Too poor to pay: charging for insecticide-treated bednets in highland Kenya


Helen Guyatt, Wellcome Trust Research Laboratories/KEMRI, Nairobi PO Box 43640, Kenya. E-mail:


WHO has proposed malaria control as a means to alleviate poverty. One of its targets includes a 30-fold increase in insecticide-treated nets (ITNs) in the next 5 years. How this service will be financed remains unclear. In July 2000, 390 homesteads in rural highland Kenya were interviewed on their willingness to pay for ITNs. The costs to a household of protecting themselves with ITNs were compared with current household expenditure. Homesteads expressed a willingness to pay for ITNs, but the amounts offered were not sufficient to cover the costs of providing this service without donor support to meet the difference. Furthermore, as most household expenditure was allocated to basic needs these interventions were ‘unaffordable’. The cost of protecting a household with ITNs would be equivalent to sending three children to primary school for a year. The aspiration by poor rural homesteads to protect themselves with ITNs is not compatible with their ability to pay. One option to have an immediate equitable impact on ITN coverage and break the cycle between malaria and poverty is to provide this service free of charge.


There is overwhelming evidence that insecticide treated nets (ITNs) have enormous potential for alleviating the burden of malaria in Africa. The challenge is to move beyond clinical and experimental trials to implementation at a national level. As with most effective public health interventions in Africa, a major barrier in going to scale will be economic. Most governments in countries with endemic malaria would not be able to afford the provision of these preventative measures from existing budgets. One approach to financing health care in developing countries has been to ask communities to pay for these services. Willingness to pay (WTP) studies have emerged as the fashionable approach to establishing the potential of such cost recovery schemes. They have been widely used in the health field as a marketing tool to help in setting prices prior to introducing user fees, and to provide information on potential demand and revenue (Russell et al. 1995).

Most research on cost recovery treats WTP as synonymous with ability to pay as ‘Consumers are assumed to be able to afford whatever they are willing to pay, because they know best how to allocate their resources’ (Russell 1996). Guidelines for defining affordability do not exist, but one approach has been to consider the opportunity costs of payment on other household expenditures (Russell 1996). This paper presents a comparison of the WTP responses for ITNs in rural homesteads of highland Kenya with the ability to pay as assessed from current household expenditure.


Survey population

Kisii and Gucha Districts (both formerly Greater Kisii) are in Nyanza Province of western Kenya. Both lie within the western highland belt, with altitudes ranging between 1400 and 2200 m. They are rural farming areas with a high population density (700 people per square km) (CBS 2001). In 1999, Merlin (Medical Emergency Relief International) initiated the provision of ITNs in these districts through 32 organized community groups. Each group was trained by Merlin and the Ministry of Health (MoH) staff, and provided with nets, K-Othrine 1% SC (deltamethrin; Aventis) and all the materials necessary for net treatment. A total of 15 000 bednets were distributed amongst the groups. A first quota of nets (160 per group in Kisii and 125 in Gucha) were sold at a highly subsidized price of Ksh 50 (US$ 0.64) to households with pregnant women and young children who were considered unable to pay Ksh 350. These ‘vulnerable’ individuals were identified in consultation with the local chiefs. All subsequent nets were sold for Ksh 350 (US$ 4.49) (53% of the estimated total cost of ITNs at US$ 8.42) (Guyatt et al. 2002a).

In July 2000, a homestead survey was undertaken in Nyamache division (Gucha district) to evaluate the impact of the vector control activities by Merlin on Plasmodium falciparum infection. Nyamache division is one of the seven divisions in Gucha district with a population of 54 722. The survey took place in five of the 12 sublocations (Turwa, Omokonge, Kiobegi, Masisi and Emenwa). The community group providing bednets for the study homesteads was based at Kionyo dispensary (Turwa sublocation).

Two types of homesteads were interviewed: (1) those with at least one ITN from a Merlin trained community group (n = 190) and (2) those with neither a Merlin ITN nor targeted for any other intervention (n = 200). All homesteads (n = 390) were interviewed regarding basic socio-economic variables. Details of the survey and the cost-effectiveness analysis of the vector control interventions are available elsewhere (Guyatt et al. 2002a, b), with ITNs reducing the risk of P. falciparum infection by 63%.

Willingness to pay assessment

The WTP assessment involved both an unprompted open-ended question (How much would you pay?) and a ‘take-it-or-leave-it’ option following a brief statement on the high cost of ITNs [These are expensive. Would you pay Ksh 350 (US$ 4.49)?].

