Abstract

There is growing concern that health policies and programmes may be contributing to disparities in health and wealth between and within households in low-income settings. However, there is disagreement concerning which combination of health and non-health sector interventions might best protect the poor. Potentially promising interventions include those that build on the social resources that have been found to be particularly critical for the poor in preventing and coping with illness costs. In this paper we present data on the role of one form of social resource—community-based organizations (CBOs)—in household ability to pay for health care on the Kenyan coast. Data were gathered from a rural and an urban setting using individual interviews (n = 24), focus group discussions (n = 18 in each setting) and cross-sectional surveys (n = 294 rural and n = 576 urban households). We describe the complex hierarchy of CBOs operating at the strategic, intermediate and local level in both settings, and comment on the potential of working through these organizations to reach and protect the poor. We highlight the challenges around several interventions that are of particular international interest at present: community-based health insurance schemes; micro-finance initiatives; and the removal of primary care user fees. We argue the importance of identifying and building upon organizations with a strong trust base in efforts to assist households to meet treatment costs, and emphasize the necessity of reducing the costs of services themselves for the poorest households.

KEY MESSAGES

  • Existing social resources provide a basis for protecting low-income households from the impoverishing effects of illness-related costs. CBOs are one form of social resource that has considerable potential.

  • Working with CBOs to strengthen health care affordability for households has challenges, which include: identifying and building CBOs with a strong internal trust base; and co-ordination and collaboration among CBOs, and between CBOs and other non-governmental organizations and governmental structures.

  • Protecting the poorest households through CBOs has particular challenges. These households are less likely to belong to any CBO, and the CBOs they have formed themselves involve others of a similarly low socio-economic status and are therefore relatively fragile.

Introduction

Protecting the poor through interventions within and beyond the health sector

Economic changes, structural adjustment policies and public sector reforms in developing countries have had a profound impact on health systems, the organization of health service delivery and the health seeking behaviour of communities and households (Bloom et al.2000). Low-income households face growing and competing demands on their resources, and have to make choices between a bewildering array of health providers, many of whom are poorly regulated or supervised (Davies and Sanders 1993; Gilson and Mills 1995; Brugha and Zwi 1998; Bloom et al.2000). There is growing concern that in these settings many health policies and programmes may not only be failing to reach the poorest groups, but might also be contributing to disparities in health and wealth between and within households (Stierle et al.1999; Bloom and Standing 2001). While there is consensus that disadvantaged groups need protection, there is real disagreement concerning which combination of possible health and non-health sector interventions or approaches might be employed.

Health sector interventions

Recognition that many people, including the poor, seek treatment from a range of private for-profit providers, sometimes paying substantially, has been used to promote increasing privatization and private provider interventions, and in defence of public sector user fees (Brugha et al.1999; Brugha and Zwi 1999). However, these options, and their alternatives, remain highly contested, particularly on equity grounds (payment on the basis of ability to pay, and equal opportunity of use for equal need) (Bennett et al.1997; Gilson 1997; Giusti et al.1997). These concerns have led to a shift in interest among donors towards insurance or risk-sharing mechanisms in cost recovery (Steirle et al.1999; World Bank 2000; Preker et al.2002). National tax-based and social insurance schemes have limited potential in settings where a high proportion of the population operate in the informal or agricultural sector, and private or employment-based schemes can lead to the neglect of poorer groups (Ensor 1999). In such settings—where formal sector schemes effectively cover only the relatively wealthy—appropriately designed and managed community-based health insurance schemes (CBHIs) are seen as a means of improving health service access, affordability, quality and community participation. CBHIs can be both insurance (where financial risks are shared among groups of people) and prepayment schemes (where an individual pays in advance for a set of defined, non-transferable health care benefits, such as five outpatient visits).

Interventions beyond the health sector

Studies exploring the impact of cost of illness on households (see, for example, reviews by Russell 2001, 2004; McIntyre et al.2006) highlight that protecting households from illness-related costs requires interventions that build on the resources that are critical to preventing and coping with illness costs. Aside from access to cash and availability of assets that can easily be sold or pawned, access to social resources may be key. Russell's reviews of the data (Russell 1996, 2001) show that social resources enable disadvantaged groups to gain access to the financial, human, physical, social and natural resources of others; that drawing on social resources is one of the most common strategies for minimizing costs and coping with the costs that do arise; and that social network members play an important role in informing each other about and deciding upon which different sources of care to use. Micro-finance initiatives (MFIs) encourage group borrowing and joint liability, and therefore depend on, and arguably even strengthen, social resources. MFIs have gained enormous support from governments, donors and ‘social entrepreneurs’ over the last 30 years (Marr 1999; Martin et al.1999). There is great potential in MFIs protecting against illness-related financial shocks indirectly through, for example, helping the poor to set up or maintain income-earning activities, to improve their homes and environments, to have security against short-term shocks and strengthen long-term welfare (UNFPA 2006). There is now also a growing interest in developing more explicit linkages between MFIs and health insurance and promotion schemes (Interagency Coalition on Aids and Development 1998–2001; Amin et al.2001; UNFPA 2006).

Community-based organizations as social resources

Social resources lie along a spectrum, from visible institutions such as registered community-based organizations (CBOs), through unregistered groups of personal friends, to the least visible familial relations. Given that policies and programmes are increasingly seeking to engage and utilize resources within civil society for health (Loewenson 2003), associations and groups beyond the ties of immediate family, whether registered or unregistered, are of particular interest in their potential to provide protection for households from illness costs. Such organizations are likely to include both those that are intended at impacting on household-level expenditure levels and affordability and those that are not. The successes and challenges that indigenous social arrangements and pre-payment and risk-sharing mechanisms have are likely to offer important lessons for, and potential entry points for working with, civil society to minimize affordability problems in different ways, including through CBHIs and MFIs.

Drawing on Uphoff (1986), organizations may be distinguished by whether they operate at the group, community or locality level. At group level, organizations are self-identified sets of persons with some common interests such as neighbourhood, occupation or gender. At community level, organizations are relatively self-contained socio-economic residential units, and at locality level they are sets of communities having socio-economic relations. Blackmore and Ison (1998) argue that organizations operating at different levels incorporate different boundaries of activities and decision making, with many non-government organizations (NGOs) now preferring to work through locality-level organizations. Drawing on the social capital literature, community organizations can also be categorized according to whether membership suggests relationships that are bonding, bridging or linking (Lyon 2000): These ways of categorizing organizations are complementary. Group- and community-level organizations are more likely to have bonding and bridging relationships, and locality-level organizations more likely to include linking relationships, including to and among larger NGOs.

  • Bonding relationships are strong, and connect people from the same immediate group: family members, neighbours, close friends and business associates sharing similar socio-demographic characteristics.

  • Bridging relationships are weaker, bringing together people from the same socio-economic status but different ethnic, occupational or geographic groups.

  • Linking relationships are vertical ties between people from different socio-economic groups and positions of power, such as the links between poor people and actors in positions of influence in formal organizations, such as political parties, banks, schools, hospitals, housing authorities or the police.

