The impact of health financing strategies on access to health services in low and middle income countries

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

This review aims to assess the effectiveness of a range of financial mechanisms in improving access to care in low and middle income countries and where possible, health outcomes.

Background

National patterns of health financing depend to differing degrees on resources from governments, social and private insurance schemes, foreign donors, non‐governmental organizations, communities and households. Low and middle income countries (LMIC) tend to rely on a combination of scarce government resources, donor funded projects and typically high levels of household contributions.

As a result of increasing budget deficits throughout the developing world in the 1980s, a decline in quantity and quality of publicly subsidized health services led to higher dependence on patient payments through the introduction of user fees, supported both by UNICEF (through the Bamako Initiative which promoted 'community financing' of primary health care) and the World Bank (World Bank 1987; Griffin 1988; Shaw 1995). In the 1990s issues were soon raised about the negative impacts that user fees could have on equity and access (Creese 1991; Gilson 1995; Gilson 1997). African countries gradually favoured the introduction of a number of financial mechanisms aiming to reduce the risks of 'catastrophic' payments at point of use (Mwabu 1990, World Bank 1997). More recently, attention has been drawn to new innovative schemes to address equity issues. Performance‐based contracts have been introduced to improve both efficiency and equity in health on the supply‐side (Loevinsohn 2005; Jacobs 2006), and demand‐side incentives targeted at poor people are attracting increased attention amongst policy‐makers and donors (Rubio 2003; Rivera 2004).

However, it is recognised that the poorest populations remain excluded from basic health care in a lot of countries. Even services established as highly cost‐effective are failing to reach those in need (Gwatkin 2004), and there is increasing evidence that those with least access are the poorest and most vulnerable groups (Castro‐Leal 2000; Gwatkin 2001). Limited access and low utilization of important basic health services are in turn contributory factors for the persistence of disease and low life expectancy.

Policy makers in low income countries and donor organisations need good quality information on the effects of alternative financing approaches. Systematically reviewing existing literature is an important first step in discovering existing evidence available. However, despite some decades of experience and research on this subject, few reviews have adopted a systematic approach, with the notable exception of Ekman 2004. Yet this review has a precise narrow scope. It encompasses only one type of financial mechanism (community‐based health insurance) and specifically focuses on resource mobilization and financial protection. In contrast, our proposed review will have a broader perspective as far as the interventions are concerned and will concentrate on other outcomes which are related to improving access or equity of access to health services.

Due to increasing recognition of the lack of access of most vulnerable populations to adequate health services, the aim of this review is to gather evidence on the impact of different financial mechanisms on access to health services. We assume that there are two main drivers of such access 1) availability of services geographically and 2) the cost to the household of accessing these services. We are therefore principally interested in evidence on how financing mechanisms can help extend coverage of services or modify economic barriers to access to health care.

Health financing mechanisms are often delineated into the three main functions they are supposed to fulfil : collection of revenues, pooling of funds and purchase of services (Schieber 1997; WHO 2000; Preker 2004). The Figure 1 expands upon these categories to demonstrate financial flows that occur in low income countries between donors, governments, healthcare providers and healthcare consumers. The principal features of financing mechanisms in low income settings are:

Figure 1.

Source: adapted from Dreschler and Jutting (2005) *PHI: Private Health Insurance **CBI: Community‐Based Insurance

• The low tax base which means that government financing is not an important source of funding relative to wealthier countries (Schieber 1997).

• Donor financing, which may go directly to providers, to households or to the government.

• A number of different healthcare providers in the public and private sectors including drug sellers, GPs, NGOs and government clinics and hospitals.

• Increasing interest in contracting with private providers or non‐governmental organizations to scale up service delivery rapidly (Loevinsohn 2005; Palmer 2006).

• The predominance of out of pocket spending by households to finance their healthcare needs. This is in the form of direct payments, payment into an insurance scheme, or by purchase of a 'health card' that gives access to services for a defined period of time.

Each of the three functions of a health financing system identified by WHO are of interest. Besides, some recently introduced schemes that may also increase access but do not lie within these categories will also be included. These original approaches have been introduced to complete 'traditional' mechanisms (seeFigure 1) in order to improve equity) in access to care, for instance by stimulating the demand side with vouchers or conditional cash transfers.

