating policy rated economic issues such as financial sustainability and cost-effectiveness as less important.

Conclusions

It is feasible to apply BWS experiments in low-income countries, although whether the technique can be adjusted to elicit preferences among non-literate respondents in these settings is unclear.

Keywords

best–worst scaling policy making user fees

Introduction

Across West African countries, user fees for antenatal care (ANC) and delivery were introduced in the 1980s but since then they have been recognized as a prominent barrier to the use of care.¹ Skilled attendance at birth is the single most effective intervention to reduce maternal and neonatal mortality.² Hence, agencies across the world, including the United Nations and the African Union, have repeatedly called for the removal of user fees for facility-based delivery.^3,4

Starting in the 1990s, a wave of public policies promoting user fee removal for ANC and delivery emerged across Africa.⁵ In West Africa, such policies, in particular those relevant to delivery, are relatively recent.⁶ These policies represent an efficient means of increasing service utilization but they face barriers to their effective implementation.^1,7,8 The decision to remove user fees may be taken rapidly, leaving little time to plan implementation. Moreover, the decision is often politically driven and taken without adequate consultation with those who will have to translate the policy into everyday practice.^9–11

The published literature has so far produced limited evidence on the factors that inform the design of user fee removal policies. Retrospective qualitative studies are useful for reconstructing the processes that have led to such decisions.¹¹ These studies, however, fail to identify the relative importance of the different criteria affecting decision-making. Moreover, they do not adequately represent the criteria that policy entrepreneurs identify as important conceptually. We adopted Kingdon’s definition of policy entrepreneurs as people who influence the policy making and the policy agenda. According to this view, public policies emerge when policy entrepreneurs seize the opportunity to couple a problem stream with a political stream.¹² Our aim was to assess the value that policy entrepreneurs attribute to different criteria that shape decisions on user fee removal policies for maternal health care.

Methods

Theoretical framework

To investigate the criteria that guide decisions on user fee removal policies for maternal healthcare services, we used best–worst scaling (BWS), a methodological approach rooted in random utility theory.¹³ More specifically, we adopted a variant of BWS that presents respondents with multiple choice sets, each containing a list of items, technically defined as attributes.¹⁴ For each choice set, respondents are asked to indicate the attribute that they value as best and the attribute that they value as worst.¹⁵

BWS rests on the assumption that preferences are ordinal in nature. It models the cognitive process in which respondents repeatedly choose the two issues in varying sets of three of more issues that they feel reflect the largest difference on an underlying continuum of interest. The choice of the two issues that are simultaneously ‘most’ and ‘least’ on each set provides sufficient statistical information to infer a scale. BWS measures all issues on a common scale with one issue serving as the origin. For this property, BWS experiments are recognized to be more appropriate than other choice experiments when researchers are interested in comparing the importance of attributes/issues.¹³

In our study, we investigated what national criteria policy entrepreneurs defined as most and least important in designing user fee removal policies. We restricted our sample to those working at a national level for two reasons. First, from a theoretical point of view, in the settings where we conducted our study, all relevant policy decisions are made at the national level, with little capacity of local policy makers (such as regional and district officers) to influence national directives. Second, from a pragmatic point of view, while we would be interested to elicit preferences of local policy entrepreneurs, we did not have enough resources to include regional and district officers.

Preparatory qualitative phase

We conducted four focus group discussions with policy entrepreneurs to identify the criteria to be included in the BWS experiment. Respondents were identified at relevant regional policy meetings held between May and October 2011. All focus groups were facilitated in French, using a semi-structured interview guide. Respondents were first invited to elicit the criteria which guided decisions on user fee removal and later probed to comment on an additional series of criteria identified by the authors through a literature review. A total of 34 respondents from nine countries took part.