CBOs and household ability to pay for health care

The complex factors and interactions that might be considered when exploring the interface between households, health systems and social resources are illustrated in Figure 1, adapted from Russell (2001). The central focus of the framework is the household, and, more specifically, the interactions between household assets, perceived illness and cost of illness (Box 1 in Figure 1). A household's asset set can influence the amount of illness, perceived cause and severity, treatment-seeking strategy, cost burden and ability to cope with the cost burdens faced. Cost burdens are both direct (fees, drug costs, food costs during hospitalization, transport to reach the facility) and indirect (lost wages or payments made to others to cover for work lost), and can influence the household's asset base. Where costs are high and coping is constrained, households may be unable to pay. Although the proportion of total expenditure spent on health care is often used to indicate ability to pay, the proportions quoted in the literature vary, and the approach itself is contested for over-simplifying complex ground realities (Chuma et al.2006).

Figure 1

Conceptual framework

Figure 1

Conceptual framework

Key influences on the household ability to pay at the local health system level (Box 2, Figure 1) are providers’ locations, transport costs and charging systems (including whether or not there are functioning exemption and waiver systems). Also important at this level are perceptions of the quality of care offered by providers, and therefore whether charges are considered acceptable, and free care valuable. At the community level (Box 3, Figure 1), access to social resources, including the range of community associations, may be a particularly important general asset, especially for the poor. Social resources are also drawn upon by the health system, at least in theory, to increase responsiveness and accountability to community members (Loewenson 2003).

We have not identified any studies looking at how different community organizations impact on household ability to pay for treatment. However, the literature reviewed above, reflected in Figure 1, suggests that group members and their activities have the potential to impact directly or indirectly on household ability to pay for treatment through:

  • Reducing the direct costs of illnesses at the time of illness by meeting treatment or transport costs, with or without expectation of repayment. This should reduce the possibility of potentially damaging cost prevention strategies (for example, delaying or refusing to seek care), cost-minimization strategies (for example, using cheaper sources of care), and cost management strategies (for example, re-prioritizing expenditure towards health care and away from other basic household needs, selling essential assets or incurring debt).

  • Reducing direct costs through advocating for or implementing change at the health system level, in particular around health service proximity to vulnerable households and around charges, waivers and exemption systems.

  • Improving the asset base of households and therefore strengthening household ability to cope with direct and indirect costs, and the consequences, in future.

We suggest that if and how CBOs assist in strengthening the affordability of health care for households will be linked to the level at which organizations operate, the type of relationships within groups and between different organizations (bonding, bridging or linking), and the aims and objectives of the organization. The first bullet point is more easily explored and described than the other two.

CBOs as a social resource in Kenya

In Kenya, a poor economic environment and rapidly increasing population has resulted in 50% of Kenyans currently living in absolute poverty. The majority of Kenyans face not only a deteriorating quality and range of health services, but also a lack of purchasing power to demand health services despite overwhelming health care needs (Oyaya and Rifkin 2003). This suggests social resources will play an important role in preventing and coping with illness costs, as well as other financial crises (Bebbington 1999).

In the mid-1990s there were an estimated 5–17 CBOs in every village, and an estimated 300 000 in rural Kenya (WB/Republic of Kenya 1996). There has been some attempt to distinguish between different types of CBOs (Table 1), but there are clearly enormous overlaps. We have not identified any studies exploring the role of registered and unregistered groups in household ability to pay for health care in Kenya.

Table 1

Community-based organizations in Kenya

Organization Description 
Merry-go-round (MGR) or Rotating Savings Association (ROSCA) • Regular daily, weekly, monthly contribution of an agreed sum, which is given to one or two of the group members each time. Everyone takes a turn to receive the money. Acts as a savings system with no interest. 
 • Allow members to bulk purchase groceries, acquire or replace essential household items, meet predictable lump sum obligations such as school fees and unpredictable ones such as payments for health care, funerals etc. 
Accumulating savings and credit association (ACSA) • As above, but members can put more than the minimum sum in. Some members will borrow and pay an interest. Others will only save. At the end of an agreed period, usually a year, the money is divided, with interest distributed according to whether or not members borrowed. 
Harambee • Harambee, or ‘effort pooling’, was originally promoted by the government across all sectors, as a form of popular philanthropy to raise funds for the development of community assets. 
 • Now mostly used to raise money for weddings, education and funerals for better off groups, acting for them as a form of MGR (entails ‘unbalanced’ reciprocal relationships). 
Co-operatives and self-help groups • Similar to women's groups, but not targeting women specifically. May be externally or locally initiated. 
 • Some of the poor have their own groups focusing on shared labour or shared tools and sometimes include a MGR. 
NGOs and church groups • Types of activities include welfare (health and sanitation, provision of water), assisting with school fees, purchase of books, assisting organizing harambees for school fees, introducing new fuel-efficient stoves. 
Women's groups • More than 23 000 registered in Kenya. Grounded in Harambeee and a social welfare tradition. 
 • Idea is for groups to become economically viable production groups that are able to reach the poorer segments of society. 
 • Men often involved or supporting women's groups. 
 • Two key activities: raising cash to pay for school fees and hospital expenses, and assisting with transport costs for burial. 
 • Some do assist the poor who cannot pay monthly contributions. 
Organization Description 
Merry-go-round (MGR) or Rotating Savings Association (ROSCA) • Regular daily, weekly, monthly contribution of an agreed sum, which is given to one or two of the group members each time. Everyone takes a turn to receive the money. Acts as a savings system with no interest. 
 • Allow members to bulk purchase groceries, acquire or replace essential household items, meet predictable lump sum obligations such as school fees and unpredictable ones such as payments for health care, funerals etc. 
Accumulating savings and credit association (ACSA) • As above, but members can put more than the minimum sum in. Some members will borrow and pay an interest. Others will only save. At the end of an agreed period, usually a year, the money is divided, with interest distributed according to whether or not members borrowed. 
Harambee • Harambee, or ‘effort pooling’, was originally promoted by the government across all sectors, as a form of popular philanthropy to raise funds for the development of community assets. 
 • Now mostly used to raise money for weddings, education and funerals for better off groups, acting for them as a form of MGR (entails ‘unbalanced’ reciprocal relationships). 
Co-operatives and self-help groups • Similar to women's groups, but not targeting women specifically. May be externally or locally initiated. 
 • Some of the poor have their own groups focusing on shared labour or shared tools and sometimes include a MGR. 
NGOs and church groups • Types of activities include welfare (health and sanitation, provision of water), assisting with school fees, purchase of books, assisting organizing harambees for school fees, introducing new fuel-efficient stoves. 
Women's groups • More than 23 000 registered in Kenya. Grounded in Harambeee and a social welfare tradition. 
 • Idea is for groups to become economically viable production groups that are able to reach the poorer segments of society. 
 • Men often involved or supporting women's groups. 
 • Two key activities: raising cash to pay for school fees and hospital expenses, and assisting with transport costs for burial. 
 • Some do assist the poor who cannot pay monthly contributions. 

Source: World Bank/Republic of Kenya (1996).