This protocol presents the methodology that will be followed for four related systematic reviews. It is expected that studies of interest for the overarching question which this review seeks to address (the impact on access of financial mechanism) will fall into four sub‐categories. Each of these sub‐categories will be treated as a separate systematic review. They will be:

1. introduction, removal or change in the level of user fees;

2. risk protection mechanisms, including community‐based, social or private health insurance, and pre‐payment schemes;

3. contracting out and other forms of privatization/use of private sector providers to improve access to care (Lagarde 2009a);

4. demand side interventions such as vouchers or conditional cash transfers (Lagarde 2009b).

Aspects of financing shown in the diagram that will not be included in the review are those related to the raising of funds through the national tax system. Disbursement of funds raised through taxation may be included, but only if it is related to one of the interventions highlighted above e.g. conditional cash transfers.

Objectives

This review aims to assess the effectiveness of a range of financial mechanisms in improving access to care in low and middle income countries and where possible, health outcomes.

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials. Interrupted time‐series analyses provided that:

• the point in time when the intervention/change occurred is clearly defined;

• there are at least 3 or more data points before and after the intervention.

Controlled before and after studies provided that:

• pre and post intervention periods for study and control groups are the same;

• the choice of the control site is appropriate, e.g. similar socioeconomic characteristics and/or no major differences in the baseline.

Types of participants

The review will only include studies taking place in low and middle income countries as defined by the World Bank (World Bank 2008).

Units of study will be the populations who would potentially access health services. The population can be either well delineated (e.g. members of a health insurance scheme, pregnant women targeted by a voucher scheme) or more broad (user fees and their concurrent exemption schemes have an impact on the whole population). Issues of interest are the populations' access to health services, their utilization patterns, and possibly their health status. Hence, 'participants' will include users and non‐users of health services, as well as health facilities, where outcome data could have been collected.

Study designs may use facilities or districts as units of allocation e.g. RCT between different districts. We will include studies on all types of providers (private, governmental, NGOs.). We will not limit the scope of our study to a particular levels of healthcare delivery. We will not limit the scope of our study to a particular level of healthcare delivery and all type of health services will be included in our review. Indeed, as displayed by many reports and ambitious objectives set by international organizations, needs of health care definitively exceed what is actually provided to populations in low and middle income countries. It is thus highly relevant to inform key stakeholders of the potential consequences of existing financing mechanisms on any kind of health services and goods, as the most basic ones still need to be provided to the large majority (WHO 1995; United Nations 2000).

Types of interventions

Our review question focuses on the impact of health financing policies on households' access to health services. We will thus be interested in any financing instrument that influences the financial burden borne by individuals, whether it is targeted at particular groups of the population or not. We will therefore concentrate on methods of raising and pooling revenue for health care directly from households such as user fees or prepayment schemes and health insurance funds. We will also focus on ways of using revenues that may impact on access for different socioeconomic groups, such as voucher schemes or conditional cash transfers, or contracting out provision to the private sector. Therefore there will be four different groups of interventions studied.

• User fees which consist of direct payment by the patients at the time of healthcare delivery.

• Pre‐payment schemes, voluntary health insurance (private or community‐based) or social health insurance. These schemes allow patients to access care free at the point of use or for a small co‐payment.

• Patient incentives and allowances such as conditional cash transfers, voucher schemes or fee exemptions.

• Institution or provider incentives such as performance‐based payment.

Types of outcome measures

Primary outcomes will be changes in access to care or healthcare expenditure.

• Only objective measures relating to the final consumption of health services will be taken into consideration. Access to care can be measured by changes in utilization patterns of health facilities or services (immunization coverage, number of visits, rates of hospitalisation, numbers of people having bought an insecticide‐treated net, etc.) and/or equivalent information collected directly from the population through rigorous survey techniques. Information related to distance travelled or travel time will be out of the scope of our review.

• Healthcare expenditure will be considered when it directly reflects direct (and indirect) costs borne by the patient and/or her family.

Secondary outcomes will include equity and patient outcomes.

• Changes in equity of access ‐ increased access for disadvantaged groups or a reduction in gaps in coverage ‐ will also be an important outcome measure. This will require a preliminary analysis and categorisation of the population of interest along a socioeconomic scale. Any methodology (wealth/asset index) will be accepted provided it is rigorous and detailed.