The focus groups were audio-recorded and verbatim transcribed. Three authors independently analysed the transcribed text using a combination of inductive and deductive coding. Triangulation followed by a discussion among all authors led to the identification of the 11 criteria to be included in the BWS experiment. All criteria identified were retained for the design of the BWS experiment, irrespective of the relative value attributed by the respondents. To confirm the relevance and coherence of the 11 criteria, we assessed them against those reported in the literature reflecting relevant process evaluations and policy analyses.^6,8,16 This confirmed that the criteria were exhaustive of the criteria previously identified in the literature.

Quantitative phase

The BWS experiment was conducted using a structured self-administered questionnaire that included four sections (see Supplementary material online). The majority of respondents (65) were identified at relevant meetings of the Community of Practice (CoP) Financial Access in 2012,¹⁷ a network bringing together experts and decision-makers working on health financing from government, cooperation partners and academia.¹⁸ An additional 24 policy entrepreneurs filled in the questionnaire through the online platform of the CoP. The BWS experiment constituted the first section and was introduced by a brief description of the task.

A Balance Incomplete Block Design was used to randomly distribute combinations of the 11 criteria into two comparable and complementary sets.¹⁵ Half the sample was presented with 11 sets of five criteria and half the sample with 11 sets of six criteria. For each set, respondents were asked to identify the criterion which they deemed to be the most important (i.e. best) and the criterion which they deemed to be the least important (i.e. worst) to investigate respondent’s preferences in a positive sense. Respondents were provided with a precise operational definition of each of the 11 criteria.

The second section of the questionnaire asked participants to indicate which of the 11 criteria should be considered and which criteria should not be considered when deciding on user fee removal policies (multiple responses were allowed), so as to investigate policy entrepreneurs’ opinions in a normative framework. The third section asked participants to indicate which of the 11 criteria had actually guided decision-making on user fee removal in their own country. The fourth section collected information on the respondents’ socio-demographic characteristics.

The original questionnaire was developed in English, translated into French and back-translated into English by the authors. To verify that all questions were understandable, both the English and the French versions were piloted on 10–15 policy entrepreneurs known to the authors. Respondents chose the English or the French version of the questionnaire. Responses to hardcopy questionnaires were entered in Excel using double-data entry and consistency checks. Information from questionnaires filled in online was made available in comma separated value (CSV) format.

The analysis was conducted using Stata 12. First, we compiled descriptive summary statistics of both individual priority scores and the information collected in sections two to four of the questionnaire. Individual scores were derived by the analysis of the individual preferences expressed by the respondents on the single choice sets. By design, each criterion appeared a maximum of six times in each block. Preferences were analysed over a cardinal utility scale bound by −6 and +6 so that each respondent’s preference about a given criterion was determined by subtracting the number of times that criterion was chosen as least important (−1) from the number of times it was chosen as most important (+1) over the six (or five) subsets in which it appeared. The individual scores were averaged across criteria to provide a crude aggregate measure of the value assigned to each criterion.

Second, we employed ordered logit model to investigate the relative value of the 11 criteria in the overall sample (Model 1). The standard deviation of the coefficients was estimated by taking into account clustering within a participant (11 observations per participant) and within choice sets (the presence of 6 or 5 criteria out of 11 in each set). The dependent variable (y) in the ordered logit regression model was created by coding the criterion chosen as most important as equal to 3, the criterion chosen as least important as equal to 1 and all remaining criteria (i.e. those neither selected as most important nor as least important) as equal to 2 (intermediate rank). Following an approach previously used by Louviere and Flynn in 2010, the categorical variables representing the 11 principles were dummy coded and inserted as predictors in the regression model (Model 1).¹⁵ Consequently, the sign of the coefficients in the model reflects the change in the probability of a criterion being chosen as more important over the reference criterion, holding all other independent variables constant.

Third, in order to investigate heterogeneity in preferences across different subgroups of respondents, we extended the initial ordered logit model to include a series of interaction terms between a respondent’s characteristics (age, education, role in policy making) and the 11 criteria included in the experiment (Model 2). Wald test was used to exclude insignificant interaction terms from the extended model and to test the final reduced model (Model 2), including only significant interactions, against the restricted model with no interactions (Model 1).