In this paper we present data from a study that aimed to: 1) describe the range of CBOs in a rural and an urban setting on the Kenyan coast in terms of key functions, membership, cooperation and links with other CBOs; and 2) determine the role of CBOs in reducing costs of illness to low-income households through meeting transport or treatment costs, or through advocating for and implementing change at the health system level. In presenting data on household-level patterns of membership in organizations, and the potential and challenges faced in the range of organizations identified, we contribute to the debates on formal and informal mechanisms that might be strengthened or reconsidered by governments, NGOs or donors to protect groups who are struggling to pay for basic health services.

Methods

This study was conducted between 2003 and 2005 in a rural and an urban area within Kilifi district following national scientific and ethical approval (SSC No. 708). Kilifi district is the second poorest district in Kenya and has the highest female illiteracy rates in the country. Sixty-four per cent of Kilifi residents cannot afford to meet the minimum food requirements even if they were to spend all their income on food alone. Within Kilifi, Ganze is the rural study area, recently identified as the poorest division in the country. Ganze is located 35 km west of Kilifi town on the coast, and more than 4 hours bus ride from Malindi and the city of Mombasa. Mtwapa is the urban study area: a very busy, densely populated low-income area approximately 14 km north of Mombasa.

The work presented in this paper is part of a larger study looking at household ability to pay for health care in both areas; a study which included preliminary individual interviews and focus group discussions (FGDs), two cross-sectional surveys in each area (one in the wet and one in the dry season), and a longitudinal follow-up for 10 months of 30 households in each area. Consent from all people involved in the study was sought prior to interviews. Methods of particular interest for this paper are summarized in Table 2. Household expenditure data from the survey were weighted for household size and age to divide households into wealth quintiles.

Table 2

CBO information gathered through wider study

Methodology Information on community groups 
Preliminary qualitative work • Reasons for and against being involved in CBOs 
 • Role of CBOs in paying for health care 
Cross-sectional surveys:* • Household socio-economic status 
• Wet season: 294 rural and 576 urban households • Household (ex)membership in CBOs and details on groups (e.g. type of group, number and type of other members, main benefit, perceived success) 
• Dry season: 285 rural and 467 urban households • Reported potential sources of cash for various health-related costs 
 • Actual sources of cash to assist in meeting treatment costs 
Longitudinal follow-up of 60 case study households (30 rural and 30 urban) • As with surveys but enabling more detailed information 
 • Reported role of community groups as part of social networks 
Methodology Information on community groups 
Preliminary qualitative work • Reasons for and against being involved in CBOs 
 • Role of CBOs in paying for health care 
Cross-sectional surveys:* • Household socio-economic status 
• Wet season: 294 rural and 576 urban households • Household (ex)membership in CBOs and details on groups (e.g. type of group, number and type of other members, main benefit, perceived success) 
• Dry season: 285 rural and 467 urban households • Reported potential sources of cash for various health-related costs 
 • Actual sources of cash to assist in meeting treatment costs 
Longitudinal follow-up of 60 case study households (30 rural and 30 urban) • As with surveys but enabling more detailed information 
 • Reported role of community groups as part of social networks 

*The same households were visited in both seasons.

Additional interviews carried out specifically for this paper included 12 key informants from a range of CBOs operating within the study areas:1 representatives of two international NGOs, two indigenous NGOs, three health-service-based organizations, two CBOs linked to the Kenyan Government, and two local CBOs. Two informants were chairs of dispensary committees. Besides Village Health Committees, dispensaries are the lowest level structures in place in the health system. These committees are registered as CBOs with Social Services. With locally elected members and responsibility for deciding upon local health service fees and exemptions, they were regularly held up as being excellent opportunities for community participation and public accountability in the health system at the time of planning the study. In-depth interviews were also held with 12 households purposively selected from the 60 longitudinal case study households. Participants represented both study areas and different wealth quintiles, although selection was primarily based on indications of the participants’ knowledge and experience of CBOs through previous contacts. Interviews included households who were members of CBOs, non-members and dropouts.

The concept of data analysis as a continuous and iterative process was followed throughout the research, and was guided by the data analysis framework described by Ritchie et al. (2003). As raw data were collected and synthesized, a conceptual framework, or index, was devised in order to categorize recurring themes and issues. The thematic categories were also influenced by the research aim and objectives. The validity of the data were maximized through the building of long-term relationships between the research team and longitudinal follow-up households, regular meetings among research team members to discuss translation and interpretation of data, and triangulation of data gathered from different participants and using different methods. Validity of the thematic categories was strengthened through the development of separate indices by two researchers, followed by discussion of similarities and differences. Quotes presented in this paper are selected from the range of methods in Table 2 and aim to highlight common views expressed under key themes. Quantitative data were analysed using SPSS, using simple frequencies and chi-squared tests for significant differences between key variables.

Findings

Number and complexity of CBOs

The study found a wide range of CBOs operating in Kilifi District. The overall picture is complicated, but three distinct levels emerged: strategic (international and indigenous NGOs and government departments operating across a relatively large geographical area), intermediate (organizations that implement for and are often set up by strategic organizations) and local (usually small groups set up by individuals and households themselves and operating over a relatively small geographical area). CBOs overall present a pyramidal shape of functions, membership, linkages and tangled networks, with the strategic CBOs at the top and the local CBOs at the bottom. Key characteristics of each level are summarized in Table 3. There are parallels to Uphoff's categorizations (1986), with our local level equating to the group level, and our intermediate level crossing community and locality levels.

Table 3

Characteristics of CBOs at three operational levels

 Strategic level (e.g. KDDP, Plan International, NCCK, KWFT) Intermediate level (e.g. Village Bank, Ganze CBO, dispensary committees) Local level (e.g. women's groups, Farmers’ Field School groups, merry-go-rounds) 
General Includes international and indigenous NGOs, and government departments. Usually established by strategic CBOs to implement or facilitate their functions, but could be an existing CBO. Typically small groups set up by households known to each other. Includes groups set up by strategic or intermediate CBOs to implement activities. 
Membership and  management Comprised of full-time salaried staff. Usually registered as NGOs. Report to board of trustees, donors or government officials. Membership includes smaller CBOs, individuals or shareholders. Officials typically volunteers elected by the local community. Usually registered as CBOs with Social Services, and report to strategic level. Households provide financial contribution, and/or labour voluntarily or for small remuneration. Mixture of registered and unregistered; need to be registered to receive support from other levels. Report to each other, elected leaders, intermediate CBO or occasionally to strategic CBO. 
Functions and  implementation Mainly aim to achieve long-term sustainable improvement for the whole community. Often implement indirectly through intermediate CBOs or establish user groups at the local level. Mainly aim to facilitate initiatives at the local level through community participation for the benefit of the whole community Implement directly or indirectly through local-level CBOs. Mainly aim to benefit individual households (providing cash or other benefits and managing loan repayments). Also include community improvement initiatives. 
Co-operation with  other  organizations Have vertical links through intermediate CBOs or directly to local level. Strong horizontal links with other strategic-level organizations, with exception of government where links can be weak. Have vertical links with strategic level often restricted to one umbrella CBO. Usually, but not necessarily, stronger vertical links with local level. Horizontal links with other intermediate CBOs tend to be weak. May have vertical links with intermediate or directly to strategic level. Few horizontal links with other local-level CBOs. 
 Strategic level (e.g. KDDP, Plan International, NCCK, KWFT) Intermediate level (e.g. Village Bank, Ganze CBO, dispensary committees) Local level (e.g. women's groups, Farmers’ Field School groups, merry-go-rounds) 
General Includes international and indigenous NGOs, and government departments. Usually established by strategic CBOs to implement or facilitate their functions, but could be an existing CBO. Typically small groups set up by households known to each other. Includes groups set up by strategic or intermediate CBOs to implement activities. 
Membership and  management Comprised of full-time salaried staff. Usually registered as NGOs. Report to board of trustees, donors or government officials. Membership includes smaller CBOs, individuals or shareholders. Officials typically volunteers elected by the local community. Usually registered as CBOs with Social Services, and report to strategic level. Households provide financial contribution, and/or labour voluntarily or for small remuneration. Mixture of registered and unregistered; need to be registered to receive support from other levels. Report to each other, elected leaders, intermediate CBO or occasionally to strategic CBO. 
Functions and  implementation Mainly aim to achieve long-term sustainable improvement for the whole community. Often implement indirectly through intermediate CBOs or establish user groups at the local level. Mainly aim to facilitate initiatives at the local level through community participation for the benefit of the whole community Implement directly or indirectly through local-level CBOs. Mainly aim to benefit individual households (providing cash or other benefits and managing loan repayments). Also include community improvement initiatives. 
Co-operation with  other  organizations Have vertical links through intermediate CBOs or directly to local level. Strong horizontal links with other strategic-level organizations, with exception of government where links can be weak. Have vertical links with strategic level often restricted to one umbrella CBO. Usually, but not necessarily, stronger vertical links with local level. Horizontal links with other intermediate CBOs tend to be weak. May have vertical links with intermediate or directly to strategic level. Few horizontal links with other local-level CBOs. 