In homesteads with an ITN, two groups were identified: (1) those with a single Ksh 50 net (i.e. chosen by the selection committee as a vulnerable homestead) (n = 116) and (2) those with a single Ksh 350 (n = 52) or a single shop net treated with insecticide (n = 2) or multiple ITNs (n = 20). Chi-squared analysis was performed to test for differences in the WTP responses by four household wealth indicators: (1) possession of a radio, (2) possession of any cattle, (3) a shamba size (farming land) of > 2 acres and (4) the presence of any semi-permanent buildings. Homesteads without an ITN were asked why they did not have one, and the response was recorded verbatim.

Costs to a household and ability to pay

In 1997, the third Welfare and Monitoring Survey was conducted in Kenya. This covered 10 879 households throughout the country and 419 households in Greater Kisii district (Ministry of Finance and Planning 2000a). Households were classified as poor based on an overall poverty line of < Ksh 1239 (US$ 15.88) expenditure per adult person per month in rural areas. Three-quarters of this expenditure were allocated to food requirements for the recommended daily calorie requirement of 2250, with only Ksh 312 (US$ 4) for non-food expenditures. The expenditures of poor households in Greater Kisii were assessed (Ministry of Finance and Planning 2000b), alongside income data from this area (Kinnear 2000) to provide an indication of a household's ability-to-pay.

The costs to a household of protecting all its residents with ITNs were estimated assuming three ITNs would be needed. This was based on average household size of six (Ministry of Finance and Planning 2000a) and two people sleeping under a net (Guyatt et al. 2002b).


Willingness to pay for ITNs

Of all the homesteads interviewed (n = 390), 97% were willing to pay for an ITN (Table 1). The most frequent amount offered was Ksh 50 (US$ 0.64) (46% of responses). Only 4% of homesteads initially offered Ksh 350 or more, but after being told that nets are expensive, 26% of homesteads said they would pay Ksh 350 for an ITN. There were clear differences in the WTP responses from homesteads with different prior bednet experiences. Homesteads with a single ITN purchased at the highly subsidized price of Ksh 50 were significantly less likely to be willing to pay Ksh 350 (18%) than those who had bought at least one net at Ksh 350 or a shop net (49%) (P < 0.001). Although there was a trend for wealthier households to be more willing to pay Ksh 350 for an ITN (data not shown), the only significant difference was observed for radio ownership in ITN homesteads (15%vs. 36%, P = 0.07).

Table 1.  Willingness to pay for ITNs in 390 rural homesteads in Gucha district, Kenya (in Ksh in the year 2000) 1
with a single
Ksh 50 ITN
with Ksh 350
or shop ITNs
with no ITNs
  • 1

     1 US$ = 78 Ksh.

Number examined11674200390
Number willing to pay for ITNs (%)111 (96%)71 (96%)195 (98%)377 (97%)
Amount willing to pay (Ksh.)
  Median (lower, upper quartile)50 (50, 100)100 (50, 250)50 (50, 100)50 (50, 150)
  Mean (SE)90 (6.96)150.7 (12.2)116.2 (6.06)107.3 (4.34)
  Mode (minimum, maximum value)50 (20, 350)50 (50, 350)50 (5, 350)50 (5, 350)
Number willing to pay Ksh 350
 after prompting (% of total examined)
21 (18%)36 (49%)45 (23%)102 (26%)

For the homesteads without ITNs that responded to the question of why they did not have one (n = 195), the most frequent answer was financial (87%). The following responses from three homesteads: ‘cannot afford to buy a net because of school fees and basic needs’, ‘don’t have money to buy a net for everyone in the household' and ‘planning to buy at end of month when they get the money from the tea’ suggested that availability of cash is a problem.

Ability to pay

The third Welfare Monitoring Survey found that 52% of the Kenyan population and 48% of households in Greater Kisii were classified as poor. The average monthly expenditure by poor households in Greater Kisii was Ksh 3250 (US$ 42). Most expenditure went on food (86%), with only Ksh 126 (3.9%) being spent on education per month and Ksh 46 (1.4%) on health (Fig. 1). This profile is typical of the rural poor throughout Kenya. Most male heads of poor households in Greater Kisii are unpaid family workers (63%) and 39% operate their own trade without employees, suggesting little opportunity for earning cash through regular full-time employment.

Figure 1.

Household expenditure in Greater Kisii and the opportunity costs of purchasing ITNs. The bars represent the cost of ITNs for a household (US$ 13.5) as a percentage of the annual household expenditure on different items (given by the squares on the line graph).