Results

Table 1 reports the 11 criteria, their operational definition and a few relevant citations emerging from the focus groups. Respondents were: 74% male; average age was 44 years; most (68%) were from West Africa and almost all were highly educated.

Table 1.

Criteria identified during the focus group discussions, operational definition, and relevant quotations.

Criteria	Definition	Focus group quotations
International pressure	Pressure from the international community to achieve specific population health targets and/or to abolish/reduce user fees	‘It should be noted that the politicians know the international context has a huge influence on national politics even if it is an indirect influence’
Donor money	Funds are available from bilateral and multilateral partners to support interventions in maternal care and/or to abolish/reduce user fees	‘The financial support from partners highly contributes to decisions on gratuity policy’
Political commitment	National government and/or national politicians are committed to abolish user fees for maternal healthcare services	‘Politicians must judge each action’s usefulness in terms of getting reelected’ ‘Political will is required for improvements in reproductive health’
Financial sustainability	Funds are available, either from internal or external sources, to sustain user fee abolition/reduction in the long term	‘If we fund the policy from non-sustainable resources, it will failat some point. But if we base it on renewable and accessible resources without need for a lot of work, it can be very durable’ “It is necessary that policies are funded for sufficient time to have adequate and long lasting results’.
Equity	Ensuring access to services free of charge based on need and not on ability to pay or geographical location	‘Since we don’t have an unlimited budget we have to provide care where it is most needed by those who have little or no money for healthcare’
Increase in service utilization	Increase service utilization by lifting the financial barrier	‘As services are not being used to their maximum potential, question is will making the service available without cost result in increased utilization?’
Institutional capacity	User fee abolition/reduction can be implemented within the institutional set up of the existing system and foreseen changes in the demand for services can be managed adequately	‘As a first step we must analyze the current status of health facilities in the community: do we have the required institutions, facilities for the public and healthcare providers with the capacity for new projects?’
Quality of care	The quality of the services on offer can be maintained, and/or improved, even after user fees are abolished/reduced	‘We have to evaluate the project cost in terms of personnel and equipment because if abolishing costs leads to an increase in usage, there will be increased investment required in human resources, drugs and equipment’.
Impact on health	Evidence is available to demonstrate the positive impact of abolishing/reducing user fees on population health	‘We face women whose own life and the life of the child are at risk when they give birth. These conditions may play a major rule in changing policy to provide care without user fees’
Cost effectiveness	Evidence is available to demonstrate that the intervention is cost-effective i.e., is good value for money	‘We have to use the minimum resources required to reach the maximum effectiveness. We can benefit from studying what has worked in other places and adopting the most effective and efficient policies and methods’
Burden of disease	Evidence is available on large disease burden induced by the condition associated with the service for which user fees are abolished/reduced	‘We need to know the health problem importance. We should be based on scientific report from different hospitals for the degree and frequency of some diseases and child mortality rates’

Table 2 reports the total scores for ‘most important’, ‘least important’ and individual averages for all 11 criteria. ‘Political commitment’ and ‘Impact on health’ were identified as the most important criteria guiding decisions on user fee removal (average individual score 1.539 (95% CI 1.107; 1.971) and 1.303 (95% CI 0.884; 1.723) respectively), while ‘International pressure’ and ‘Donor money’ were identified as the least important criteria (average individual score −3.236 (95% CI −3.684; −2.788) and 2.303 (95% CI −2.666; −1.194), respectively).

Table 2.

Results of best–worst scaling: descriptive scores (N = 89).