Source: Key informant interviews.

The number of CBOs operating in Kilifi District is uncertain.2 We identified through snow-balling some 30 CBOs functioning at the strategic or intermediate level. At the local level, the best indication of the number of groups is the proportion of households claiming membership in groups. In our wet season surveys, 37% of rural and 38% of urban households reported having at least one household member in a group; 25% and 26%, respectively, reported member(s) having dropped out within the last 2 years. Within both study areas, the poorest households were less likely to have membership in groups than the least poor (Table 4). The type of group that households belonged to differed by setting (Table 5), but of note is that in both settings very few groups were specifically formed to deal with health expenditure costs. Nevertheless, many groups have the potential to assist with transport and treatment costs as one of many types of financial crisis facing households, and to impact on affordability of treatment indirectly through strengthening aspects of the household's inter-related asset base.

Table 4

Group membership by socio-economic status

Income quintile Rural*n (proportion of all households in quintile) Urban*n (proportion of all households in quintile) 
1 (poorest) 15 (28%) 27 (27%) 
16 (30%) 31 (31%) 
19 (35%) 37 (37%) 
20 (37%) 41 (41%) 
5 (least poor) 27 (50%) 51 (51%) 
Income quintile Rural*n (proportion of all households in quintile) Urban*n (proportion of all households in quintile) 
1 (poorest) 15 (28%) 27 (27%) 
16 (30%) 31 (31%) 
19 (35%) 37 (37%) 
20 (37%) 41 (41%) 
5 (least poor) 27 (50%) 51 (51%) 

*Includes households with complete socio-economic status data with at least one member in a CBO: 110 rural households, 216 urban households.

Table 5

Types of CBOs that households belonged to in each study area

Type of group Number (proportion) of households with at least 1 member (n = 294) Number (proportion) of households with at least 1 member (n = 576) 
Merry-go-rounds 75 (25.6) 114 (19.8) 
Financial or  credit group 41 (14.1) 79 (13.7) 
Employment related 33 (11.1) 50 (8.8) 
Religious 19 (6.5) 32 (5.6) 
Sports/dance 22 (7.5) 23 (4.0) 
Education related 12 (4.1) 5 (0.9) 
Other 92 (31.3) 122 (21.2) 
Type of group Number (proportion) of households with at least 1 member (n = 294) Number (proportion) of households with at least 1 member (n = 576) 
Merry-go-rounds 75 (25.6) 114 (19.8) 
Financial or  credit group 41 (14.1) 79 (13.7) 
Employment related 33 (11.1) 50 (8.8) 
Religious 19 (6.5) 32 (5.6) 
Sports/dance 22 (7.5) 23 (4.0) 
Education related 12 (4.1) 5 (0.9) 
Other 92 (31.3) 122 (21.2) 

Merry-go-rounds (MGRs) were more common in the urban area, and groups set up to implement community-based projects such as piped water and construction of latrines were more common in the rural area. This is likely to be linked to better access to small amounts of money in the urban setting, and NGO activities in the rural area. Group membership is reportedly more common among women than men, possibly because there is greater need, and may be kept secret from husbands:

‘It is easier for women to save because men usually go for drinking sprees when they have extra cash in their pockets.’ (Young man, FGD, 24/01/03)

‘Actually I am not sure if they have started [referring to MGRs] because they [women] can be secretive sometimes … they fear that when you come to know of it, you might borrow their money, drink it and never return it back … And that's why the women can even dig holes in the house [to hide their money].’ (Young man, FGD, 24/01/03)

CBOs’ role in assisting households with meeting health care costs

CBOs interviewed reported a range of activities to assist households directly and indirectly with health care costs (Table 6). Aside from the provision of general emergency loans, of note is that relatively few organizations were involved in assisting households with meeting transport and treatment costs, or directly advocating for or implementing cost-related changes at the health system level. The latter were reported only by the two dispensary committees operating in the study areas at the time.

Table 6

CBO assistance in relation to health care

Direct Indirect 
  • ▪ Emergency loans

  • ▪ User fee exemptions

  • ▪ Lower charges at dispensaries and health facilities

  • ▪ Procure drugs for mission hospitals

 
  • ▪ Reducing poverty

  • ▪ Provide additional services locally

  • ▪ Increase skills of local health staff

  • ▪ Construction of health facilities

  • ▪ Procure equipment

 
Direct Indirect 
  • ▪ Emergency loans

  • ▪ User fee exemptions

  • ▪ Lower charges at dispensaries and health facilities

  • ▪ Procure drugs for mission hospitals

 
  • ▪ Reducing poverty

  • ▪ Provide additional services locally

  • ▪ Increase skills of local health staff

  • ▪ Construction of health facilities

  • ▪ Procure equipment

 

Source: Key informant interviews (n = 12).

Survey questions indicating direct assistance by CBOs to households included whether or not free health care had ever been received at facilities, health insurance coverage, and assistance from CBOs in paying for treatment. The following figures for both the rural and the urban settings support that there is little direct assistance from any level of CBO to households in meeting transport and treatment costs:3

  • 2% of rural and 5% of urban households reported having at least one person covered by health insurance, typically a work-based scheme;

  • 2% of rural and 7% of urban households reported having ever received free treatment at public health facilities;

  • nobody in 614 rural and 505 urban illnesses reported in the 2 weeks preceding the interview said they had used financial assistance from a CBO to meet treatment costs.

In the following sections we draw on qualitative data to discuss this overall finding, firstly for different levels of organizations without a specific remit to impact directly on treatment costs, and then for the special case of dispensary committees.