• Changes in health outcomes, measured by morbidity and mortality rates (broken down by age group, sex, etc.) will also be considered.

There will be objective measures of utilization, performance or patient outcomes. Studies based only on measurements of attitudes, beliefs or perceptions will not be included. Analyses reporting secondary outcomes will not be given the same weight as those reporting primary outcomes.

Search methods for identification of studies

A comprehensive search will be performed in order to avoid both selection bias of published articles and language bias of publications. Academic journals (peer‐reviewed) and grey literature (non published/internal or non‐reviewed papers, reports) will both be searched.

The following databases will be searched. Bibliographic databases

• PubMED, EMBASE (Athens), Popline, CAB‐Direct (Global Health), Healthcare Management Information Consortium (HMIC), World Health Organization Library Information System (WHOLIS), African Healthline (bibliographic databases on African health issues), International Bibliography in Social Sciences (Athens).

• The Cochrane Central Register of Controlled Trials (CENTRAL), the Database of Abstracts of Reviews of Effectiveness and the Cochrane EPOC Specialised Register (and the database of studies awaiting assessment) will be reviewed.

Development studies databases

• ELDIS database ‐ database of development references developed by the Institute of Development Studies (IDS).

• British Library of Development Studies (BLDS) ‐ a database on economic and social issues in developing countries.

• ID21 ‐ database on international development research from the UK.

• The Antwerp Institute of Tropical Medicine database.

Economic databases

• Jstor, Inter‐Science (Wiley), ScienceDirect.

• IDEAS (Repec) for economic working papers.

Spanish and South American databases

• LILACS ‐ Latin American and Caribbean health sciences.

• MEDCARIB ‐ Caribbean health sciences literature.

• ADOLEC ‐ literature on adolescence health.

• PAHO ‐ PAHO HQ Library Catalog.

French databases

• FRANCIS ‐ bibliographic database in social sciences.

• BDSP (Banque de Donnees en Sante Publique) ‐ French database on public health literature.

The websites and online resources of UNICEF, USAID and the World Bank. The websites and online resources of a number of consultancy firms and NGOs : Partnerships for Health Reforms, Abt Associates, Management Sciences for Health (MSH), Oxford Policy Management, Save the Children, Oxfam, relevant donors‐funded projects, etc. The websites and online resources (working papers) of university research centres. Search methodology Search strategies for electronic databases will use selected MeSH terms and free text terms relating to printed health financing literature for developing countries. Some pilot searches lead us to use quite general (exploded) MeSH terms in the search strategy, as we realized that a number of relevant articles were indexed with very different MeSH terms. However, to limit the inclusion of irrelevant references, we will exclude a few specific MeSH terms that are irrelevant. The search in PubMed will also be restricted to all the developing countries (see World Bank website above), by selecting all relevant geographical categories as exploded terms.

• Search "Developing countries"[MeSH] OR "Africa"[MeSH] OR "Central America"[MeSH] OR "South America"[MeSH] OR "Latin America"[MeSH] OR "Mexico"[MeSH] OR "Asia"[MeSH] OR "Commonwealth of Independent States"[MeSH] OR "Pacific Islands"[MeSH] OR "Indian Ocean Islands"[MeSH] OR "Europe, Eastern"[MeSH]

• Search ("Economics"[MeSH] OR "Economics"[SH] OR "socioeconomic factors"[MeSH]) AND ("Delivery of health care"[MeSH] OR "health services research"[MeSH] OR "health planning"[MeSH] OR "health services "[MeSH] OR "utilization"[SH])

• Search "Personnel Downsizing"[MeSH] OR "workplace"[MeSH] OR "health planning guidelines"[MeSH] OR "patient freedom of choice laws "[MeSH] OR "preferred provider organizations"[MeSH] OR "provider‐sponsored organizations"[MeSH] OR "emergency Medical Service Communication Systems"[MeSH] OR "Genetic Services"[MeSH] OR "Medical Errors"[MeSH] OR Chemicals and Drugs Category[MAJR] OR "Drug industry"[MAJR] OR "epidemiology"[MAJR] OR "Patents"[MAJR] OR "War"[MAJR] OR Anatomy Category[MAJR] OR "Child Abuse"[MeSH] OR ("Technology and Food and Beverages Category"[MAJR] NOT "food supply"[MeSH])