Criterion	Most important	Total (%)	Least important	Total (%)	Average individual score	SD	95% Confidence Interval
International pressure	11	1.1	299	30.5	−3.236	0.225	−3.684	−2.788
Donor money	16	1.6	221	22.6	−2.303	0.182	−2.666	−1.941
Political commitment	167	17.1	30	3.1	1.539	0.217	1.107	1.971
Financial sustainability	138	14.1	40	4.1	1.101	0.231	0.642	1.560
Equity	123	12.6	48	4.9	0.843	0.232	0.382	1.304
Increase in service utilization	102	10.4	48	4.9	0.607	0.181	0.248	0.965
Institutional capacity	61	6.2	84	8.6	−0.258	0.204	−0.665	0.148
Quality of care	63	6.4	39	4.0	0.270	0.162	−0.052	0.592
Impact on health	141	14.4	25	2.6	1.303	0.211	0.884	1.723
Cost effectiveness	56	5.7	104	10.6	−0.539	0.198	−0.933	−0.145
Burden of disease	101	10.3	38	3.9	0.708	0.196	0.319	1.097

The results of Model 1 (Table 3) largely mirror those of the descriptive analysis (Table 2). Using ‘International pressure’ as reference category (the criterion identified as the least important by the descriptive analysis), ‘Political commitment’ and ‘Impact on Health’ were confirmed as most important criteria, followed by ‘Financial sustainability’ and ‘Equity’. ‘Donor money’ was confirmed as least important. All coefficients were positive and statistically significant, confirming all other criteria were rated above ‘International Pressure’.

Table 3.

Clustered-ordered logit models (with and without interactions).

	Model 1: Model no interactions			Model 2: Model with significant interactions
Criteria^a	Coefficients	Robust SE	P > z	Coefficients	Robust SE	P > z
Donor money	0.547	0.144	0.000	0.551	0.145	0.000
Political commitment	3.068	0.151	0.000	2.903	0.168	0.000
Financial sustainability	2.848	0.155	0.000	3.203	0.187	0.000
Equity	2.736	0.173	0.000	3.017	0.201	0.000
Increase in service utilization	2.434	0.150	0.000	2.451	0.151	0.000
Institutional capacity	1.970	0.151	0.000	2.228	0.183	0.000
Quality of care	2.231	0.133	0.000	2.384	0.151	0.000
Impact on health	2.853	0.143	0.000	3.116	0.165	0.000
Cost effectiveness	1.524	0.154	0.000	2.141	0.223	0.000
Burden of disease	2.382	0.139	0.000	2.400	0.140	0.000
Interaction terms
Age > 50 × Political commitment				0.499	0.198	0.012
Doctorate				0.234	0.058	0.000
Doctorate × Equity				− 0.664	0.252	0.008
Doctorate × Institutional capacity				−0.614	0.248	0.013
Doctorate × Quality				−0.343	0.191	0.073
Doctorate × Impact on health				−0.612	0.221	0.006
Doctorate × Cost effectiveness				−0.729	0.228	0.001
Policy_out × Financial sustainbility				− 0.616	0.203	0.002
Policy_out × Cost effectiveness				−0.537	0.211	0.011
/cut1	−0.035	0.090		0.056	0.091
/cut2	4.075	0.110		4.199	0.112

Reference category: international pressure.

The results of Model 2 (Table 3) indicated that ‘Political commitment’ was important to the general sample, but particularly so to respondents above 50. Those more involved in policy advising than formulation (i.e. policy out: academics and partner agencies) rated economic issues such as ‘Financial sustainability’ and ‘Cost-effectiveness’ as less important than those involved in direct policy formulation (policy in: policy technicians at central and local level).

In section 2 of the questionnaire, the vast majority of respondents indicated ‘Political commitment’ (76.7%), ‘Financial sustainability’ (74.2%) and ‘Equity’ (71.9%) as the criteria that should be considered when designing user fee removal policies. Only 20–25% indicated that ‘Donor money’ and ‘International pressure’ should be considered, while over 70% questioned why these two criteria were even suggested. Most respondents (62.9%) reported that ‘Political commitment’ had actually guided decision-making in their country.

Discussion

Main findings

This is the first study undertaken to identify the criteria guiding policy decisions on user fee removal for maternal healthcare services and ascertain their relative value among African policy entrepreneurs. It is also the first study applying the BWS methodology in a low-income country and the only one doing so to examine health policy decision-making. Our respondents consistently recognized in ‘Political commitment’ the most important factor shaping the relevant policy decision and also identified ‘Political commitment’ as the factor actually leading to user fee removal in their own country. Older respondents, i.e. those with more extensive experience in policy making, appeared to be even more aware of the importance of this single factor.