Strategic-level organizations

There is little evidence that strategic-level CBOs directly affect household ability to pay. This is in part due to the long-term development approach adopted by these organizations, rather than meeting the immediate financial needs of households. MFIs might be considered most likely to have a direct impact through providing emergency or other loans that can be used to meet treatment costs. However, we found no households reporting having benefited as a result of such schemes. In fact, for many poor households, regular re-payments were considered potentially crippling:

Fieldworker case study report: ‘The wife had a loan [from a MFI] but decided to quit due to money problems; she said that the running of a group like that is very tough; that it can make somebody sick due to thoughts about how she is going to be able to pay back the loan.’ (Rural case study HH: RD080)

For lenders, loans to pay for health care were reportedly too risky:

‘Initially there were special loans to cover health care and treatment costs, but this had to be stopped because it was too great a risk. The main purpose of loans is to enable people to make money. If large loans are given for health care costs, there is no certainty of repaying the debt because money has been spent with probably no way of recouping it.’ (Key informant 11)

There may be indirect assistance from these organizations in the future, as they achieve their aims of increasing community mobilization and participation and alleviating poverty. However, they report inter-related challenges. First, there appear fundamental differences between the expectations of strategic-level CBOs and those of households: strategic CBOs’ functions are moulded by the motivations of their donors, with the majority focusing on training and community-based projects for long-term, sustainable socio-economic improvement; households are more concerned with their immediate individual material needs. The poorest households therefore do not join such projects, or drop out, focusing instead on daily survival, particularly in times of hardship (see below).

Second, there seem to be differences between strategic-level organizations regarding how to implement activities, for example regarding allowances for collaborating CBOs to attend meetings, whether community representation is best operationalized at the village or higher levels, how to reach community members, and whether local-level CBOs are expected to contribute in cash or in kind to development initiatives that benefit their communities. This can result in multiple structures and processes being put into place by NGOs and CBOs, often in addition to existing government structures. An outcome is conflict, confusion and replication within the community.

Third, there can be difficulties in ensuring the independence and sustainability of intermediate- or local-level organizations that strategic organizations implement through, because of the challenges faced by those organizations (see below).

Intermediate-level organizations

Intermediate-level organizations, depending on size and on how they were established, share the challenges of both other levels of organization, and face additional inter-related problems around dependency, representation, remuneration and sustainability. As strategic-level key informants explained:

‘It can easily happen that the CBO becomes dependent on us … It can be difficult for the CBO to understand the concept [of sustainable development]. They are used to relying on handouts—a precedent that has been set. There is a lack of belief in their own ability or they say they are too poor and cannot do anything.’ (Key informant 09)

‘For example, maybe the [CBO] constitution says there should be re-election of the chairperson every 1–2 years but the chairman and/or other members stay for five years, we have no power to force them to change.’ (Key informant 01)

Heavy dependence among implementing CBOs on voluntary participation, with little or no remuneration, and the substantial level of responsibility attached to some voluntary posts, is a particular concern. For households faced with daily hardship and regular financial crises, contributing time on a voluntary basis to community projects constitutes yet another burden, and may be unmanageable or lead to questionable alternative strategies for financial support:

‘The use of volunteers is a tricky issue. A volunteer has to be someone who is not employed full-time so he/she has time to do one or two things. Volunteers need to be reimbursed travel costs. If not remunerated, it becomes tricky to sustain volunteers in the community.’ (Key informant 05)

‘Selfish leaders can mess up funds. It's mainly the better educated who can take advantage of the poor who are uneducated. They are not transparent with funds.’ (Key informant 07)

Local-level organizations

It is the local-level CBOs, set up by households themselves, which appear to have the greatest potential to improve household ability to pay for expenses, including health care. For reasons of trust, kinship and convenience, these organizations are typically comprised of relatives, friends and neighbours. They almost always include a MGR. MGRs are popular for both assisting with savings and acting as part of a safety net to allow households access to essential day-to-day items, and to meet unexpected expenses (including treatment costs). In some cases they also allow members to take the proceeds before their turn:

Fieldworker report on one household member: ‘In one of the MGRs that she has been in, they even had a constitution rule stating that severe illness cases, and death cases of members, will be treated specially, in that irrespective of whether it's the affected member's turn or not, one will get money to cater for the relevant expenses.’ (Urban case study HH: UV005)

‘… when your child is taken ill suddenly and maybe it is another person to be paid that time, you go to the treasurer and explain the situation to her, and then you will go to the person to be paid and explain to her your situation. She might give you her number and you will be given the money to take your child to the hospital.’ (Young woman, FGD, 15/02/03)

While the potential of such groups is clear, lack of money and low levels of trust often undermine the success of groups (Table 7). Numerous people report having personally lost money because of fellow members being unable to pay their contribution or because of corrupt group leaders. For some this is the main reason for never joining:

‘The leaders were coming for credit from this money without the other members being informed; especially the chairman was the one who usually borrowed this money for his own use.’ (In-depth interview, rural HH, RC019)

‘People would contribute and give their money to someone and when it comes to his turn he says he has a problem in getting money to contribute. Then it becomes a problem for such a group to operate … Why didn’t he say he had a problem before he got his share?’ (Young man, FGD, 29/01/03)

‘There are others who will be given money and they fail to pay the others … Nothing is done to such people because they have no money. You cannot make them pay by force.’ (Young woman, FGD, 31/01/03)

Table 7

Why members drop out, and why CBOs collapse in rank order (high = 1)

Why members drop out Why CBOs collapse 
  • • Cannot maintain contributions/repay loan instalments (1)

  • • Untrustworthy officials (2)

  • • Delays in receiving payment (3)

  • • Illness (4)

  • • Lack of permission from head of household (5)

 
  • • Untrustworthy officials (1)

  • • Cannot maintain contributions (2)

  • • Lack of external funding (3)

  • • Arguments between members (4)

  • • Expectations not met (5)

  • • Member leaves after receiving payment, i.e. MGR (6)td/

 
Why members drop out Why CBOs collapse 
  • • Cannot maintain contributions/repay loan instalments (1)

  • • Untrustworthy officials (2)

  • • Delays in receiving payment (3)

  • • Illness (4)

  • • Lack of permission from head of household (5)

 
  • • Untrustworthy officials (1)

  • • Cannot maintain contributions (2)

  • • Lack of external funding (3)

  • • Arguments between members (4)

  • • Expectations not met (5)

  • • Member leaves after receiving payment, i.e. MGR (6)td/

 

Source: In-depth interviews (n = 12).