• Search Practice Guideline[ptyp] OR Letter[ptyp] OR Editorial[ptyp] "Clinical Trial"[ptyp] OR "Clinical Trial, Phase I"[ptyp] OR "Clinical Trial, Phase II"[ptyp] OR "Clinical Trial, Phase III"[ptyp] OR "Clinical Trial, Phase IV"[ptyp]

• Search "Japan"[MeSH] OR "Korea"[MeSH] OR "Taiwan"[MeSH] OR "New Zealand"[MeSH] OR "Singapore"[MeSH] OR "Israel"[MeSH]

• Search "1985"[PDAT] : "2005"[PDAT]

• Search #1 AND #2 AND #3 AND #7 NOT #4 NOT #5 NOT #6

• Search "Fees and charges"[MeSH] OR user fee[TIAB] OR user fees[TIAB] OR social insurance[TIAB] OR health insurance[TIAB] OR community‐based insurance[TIAB] OR prepayment plan[TIAB] OR prepayment plans[TIAB] OR prepayment scheme[TIAB] OR prepayment schemes[TIAB] OR conditional cash transfers[TIAB] OR cost recovery[TIAB] OR prepayment[TIAB]

• Search #1 AND #2 AND #9 AND #7 NOT #4 NOT #5 NOT #6

• Search #10 OR #8

This search strategy will be translated into the other databases using the appropriate controlled vocabulary as applicable.

Checking of references (i) Reference lists of all papers and relevant reviews identified and a search in Web of Science (Science Citation Index and Social Science Citation Index) will be checked for additional references. (ii) Authors of relevant papers will be contacted regarding any further published or unpublished work. (iii) Authors of other reviews in the field of effective professional practice will be contacted regarding relevant studies that they may be aware of.

Data collection and analysis

Screening Both review authors will review the list of references and abstracts and likely relevant studies will be retrieved. The two review authors will then independently assess the short‐listed studies using a checklist of eligibility based on inclusion criteria listed above.

Data abstraction The following elements will be abstracted independently from each study by two review authors: ( 1 ) Study references

• Name of the first author and date of publication

• Date of the study

• Location of the study.

( 2 ) Described intervention(s) and context

• Nature of intervention

• Intervention group

• Control group

• Broader context/reforms in place if mentioned in the article.

( 3 ) Study characteristics and inclusion criteria

• Type of study: ITS, BACS, or RCT

• List of quality criteria by type of study (see below).

( 4 ) Results

• Main outcomes measured

• Effect.

Quality assessment The quality of the selected studies will be assessed independently by the two review authors. The studies will then be divided into three categories according to their risk of bias: A = low risk of bias, unlikely to seriously alter the results; B = moderate risk of bias, raises some doubts about the results; C = high risk of bias, seriously weakens confidence in the results.

Given the specificities of the research area, traditional EPOC criteria will be slightly modified as follows:

• For all types of studies: the reliability of primary outcome measure(s) will be scored DONE if the outcome is obtained from some automated system (e.g. length of hospital stay) or comes from another objective source; NOT CLEAR if reliability is not reported for outcome measures that are obtained by chart extraction or collected by an individual (will be treated as NOT DONE if information cannot be obtained from the authors); and NOT DONE if the primary data is reportedly of a poor quality.

• For Randomised Trials: two new criteria have been added to account for the rigor of the sampling and the statistical analysis, particularly to take into account potential clustering effects. The two criteria on the follow up of professionals and patients are replaced by one single evaluation of potential exclusion bias.

• For CBA studies: the protection against exclusion and selection bias will be scored DONE if outcome measures obtained from the whole population or a representative sample of the population (and the control group) which is studied; NOT CLEAR if not specified in the paper; NOT DONE if outcome measures were not obtained from a representative sample.

• For ITS analyses: the protection against exclusion and selection bias will be scored DONE if outcome measures are obtained from the whole population or a representative sample of the population studied; NOT CLEAR if not specified (will be treated as NOT DONE if information cannot be obtained from the authors); NOT DONE if data set is not drawn from a representative sample.