Unexpectedly, respondents attributed little importance to the role of institutional health system capacity. This largely reflects a generalized attitude to consider the implementation of a policy as something that can simply follow from a political decision, in accordance with the theories of rational choice in political sciences.¹⁹ This finding is worrisome, as it is well established that the implementation of such policies frequently faces serious difficulties caused by deficient preparation, lack of funding and a health system that is not always able to respond to the increase in demand induced by user fee removal.^8,20

Finally, ‘Impact on health’ was ranked at the top of the list while ‘Equity’ was surprisingly not ranked as high. The ‘impact on health’ ranking might be due to the fact that the targets set by the Millennium Development Goals (MDG) have driven much of the health policy discourse in the last few years, privileging ‘high impact policies’. Similarly, the relative low importance attributed to ‘Equity’ is well aligned with the experience of policies that operate on an egalitarian principle, targeting vulnerable groups (e.g. women), but not addressing existing inequities within these groups.²¹

Limitations

Due to the limited number of respondents, we must be cautious to claim that the results are generalizable to all policy entrepreneurs working in West Africa. However, the CoP Financial Access, which provided us the setting to conduct this research, brings together policy entrepreneurs from very diverse backgrounds. This diversity emerged clearly in our sample, with over 10 countries being represented and with people of different professions and academic training, to indicate that we largely captured the diversity of the African policy constituency. Together with the consistency the results displayed across alternative analytical models and with the fact that our sample was comparable in size to that of many other choice experiments, we are confident in affirming that, albeit with some caution, our results can be judged to be representative of policy entrepreneurs across West African settings.^22,23

It may be argued that a major limitation of the present study lies in its investigation of stated preferences rather than actual ones. This means that, in our BWS experiment, hypothetical behavior was investigated rather than factual evidence. One of the greatest challenges of all stated preference methods is to establish their external validity. Would policy entrepreneurs have made the same choices if they had been confronted with the same situation in real life? In order to answer this question with certainty, a qualitative study in which policy entrepreneurs are observed while making decisions on user fees removal would be desirable. However, our study aims to overcome this gap – at least partially – by investigating also factual evidence in a retrospective manner. We find great consistency between the two.

Relationship to previous studies

Our study is well aligned with prior empirical evidence indicating that historically political commitment has been at the core of the decision to remove fees for maternal healthcare services in many African settings.^5,8,11

The fact that ‘Donor Money’ and ‘International pressure’ were identified as the least important factors shaping policy decisions is consistent with reality and existing evidence in two ways. First, our data reflected the fact that user fee removal policies in most countries have been taken by politicians without consulting the technical advisors and donors.^5,8,11 Second, our respondents rejected the suggestion that their work is influenced by donors. This finding is consistent with a recent empirical review observing that ‘all reforms benefit from strong ownership at governmental level … [and] the role of donors was quite limited’ and reaffirming that most initiatives were financed directly by national budgets.⁸ For instance, in Burkina Faso, the World Bank played a facilitation role, if nothing else.²⁴ The lack of importance attributed to the influence of donors could indicate that, because they had not taken part in the initial decision-making process, many international aid agencies waited for political authorities to take the lead on such matters. This consistency between what has been observed through retrospective policy analysis and what our respondents identified as important suggests that policies have actually been implemented according to the criteria deemed important on a conceptual level by the policy entrepreneurs themselves. Further qualitative inquiry to be carried out in the respondents’ respective countries could confirm or disclaim this hypothesis.²⁵ The alternative explanation is that the BWS results merely reflect the reality observed over the years and do not represent what is valued as important from a conceptual standpoint.