The finding that membership of MGRs is more common in the urban area suggests that better employment opportunities, providing more disposable income, enables these households to better maintain CBO membership contributions. Rural households are experiencing hunger and having to under-go strategies to cope with basic survival, undermining any efforts at even small-scale saving:

Fieldworker report from in-depth interview: ‘… they haven’t registered the group yet due to the hunger [famine] which these people are facing. They cannot contribute the amount they are required to contribute, and what they have already gathered is having to be given away as credit to assist with urgent problems like food.’ (Rural household: RC019/3)

For some of the above reasons, local-level CBOs serve as only one of a variety of sources of informal credit for households. Other sources include shopkeepers, private health care providers and borrowing from other members of their social network. The latter option is attractive in comparison to a (formal) credit-giving facility because ‘friends can give financial assistance on the spot, depending of the availability of the required amount of money’. It is therefore an extremely common strategy used (see also Chuma et al.2006). Nevertheless, even borrowing from friends will be avoided where possible because failure to repay strains social relationships and future ability to deal with crises:

‘If you fail to pay back in time, in the village she will start talking behind your back that she gave you credit three months ago and you have not paid back …’ (Young woman, FGD, 29/01/03)

Credit from shops or health facilities may therefore be preferred. However, not everybody has easy access to credit from friends, neighbours, shops or facilities. Being able to borrow or receive treatment on credit depends largely on trustworthiness, which is linked in large measure to ability to pay back the debt. The poorest experience difficulties in trying to access any form of credit because of their lack of resources to repay debts. As their poverty increases, their ability to draw on the range of strategies employed by community members to pay for health care decreases.

The special case of dispensary committees

A set of CBOs that the poorest could in theory turn to for assistance in payment for health care are dispensary committees, considered intermediate CBOs in this study. Consisting of elected representatives and with their own bank accounts, committees until recently kept 75% of revenue raised through user fees at their facility for re-investment into the dispensary. Health sector staff and committee members reported that this money was used to employ cleaners and watchmen, to carry out basic maintenance, to pay small incentives to committee members such as a sitting allowance, and to buy some drugs. Committee members were also able to make decisions and act ‘on the spot’, including with regard to costs of services and who may be exempted. Nevertheless, and as noted above, relatively few people in the survey reported receiving free treatment at facilities. Offering exemptions and waivers in a setting where so many are poor, and where the money raised can impact on the quality of care provided, is clearly challenging. It also influences the ability to keep charges at affordable levels (Chuma et al.2006).

Mid way through our study the Kenyan government announced that it was abolishing user fees for primary health care. Patients aged over 5 years visiting dispensaries and health facilities would still pay a consultation fee (10 KShs and 20 KShs, respectively) but the move was largely referred to by the government as towards ‘free’ health services. Although potentially exciting in terms of protecting the poorest groups from being unable to pay for basic health care, this change was described negatively by the majority. Public expectations were raised and dispensaries were reportedly flooded with people seeking free care, who were only disappointed to note long queues, staff shortages and inadequate drug supplies. Drugs and other essential resources previously purchased by dispensary committees from user fees were reportedly not being supplied. Whether or not this was indeed the case was beyond the scope of this study, but certainly perceived resource constraints and a lack of clarity around the consultation fee appeared to be undermining the relationship between committee members and communities:

‘Dressing kits, syringes and drugs are in short supply. The dispensaries don’t have cotton wool, bandages. How do you work?’ (Key informant 06)

‘The introduction of free services has resulted in a very small range of drugs being available. If patients want/need other drugs they have to purchase them from the existing drug stock purchased by the committee with user fees before the changes. The committee cannot afford to give these drugs free because they have to account for the money and will have no other funds to replace the drugs. People have been heard in all corners saying that the dispensary has no drugs and the committee is doing nothing. So the committee has had a rough time whenever they come into the streets, being backbitten.’ (Key informant 02)

‘The public doesn’t understand that 10/= is the registration fee and that it doesn’t include drugs. They link the fee to drugs. So if they pay and get no drugs it's a big problem.’ (Key informant 03)

‘The community now say “the government is taking care of our health …” community expectations are high and they believe that “there's money to be eaten” ’ [hinting at untrustworthy behaviour by committee members] (Key informant 05)

Discussion

We have described a complex hierarchy of CBOs operating at the strategic, intermediate and local level and have presented an overview of their role in helping households meet their direct treatment costs. Here we make several concluding points, and links to the wider literature, regarding CBOs’ role in households’ ability to meet treatment costs at each level, and the potential of CBHI.

The way in which the strategic, umbrella-type organizations implement their activities through partner organizations appears confusing, and is often in conflict with the approach of other, similar NGOs. This situation leads to replication, conflict and confusion amongst communities. For example, if some organizations establish committees at the village level and others at the location level, and there are poorly defined relationships between the two, parallel systems emerge with community members aligning themselves with different organizations. These problems are exacerbated where one organization appears more generous than another in, for example, providing aid with minimal resource requirements from community members, giving a greater range of and higher allowances to committee members, and other perks such as holding training programmes in relatively expensive hotels. Such differences have implications for the support and sustainability of activities, and ultimately for the ability of these organizations to build up the resources that are critical for households to prevent and cope with illness costs. It suggests that greater cohesion between strategic CBOs would improve their potential to directly, and especially indirectly, impact on household ability to pay for health care. Recognition of the importance of greater co-ordination and collaboration in the health sector in Kenya at the district level is reflected in the introduction of District Health Stakeholder Forums (DHSFs) (Oyaya and Rifkin 2003). Our findings would support such efforts. MFIs are the strategic CBOs that offer greatest potential to bridge the gap between the long-term aims of strategic CBOs and the short-term needs of households. However, we would support the concerns of others regarding MFIs that perceived uses of the credit can sometimes be illusionary, that repayment of loans may be putting significant stresses on some individuals and households, and that these initiatives may not be reaching those most in need (Buckley 1997; Rahman 1999; Pankhurst 2002).

Intermediate- and local-level CBOs are in theory better placed to fulfil household expectations. However, links between intermediate CBOs and other organizations operating at the same level or at the strategic level are generally weak. In their present form, these CBOs appear to lack the capacity to sustain community participation at the local level, and have little clout in terms of being able to demand accountability from political and administrative officials who should be responsive to local needs and priorities. There appears to be enormous potential for CBOs established by strategic organizations at the intermediate level to expand their role, in particular through greater co-ordination at the strategic level (as described above), and through moving away from a tendency for each new initiative to establish its’ own new set of intermediate CBOs. Nevertheless, perhaps the most important constraint in the broader potential of these groups is their existence in a political system in which there has been a history of ‘ethnic rivalries, patronage and corruption …[being]… the tools used to advance the economic and political interests of the small groups controlling land, business and capital’ (Rakodi et al.2000).

Local-level CBOs are another potential source of direct and indirect support in paying for health care, and working through established groups appears a good way of reaching many community members. However, as noted by others (Feldman 1983; Masinde 1987; Casson and Giusta 2004), dishonesty among leaders and fellow members is a major problem. Identifying groups with a strong trust base appears key to the success of interventions involving these organizations: group members have to trust that others will fulfil their obligations and discuss difficulties where they are unable to; they also have to trust in the motives and activities of those leading and accounting for the group. Developing simple criteria to assess trustworthiness and sustainability may assist in identifying groups with a strong trust base. Drawing on Goodge and Gilson (2005), such assessments might include indicators of shared norms such as respect, dignity, fairness, openness, compromise and solidarity, and evidence around cooperative activities and handling of conflicts. This process may be assisted by more coordinated support from intermediate- and strategic-level CBOs.

Another issue is that groups aimed at maximizing trust tend to include only friends, neighbours and relatives. They are therefore increasingly reliant on bonding forms of social capital (Putnam 1993). This type of particularized trust precludes the mixing of households by socio-economic status and the cross-subsidy of poorer households by the better off (Putnam 2002). Finally, it appears that the poorest households and individuals are least likely to be reached through existing groups. Thus working only through existing CBOs may risk widening gaps between less poor and poorest, undermining existing bonding ties to create ‘antisocial’ capital (Campbell et al.2002). Although creating groups or networks through outside intervention is a possibility, we would support Lyon's (2000) view that this approach would be limited because successful groups tend to be embedded in social relations that take time to develop.

Several recent reviews have examined the growing experience and documentation of the diverse range of insurance schemes that either focus exclusively on, or include as part of their membership, persons outside formal sector employment (Bennett et al.1998; Musau 1999; Arhin-Tenkorang 2001; Ekman 2004). Among other concerns, such as low coverage levels and dependence on skilled external technical support, they have highlighted problems of affordability to the poorest groups. They have shown that the tendency to operate flat-rate premiums, with no sliding scale, not to allow payment in kind, to require substantial co-payments, and not to operate an exemption/waiver policy, all operate counter to affordability. Thus many of those in greatest need of increased economic access are not reached. Our research would strongly support this problem for the Kenyan coast.

For those individuals who can afford to enrol in CBHI and pay a premium, our finding regarding low levels of trust in fellow group members, and in leaders, would suggest developing structures and practices that build and maintain trust in CBHI management will be critical. As argued by Schneider (2005), consumers who pay premiums need to know that they will receive quality care in the case of illness, and that they will not lose money in default schemes. This requires trust in the management team's accounting systems, in their selection of contracted providers, and in their monitoring of these providers’ technical and interpersonal expertise. Building and maintaining such trust not only requires appropriate mechanisms to be in place, but also awareness of such mechanisms and the promotion of CBHI in the broader general community. Schneider (2005) suggests that a successful CBHI may actually foster trust in the community and contribute to wider benefits through strengthening other CBOs and social capital more generally. While this is a particularly attractive idea, it would be dependent on many of the other challenges described above being overcome.

Reducing the direct costs of services clearly remains an absolute priority for the poorest and most vulnerable households who have a limited range of social support strategies available to assist in meeting health care costs. This study, however, supports Gilson and McIntyre's (2005) argument for the need for careful implementation of this essential policy change. The sudden introduction of ‘free’ primary health care, where this is not accompanied by an increase in government provision of resources and careful information-giving to the public, can lead to existing quality of care and relationships within the community being undermined. Furthermore, depending on community efforts and participation to fill in the short-falls is not only likely to be unrealistic (Cornwall et al.2000), but it may undermine existing initiatives and relationships. Now that the policy has been better established, further research around if and how the removal of user fees has protected the poor—and implications for further actions—is urgently needed.

Conclusion

In this paper we have explored the role of one form of social resource—community-based organizations—in household ability to pay for health care. There are several limitations to the study. The first is the possibility of under-reporting of use of CBOs through survey respondents. The second is the relative difficulty of exploring the less direct impacts of CBOs on affordability to households through, for example, advocating for change at the health system level and improving households’ asset bases. Future studies will tackle these limitations. Despite these limitations we have highlighted the potential for CBOs to impact both directly and indirectly on household affordability of health care, and have drawn out some challenges for several interventions of international interest and importance at present: community-based health insurance schemes; micro-finance initiatives; and the removal of primary care user fees. Cross-cutting issues include an apparent need for greater co-ordination and collaboration between organizations at strategic, intermediate and local levels, difficulties in establishing and maintaining strong trust bases in organizations, and reaching the poorest households through existing groups.

Acknowledgements

This research was funded by The Wellcome Trust, UK and supported by the Kenyan Medical Research Institute. We wish to thank the fieldworkers who played a central role in organizing and holding numerous discussions, in particular Jane Kahindi, Isaac Kalume, Wilfred Katana, Johnson Masha and Gladys Sanga. We are particularly grateful to the organization representatives and community members who gave up their time to share their experiences and answer questions. This paper is published with the permission of the director of KEMRI.

Endnotes

1
Organizations included were Plan International, Kilifi District Development Programme, Amkeni, the National Council of Churches of Kenya, two dispensary committees, the Kenya Women's Finance Trust (a MFI), the Farmers’ Field School, ‘The Ganze CBO’ and a Village Bank.
2
In theory, all CBOs should be registered with the Social Services Department, but in practice, resource constraints mean records are incomplete and out of date. Social Service records also do not include non-registered groups.
3
A limitation regarding data on financial assistance from a CBO to pay for illness is that one of the most common treatment payment strategies was gifts and loans from friends (see also Chuma et al.2006). It was beyond the scope of the study to explore in each case whether these friends were fellow members of a CBO, and if so whether the money was a personal loan or a loan from the group. Nevertheless, the qualitative data suggest that gifts or loans from the latter were rare.

References

Amin
R
St. Pierre
M
Ahmed
A
Haq
R
Integration of an essential services package (ESP) in child and reproductive health and family planning with a micro-credit program for poor women: experience from a pilot project in rural Bangladesh
World Development
 , 
2001
, vol. 
29
 (pg. 
1611
-
21
)
Arhin-Tenkorang
D
Mobilising resources for health: the case of user fees revisited
2001
 
Commission on Macroeconomics and Health, Working Paper series, November 2001, Paper No. WG3:6
Bebbington
A
Capitals and capabilities: a framework for analyzing peasant viability, rural livelihoods and poverty
World Development
 , 
1999
, vol. 
27
 (pg. 
2021
-
44
)
Bennett
S
McPake
B
Mills
A
Private health providers in developing countries: serving the public interest?.
 , 
1997
London
Zed Books
Bennett
S
Creese
A
Monash
R
Health Insurance Schemes for People Outside Formal Sector Employment.
 , 
1998
 
Current Concerns, ARA Paper No. 16, Geneva: World Health Organization
Blackmore
C
Ison
R
Thomas
A
Chataway
J
Wuyts
M
Boundaries for thinking and action
Finding out fast: investigative skills for policy and development.
 , 
1998
London
Sage/The Open University
 
chapter 2
Bloom
G
Lucas
H
Edun
A
, et al.  . 
Health and Poverty in sub-Saharan Africa.
 , 
2000
 
IDS Working Paper 103. Brighton: Institute of Development Studies
Bloom
G
Standing
H
Pluralism and marketisation in the health sector: meeting health needs in contexts of social change in low and middle-income countries.
 , 
2001
 
IDS Working Paper 136. Brighton: Institute of Development Studies
Brugha
R
Zwi
AB
Improving the quality of private sector delivery of public health services: challenges and strategies
Health Policy and Planning
 , 
1998
, vol. 
13
 (pg. 
107
-
20
)
Brugha
R
Zwi
AB
Sexually transmitted disease control in developing countries: the challenge of involving the private sector [editorial]
Sexually Transmitted Infections
 , 
1999
, vol. 
75
 (pg. 
283
-
5
)
Brugha
R
Chandramohan
D
Zwi
AB
Viewpoint: Management of malaria – working with the private sector
Tropical Medicine and International Health
 , 
1999
, vol. 
4
 (pg. 
402
-
6
)
Buckley
G
Microfinance in Africa: is it either the problem or the solution?
World Development
 , 
1997
, vol. 
25
 (pg. 
1081
-
93
)
Campbell
C
Williams
B
Gilgen
D
Is social capital a useful conceptual tool for exploring community level influence on HIV infection? An exploratory case study from South Africa
Aids Care
 , 
2002
, vol. 
14
 (pg. 
41
-
54
)
Casson
MC
Giusta
MD
The costly business of trust
Development Policy Review
 , 
2004
, vol. 
22
 (pg. 
321
-
42
)
Chuma
J
Gilson
L
Molyneux
CS
Inequities in health care seeking, cost burdens and coping strategies among low-income rural and urban households in coastal Kenya
2006
 