Daily cash needs for households [soap, salt, kerosene, tea and sugar at < Ksh 50 (US$ 0.64) per day] are met through the sale of produce, such as maize, matoke and fruits, or from contract work (agricultural work or tea picking can earn Ksh 40 per day). The main cash crop is tea, which brings in Ksh 6 per kilo on collection and an additional Ksh 6 per kilo bonus at the end of the year. A monthly yield of 50 kg (from around 0.5 acre) would provide a monthly income of Ksh 300 (US$ 3.85) and an annual bonus of up to Ksh 3600 (US$ 46). The tea bonus is often set aside for items such as school fees, clothes and medical bills. The annual cost of sending a child to primary school is Ksh 300–500 (US$ 3.85–6.41). Livestock will only be sold in an emergency, such as payment of hospital bills. A chicken can be sold for Ksh 150 (US$ 1.92) and a young local milking cow for as much as Ksh 10 000 (US$ 128). In the 390 homesteads interviewed as part of the WTP survey, nearly one-quarter had no cattle, one-third had less than four chickens and half had less than 3 acres of land.

The financial outlay for a household of protecting themselves with ITNs (three nets) purchased from a Merlin trained community group would be Ksh 1050 (US$ 13.5). This would represent 70% of current annual expenditure on education and twice what is spent on all health care (Fig. 1).


User fees are a popular approach for supporting and sustaining curative health services in less developed countries, and WTP studies are considered a simple, quick and cheap approach to evaluate their potential revenue and demand (Russell et al. 1995). This methodology has also extended into payment for preventative health care measures such as ITNs or net retreatment (Desfontaine et al. 1990; Mills et al. 1994; Dgedge et al. 1999; Rowley et al. 1999; Some 1999; Onwujekwe et al. 2000). Most studies express some words of caution as to the reliability of the WTP responses. One approach suggested to validate the responses are associations with household wealth variables, with the premise that the more income available the more households should be WTP. The ultimate test has been to compare hypothetical WTP with actual payment. Although recent work in Nigeria has suggested that the WTP technique can predict more than 75% of actual payments for ITNs (Onwujekwe 2001), the performance in other settings has been far less impressive. In Mozambique (Dgedge et al. 1999) and The Gambia (Rowley et al. 1999) only half of households who said they would be willing to pay for an ITN or net treatment actually went ahead with this. The low demand for bednet retreatment following the introduction of user fees in The Gambia, Kenya and Tanzania was attributed to previously providing this service free-of-charge (Cham et al. 1997; Winch et al. 1997; Snow et al. 1999). Evident from this study was that the Ksh 50 pricing of nets for vulnerable groups has influenced the price that homesteads are prepared to pay for them in the future. Less than one-fifth of those who had received one of these highly subsidized nets were willing to pay Ksh 350, and even half of the homesteads that had paid the Ksh 350 (or more for a shop net) were unwilling to do so again.

Our study also suggests that although homesteads express a WTP for ITNs, the amount offered is not sufficient to cover all costs. This may suggest a demand for ITNs, but an inability to finance this. Half of the million people living in Kisii and Gucha districts are classified as living below the poverty line. In these households, most money is spent on food, and financing of ITNs would need to be at the expense of such basic needs. This study suggests that purchasing ITNs for a household (at Ksh 1050) could be the cost of sending three children to primary school for the year (assuming fees of around Ksh 300 per child). It has been implied that if ITNs deplete less than 2% of annual household expenditure, they are affordable based on the assumption that between 2% and 5% of household expenditure relates to health care (Onwujekwe 1999). In Greater Kisii poor rural households spend only 1.2% on health care.

The introduction of community financing mechanisms or other fund raising innovations have an important role to play in purchasing health commodities in the future. However, it is clear that ITNs are currently not affordable to most of the rural population in Kenya. The US Congress has recently supported a new amendment to US foreign aid law that opposes loans by international financial institutions that ‘would require user fees or service charges on poor people for primary education or primary healthcare, including prevention and treatment efforts for HIV/AIDS, malaria, tuberculosis, and infant, child and maternal wellbeing’ (Africa Policy Information Centre 2001). Providing bednets free of charge has been shown to be a good use of scarce public resources (Evans et al. 1997). Malaria and poverty are intrinsically linked with a recent analysis suggesting that a 10% reduction in malaria is associated with a 0.3% higher economic growth (Gallup & Sachs 2001). The control of malaria has been proposed as a means to alleviate poverty by the WHO, with ambitious targets such as a 30-fold increase in the availability of nets in the next 5 years (Yamey 2000). It is clear that governments cannot break this cycle between malaria and poverty without donor commitment. Whether it is more cost-effective to supply a highly subsidized ITN or a free ITN will depend on the operational constraints of recovering funds. We would argue that the simplest approach would be to provide ITNs free of charge; this would maintain equity and guarantee a marked and immediate increase in access to ITN services across Africa.


The authors are grateful to the Wellcome Trust (UK) and the Kenya Medical Research Institute for funding, and Prof. Julius Meme for his comments on the manuscript. HLG is in receipt of a Wellcome Trust Research Career Development Fellowship (055100) and RWS has a Senior Research Fellowship in Basic Biomedical Science (058992). This paper is published with the permission of the director of KEMRI.