Implications

This BWS study does not per se aim at providing guidance on policy development. In line with the objective of the BWS technique, this study identifies preferences.¹³ It cannot indicate what elements should or should not be considered when planning user fee removal policies because it reflects the preferences of those very same people developing and planning those policies (rather than for instance, community preferences). Knowledge of policy entrepreneurs’ preferences, however, is an important factor to be kept into consideration by other development partners in the health sector, for instance cooperation agencies and foreign governments. Without knowing and paying due respect to the preferences of the local policy entrepreneurs to understand what elements are valued to be the most important in guiding relevant policy decisions, exogenous policy initiatives, even if very well intended, are likely to fail.²⁶

From a methodological point of view, there is a growing interest in the use of preference elicitation methods to explore decision-making criteria for policy change. For example, in an explorative study, a discrete choice experiment was carried out to determine the relative importance of different criteria in identifying priority interventions in Ghana.²⁷ The results showed that cost-effectiveness was considered to be an important criterion together with reducing poverty, focusing on severe disease, or targeting younger patients. Following this pioneer work, preference elicitation methods have been used to investigate decision-making criteria of policy makers in seven low- and middle-income settings.²⁸ With specific reference to BWS methodology, we find its application in a study aimed to investigate public perceptions and preferences for healthcare reform in Australia.¹⁵

Our study is the first in its kind in sub-Saharan Africa, and it clearly indicates the feasibility of applying BWS experiments even in low income settings, at least among literature populations. Future research may wish to test whether the same technique can be adjusted, possibly through the complementary use of images and vignettes, to elicit preferences on health policies even beyond user fee removal, also among recipient communities. Efforts have been made recently to adjust other preference-eliciting techniques, such as discrete choice experiments, to elicit preferences for health service provision among illiterate populations.^29,30 Given its simpler mode of administration, there is reason to believe that adapting BWS may represent an even more suitable tool to elicit preferences among illiterate populations in low income settings, An even greater gap in knowledge applies to the recipient communities, since very little is known on their preferences for health policies in general and specifically so for exemption policies, both in relation to the services being targeted and to the implementation procedures being adopted.

Footnotes

Acknowledgements

We would like to thank the organizers of the workshop in Bamako as well as all the members of the Community of Practice Financial Access who kindly agreed to participate in this study. The research was approved by the Review Committee of the Financing Harmonization for Health in Africa Communities of Practice. We are grateful for the assistance of Olivier Kalmus, Afua Asante-Poku and Charles Kabore during data entry and piloting.

Funding

This study has been financed by the Africa Initiative at the Centre for International Governance Innovation. Valery Ridde is the recipient of a New Investigator grant from the Canadian Institute of Health Research (CIHR).

Supplementary Material

The online appendix is available at

References

Dzakpasu

Powell-Jackson

Campbell

. Impact of user fees on maternal health service utilization and related health outcomes: a systematic review. Health Policy Plan 2014; 29(2): 137–150.

Campbell

Graham

. Strategies for reducing maternal mortality: getting on with what works. Lancet 2006; 368: 1284–1299.

African U. Decisions, declarations, resolutions. Assembly of the African Union. Fifteenth Ordinary Session 25–27 July 2010, Kampala, Uganda, http://www.au.int/en/sites/default/files/ASSEMBLY_EN_25_27_July_2010_BCP_ASSEMBLY_OF_THE_AFRICAN_UNION_Fifteenth_Ordinary_Session.pdf (2010, accessed 16 May 2013).

The Global Campaign for the Health Millennium Development G. Leading by example – Protecting the most vulnerable during the economic crisis, Oslo: Office of the Prime Minister of Norway, 2009.

Gilson

McIntyre

. Removing user fees for primary care in Africa: the need for careful action. BMJ 2005; 331: 762–765.

Ridde

Morestin

. A scoping review of the literature on the abolition of user fees in health care services in Africa. Health Policy Plan 2011; 26: 1–11.

Lagarde

Palmer

. The impact of user fees on access to health services in low- and middle-income countries. Cochrane Database Syst Rev 2011; 4: Art.No. CD009094–Art.No. CD009094.