Submitted manuscript
Cornwall
A
Lucas
H
Pasteur
K
Introduction: Accountability through participation: Developing workable partnership models in the health sector
IDS Bulletin
 , 
2000
, vol. 
31
 (pg. 
1
-
13
)
Davies
R
Sanders
D
Economic strategies, adjustment and health policy: issues in sub-Saharan Africa for the 1990s
Transformation
 , 
1993
, vol. 
21
 (pg. 
78
-
93
)
Ekman
B
Community-based health insurance in low-income countries: a systematic review of the evidence
Health Policy and Planning
 , 
2004
, vol. 
19
 (pg. 
249
-
70
)
Ensor
T
Developing health insurance in transitional Asia
Social Science and Medicine
 , 
1999
, vol. 
48
 (pg. 
871
-
9
)
Feldman
R
Women's groups and women's subordination: an analysis of policies towards rural women in Kenya
Review of African Political Economy
 , 
1983
, vol. 
10
 (pg. 
67
-
85
)
Gilson
L
The lessons of user fee experience in Africa
Health Policy and Planning
 , 
1997
, vol. 
12
 (pg. 
273
-
85
)
Gilson
L
McIntyre
D
Removing user fees for primary care in Africa: the need for careful action
British Medical Journal
 , 
2005
, vol. 
331
 (pg. 
762
-
5
)
Gilson
L
Mills
A
Health sector reforms in sub-Saharan Africa: lessons of the last 10 years
Health Policy
 , 
1995
, vol. 
32
 (pg. 
215
-
43
)
Giusti
D
Criel
B
De Bethune
X
Viewpoint: Public versus private health care delivery: beyond the slogans
Health Policy and Planning
 , 
1997
, vol. 
12
 (pg. 
193
-
8
)
Goodge
J
Gilson
L
How can trust be investigated? Drawing lessons from past experience
Social Science and Medicine
 , 
2005
, vol. 
61
 (pg. 
1385
-
95
)
Interagency Coalition on AIDS and Development
Factsheet: HIV/AIDS and Microfinance
1998
 
Loewenson
R
Overview of issues from the bibliography on civil society and health
2003
Geneva
World Health Organization Training and Research Support Centre
 
Lyon
F
Trust, networks and norms: the creation of social capital in agricultural economies in Ghana
World Development
 , 
2000
, vol. 
28
 (pg. 
663
-
81
)
Marr
A
The poor and their money: what have we learned? ODI Poverty Briefing
1999
London
Overseas Development Institute
 
Martin
I
Hulme
D
Rutherford
S
Financial services for the poor and poorest: deepening understanding to improve provision. Finance and Development Research Programme
Working Paper series, no. 9.
 , 
1999
Manchester
Institute for Development Policy and Management, University of Manchester
Masinde
JRM
Determinants of women's groups’ performance and their implications for rural development planning: A study of Sirisia Division in Bungoma District Kenya
Unpublished MA thesis.
 , 
1987
Kenya
University of Nairobi
McIntyre
D
Thiede
M
Dahlgren
G
Whitehead
M
What are the economic consequences for households of illness and of paying for health care in low- and middle-income country contexts?
Social Science and Medicine
 , 
2006
, vol. 
62
 (pg. 
858
-
65
)
Musau
S
Community-based health insurance: experience and lessons learned from East Africa
Technical Report No. 34.
 , 
1999
Bethesda, MD
Partnerships for Health Reform Project, Abt Associates Inc
Oyaya
CO
Rifkin
SB
Health sector reforms in Kenya: an examination of district level planning
Health Policy
 , 
2003
, vol. 
64
 (pg. 
113
-
27
)
Pankhurst
H
Passing the buck? Money literacy and alternatives to credit and savings schemes
Gender and Development
 , 
2002
, vol. 
10
 (pg. 
10
-
21
)
Preker
AS
Carrin
G
Dror
D
, et al.  . 
Effectiveness of community health financing in meeting the cost of illness
Bulletin of the World Health Organisation
 , 
2002
, vol. 
80
 (pg. 
143
-
50
)
Putnam
R
Making democracy work: civic traditions in modern Italy
1993
Princeton
Princeton University Press
 
Cited in: Gilson L. 2003. Trust and the development of health care as a social institution. Social Science and Medicine56: 1453–68
Putnam
R
Bowling alone: America's declining social capital
Journal of Democracy
 , 
2002
, vol. 
6
 (pg. 
65
-
78
)
Rahman
A
Micro-credit initiatives for equitable and sustainable development: who pays?
World Development
 , 
1999
, vol. 
27
 (pg. 
67
-
82
)
Rakodi
C
Gatabaki-Kamau
Devas
N
Poverty and political conflict in Mombasa
Environment and Urbanisation
 , 
2000
, vol. 
12
 (pg. 
153
-
70
)
Ritchie
J
Spencer
L
O’Connor
W
Ritchie
J
Lewis
J
Carrying out qualitative analysis
Qualitative research practice: a guide for social science students and researchers.
 , 
2003
London
Sage Publications
Russell
S
Ability to pay for health care: concepts and evidence
Health Policy and Planning
 , 
1996
, vol. 
11
 (pg. 
219
-
37
)
Russell
SJ
Can households afford to be ill? The role of the health system, material resources and social networks in Sri Lanka
PhD Thesis
 , 
2001
University of London
Russell
S
The economic burden of illness for households in developing countries: a review of studies focusing on malaria, tuberculosis and human immunodeficiency virus/acquired immunodeficiency syndrome
American Journal of Tropical Medicine and Hygiene
 , 
2004
, vol. 
71
 
Suppl. 2
(pg. 
147
-
55
)
Schneider
P
Why should the poor insure? Theories of decision-making in the context of health insurance
Health Policy and Planning
 , 
2005
, vol. 
19
 (pg. 
349
-
55
)
Stierle
F
Kaddar
M
Tchicaya
A
Schmidt-Ehry
B
Indigence and access to health care in sub-Saharan Africa
International Journal of Health Planning and Management
 , 
1999
, vol. 
14
 (pg. 
81
-
105
)
Uphoff
N
Local institutional development: an analytical sourcebook with cases.
 , 
1986
West Hartford, CT
Kumarian Press
UNFPA
From microfinance to macro change: integrating health education and microfinance to empower women and reduce poverty.
 , 
2006
New York
UNFPA
World Bank
World Development Report.
 , 
2000
Washington, DC
The World Bank
World Bank/Republic of Kenya
Social dimensions of development – revised approach to human-centred development and targeted poverty interventions.
 , 
1996
Nairobi
Government Press