For CBA and ITS studies, criteria will be divided between 'essential' and 'important'. For CBA study designs, essential criteria will consist of (a) baseline measurement, (b) characteristics for studies using second site as control and (c) protection against selection and exclusion bias. For ITS analyses, essential quality criteria will be (a) protection against secular changes, (b) appropriate econometric models were used to analyse the data, (c) protection against selection bias and (d) reliability of primary outcome measures. The ITS and CBA studies will then be divided into 3 categories (A, B or C) as explained in table 1. RCT studies will be graded 'A' if all criteria are met ('DONE'), 'B' if one or more criteria are partly met ('not clear'), and C if one or more criteria not met ('not done').

Any discrepancies in quality ratings will be resolved by discussion.

Analytical approach We will use a structured, qualitative approach to analyse the results. Only the papers rated A or B will be included in the final synthesis. Each separate review will be completed and then an overview of findings across all reviews also presented.

For each intervention, summary tables of studies will be presented. Each line will summarise types of outcomes (e.g. utilisation, equity) for each study. It will also include key contextual factors as it is likely that contextual factors within and between countries, and implementation processes will have the main modifying effect on the outcomes recorded. We can differentiate between 'individual' contextual factors, related to a group of clusters or individuals (e.g. the distance of communities from a health facility, rural/urban setting, any salient trait of the community, etc.) and some wider contextual factors (economic crisis or political turmoil at the time of the intervention that could affect the results, concomitant health services reform, etc). It is possible to look within the experiment to test for statistical interaction for the first type of contextual factors but the second type will usually require some qualitative statement from the authors. In addition, we intend to document contextual factors that could plausibly have affected the generalisability of the results. The factors that might be particularly relevant are those related to governance and corruption (e.g. possible 'informal' payments), the dominance or otherwise of the state in terms of delivery of health care, external economic shocks or crisis occurring at the time of the study, or a particular epidemic affecting the country.

In addition we will attempt to standardize measures of the magnitude of effect across studies. This may be done by expressing changes from baseline as relative percentage changes. They may also be categorized qualitatively e.g. 'no difference', 'small difference', 'big difference'. For example changes in co‐payment rates or premium rates, or changes in income share dedicated to health, will be reported along with their subsequent changes in utilization or morbidity rates. However, these approaches will require caution as studies of health utilisation and financing in low income countries employ a wide range of outcome measures which it may be misleading to standardize.

For ITS studies, we plan to do further quantitative analysis where it is possible to obtain data from the authors of existing ITS studies. ITS studies that have lacked a proper econometric analysis will be re‐analysed, when the papers provide sufficient data, or if the authors are willing to share their data. Hence, standardized short and long‐term effects will be given by the changes in level and in trends (respectively) of outcome measures before and after the intervention.

Finally, we will summarize findings across the four different sub‐reviews. Reviews will be assessed in terms of the amount of evidence, the quality of evidence and the differences in the magnitude of effect. Contextual factors will also be described between reviews. In particular it will be considered whether differences in contextual factors (such as specificity of settings) are likely to have made the effects from different reviews non‐comparable.

What's new

Date	Event	Description
11 April 2018	Amended	Protocol withdrawn from publication
1 September 2010	Amended	References for the published reviews have been added

History

Protocol first published: Issue 3, 2006

Date	Event	Description
13 November 2008	Amended	Converted to new review format.

Contributions of authors

Mylene Lagarde (ML) led the writing of the protocol, Natasha Palmer (NP) provided comment, feedback, and additional subsequent input.

Sources of support

Internal sources

• Health Economics and Financing Program, LSHTM, UK.

External sources

• The Bill and Melinda Gates Foundation, USA.

Declarations of interest

None known.

Notes

This protocol presented the methodology to be followed for four related systematic reviews on health financing strategies for health services in low and middle income countries. Three of these reviews have now been published as full reviews in CDSR (Lagarde 2009a; Lagarde 2009b; Lagarde 2011). A new protocol will be published for the fourth review on risk protection mechanisms when a review team is identifed. Therefore, this protocol has been withdrawn from the CDSR.

Withdrawn from publication for reasons stated in the review