Meessen

Hercot

Noirhomme

. Removing user fees in the health sector: a review of policy processes in six sub-Saharan African countries. Health Policy Plan 2011; 26: ii16–ii29.

Agyepong

Nagai

. ‘We charge them; otherwise we cannot run the hospital’ front line workers, clients and health financing policy implementation gaps in Ghana. Health Policy 2011; 99: 226–233.

10.

Walker

Gilson

. ‘We are bitter but we are satisfied': nurses as street-level bureaucrats in South Africa. Social Sci Med 2004; 59: 1251–1261.

11.

Olivier de Sardan

Ridde

. Les contradictions des politiques publiques. Un bilan des mesures d’exemption de paiement des soins au Burkina Faso, au Mali et au Niger. Afrique contemporaine 2012; 243: 13–32.

12.

Kingdon

. Agendas, alternatives and publics policies, New York: Harper Collins, 1995.

13.

Flynn

Louviere

Peters

. Best–worst scaling: What it can do for health care research and how to do it. J Health Econ 2007; 26: 171–189.

14.

Flynn

. Valuing citizen and patient preferences in health: recent developments in three types of best-worst scaling. Expert review of pharmacoeconomics & outcomes research 2010; 10: 259–267.

15.

Louviere

Flynn

. Using best-worst scaling choice experiments to measure public perceptions and preferences for healthcare reform in australia. Patient 2010; 3: 275–283.

16.

Hercot

Meessen

Ridde

. Removing user fees for health services in low-income countries: a multi-country review framework for assessing the process of policy change. Health Policy Plan 2011; 26: ii5–ii15.

17.

Financial Access to Health Services Community of Practice, http://www.hha-online.org/hso/financing/subpillar/financial-access-co (2013, accessed 20 April 2013).

18.

Meessen

Kouanda

Musango

. Communities of practice: the missing link for knowledge management on implementation issues in low-income countries? Tropic Med Int Health 2011; 16: 1007–1014.

19.

Sabatier

. Theories of the policy process, Boulder, CO: Westview Press, 1999.

20.

Ridde

Robert

Meessen

. A literature review of the disruptive effects of user fee exemption policies on health systems. BMC Public Health 2012; 12: 289–289.

21.

Vandemoortele

. The MDG story: intention denied. Develop Change 2011; 42: 1–21.

22.

Mirelman

Mentzakis

Kinter

. Decision-making criteria among national policymakers in five countries: a discrete choice experiment eliciting relative preferences for equity and efficiency. Value Health 2012; 15: 534–539.

23.

Youngkong

Baltussen

Tantivess

. Criteria for priority setting of HIV/AIDS interventions in Thailand: a discrete choice experiment. BMC health services research 2010; 10: 197–197.

24.

Ridde

Richard

Bicaba

. The national subsidy for deliveries and emergency obstetric care in Burkina Faso. Health Policy Plan 2011; Suppl 2: ii30–ii40.

25.

Olivier de Sardan

J-P

. Anthropology and development. Understanding contemporary social change, London: Zed Books, 2005.

26.

Walt

Lush

Ogden

. International organizations in transfert of infectious diseases: iterative loops of adoption, adaptation and marketing. Governance 2004; 17: 189–210.

27.

Baltussen

Stolk

Chisholm

. Towards a multi-criteria approach for priority setting: an application to Ghana. Health Econ 2006; 15: 689–696.

28.

Baltussen

Youngkong

Paolucci

. Multi-criteria decision analysis to prioritize health interventions: Capitalizing on first experiences. Health Policy 2010; 96: 262–264.

29.

Meenakshi

Banerji

Manyong

. Using a discrete choice experiment to elicit the demand for a nutritious food: willingness-to-pay for orange maize in rural Zambia. J Health Econ 2012; 31: 62–71.

30.

Terris-Prestholt

Hanson

Macphail

. How much demand for new HIV prevention technologies can we really expect? Results from a discrete choice experiment in South Africa. PLoS ONE 2013; 8: e83193–e83193.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB