MRC policy briefs to government

Use of insecticide-treated bednets by communities reduces malaria transmission in comparison to house spraying in Kwazulu-Natal

A. E. P. Mnzava,1 B. L. Sharp,1 D. J. Mthembu,2 S. S. Dlamini,1 J. K. Gumede,2 I. Kleinschmidt,3 C. Goodman.4

1Malaria Research Programme of the South African Medical Research Council
MRC, PO Box 17120, Congella 4013, South Africa. 
Tel.: (031) 25-1481; E-mail: mnzavaa@mrc.ac.za. 
2Malaria Control Programme of the Department of Health, Jozini. 
3Division of Biostatistics and Epidemiology, South African Medical Research Council. 
4Department of Health Economics and Financing,
London School of Hygiene and Tropical Medicine, UK

Successful trials else-where in Africa have shown that for every 180 insecticide-treated bednets used, one death of an African child is prevented per year. 1 A  study involving approximately 14 000 people in KwaZulu-Natal was therefore initiated to compare the use of insecticide-treated bednets with traditional insecticide spraying of houses for malaria vector control in terms of efficacy and cost. Do people accept bednets when there is no tradition of use? Can and should the use of bednets by communities replace traditional insecticide spraying of houses? What is the best method to implement (finance and distribute) a bednet programme in South Africa? 

This study, a collaboration between the MRC and the Department of Health, Jozini, has policy implications for malaria vector control in South Africa, in a region where malaria transmission is seasonal and house spraying with insecticide has been the mainstay of vector control for over 5 decades.2 The study was initiated in 1997.

To measure significant reductions in malaria incidence between the two interventions, a Geographical Information System (GIS) was used to identify and create seven pairs of blocks of equal incidence over time in the high-risk malaria areas of Ndumu and Makanis in  Ingwavuma Magisterial District, KwaZulu-Natal.3 Individual blocks were then randomly allocated to receive either insecticide-treated bednets or house spraying with deltamethrin, a synthetic pyrethroid (Figure 1). Malaria cases were routinely collected by either surveillance agents at homes or as reported by health facilities.

Distribution, treatment and re-treatment of bednets
Prior to bednet distribution to the selected homesteads, community health information campaign meetings were conducted in the trial area. People’s sleeping patterns and preferences for colour and shape of nets were also assessed. Green and conical nets were the most preferred because they do not show dirt easily, require less washing and are easier to hang, especially in homes where people sleep on mats. The nets were distributed in January 1997 by the surveillance teams of the Department of Health, Jozini, after training by  research staff. 

Residents from the bednet-designated homesteads were invited to distribution centres (10 in Ndumu and 5 in Makanis) where they were shown how to hang and maintain nets. Families were then issued with nets of a size to adequately cover 2 persons each.

The surveillance teams in turn trained communities to re-treat their own nets in 1998, 1999 and 2000 using the same distribution centres. This method of distributing and re-treating nets by surveillance agents of the Malaria Control Programme of the Department of Health has proved very successful in terms of logistics, and high re-treatment rates were obtained. For example, it only took 2 days on average to re-treat all the nets covering approximately 7000 people, while it would have taken between 2-3 weeks to spray their houses.  Moreover, elsewhere in Africa re-treatment rates declined to below 20% when communities had to organise and pay for re-treatment services with little or no public sector support.4

Acceptance of bednets by communities
Community surveys to assess people’s knowledge, attitude and practices in relation to the bednet programme were conducted in 1997 and 1998. This provided an opportunity to assess acceptance of and compliance by communities with net usage in relation to their knowledge on malaria. A semi-structured questionnaire was randomly administered to 10% of all the  homesteads in the study area followed by focus group discussions with key informants.

It was confirmed that provision of green and conical nets was what communities wanted because they do not show dirt  and are easy to hang - “You give people what they want”. Assessment of people’s knowledge, attitudes and practices towards bednets indicated that communities had a good knowledge of malaria and had accepted nets very well. They were using bednets consistently because of a perceived benefit - “We no longer have to cover ourselves with blankets in summer”. This has been supported by an increased demand for bednets from areas earmarked for spraying. Very low frequencies of net washing have been recorded, indicating that communities had understood the importance of not washing nets as it compromised efficacy.

Impact of bednet usage on malaria incidence
Assessment of the impact of use of bednets by communities of the Ndumu and Makanis areas of KwaZulu-Natal show that this intervention significantly reduced malaria incidence (rate of transmission) both in 1997 (RR = 0.879, 95% CI 0.80 - 0.95, P = 0.04); in 1998 (RR = 0.667, 95% CI 0.61 - 0.72, P = 0.0001), as well as in 1999 (RR = 0.72, 95% CI 0.66 - 0.79, P < 0.001). These reductions were further confirmed by an assessment of the rate of change between 1996 (baseline) and 1999, with a 31% reduction in malaria incidence (228 - 158 cases/1000 person years) in homesteads using treated bednets and an increase of 18% (189 - 222 cases/1000 person years) in sprayed areas.

These results clearly indicate that use of insecticide-treated bednets significantly reduced malaria incidence in comparison to insecticide house spraying.  The results can only be explained by the fact that bednets when used do form a physical barrier between mosquitoes and people and that mosquitoes attempting to bite people sleeping under nets are either repelled and/or killed.5 This conclusion was supported by collecting a high proportion of unfed and dead mosquitoes (> 90%) in houses with nets fitted with window traps.

Cost-effectiveness of bednets
To decide whether bednets can replace house spraying for malaria control in South Africa, we needed data not only on the efficacy of bednets and their acceptance by communities but also on their cost in comparison to house spraying. A tool for malaria control has to be both effective and relatively cheap. Financial costs and effectiveness of the two interventions were compared both in trial and operational settings. The potential saving in treatment cost between the two strategies was also estimated.

This required calculating incremental costs by collecting cost data (both financial and economic costs) retrospectively for all activities involved in the implementation of bednets and house spraying - but excluding research costs. The main outcome measure was numbers of confirmed cases of parasitaemia. Calculation of total costs saved involved detailed costing of laboratory, clinic and hospital costs likely to be saved. Using the best estimates for each variable, insecticide-treated bednets were cheaper per person covered (R20/person/year) than house spraying (R38/person/year).

These results are based on efficacy data for 1999 where bednets showed a 31% reduction in malaria incidence as compared to an18% increase in incidence with house spraying. Interestingly, however, insecticide house spraying would be cheaper (not necessarily effective) than bednets if spray teams were recruited only for the period of spraying. In KwaZulu-Natal  malaria spray teams are employed throughout the year.

Cost recovery or free distribution of bednets
There is a lot of discussion both locally and internationally that bednets should not be distributed free.  It is argued that communities will not attach value to bednets issued this way.6 This is, however, contrary to our findings in northern Kwazulu-Natal from the KAP (knowledge, attitudes and practices) study in which communities objected to the sale of nets both at subsidised or at full market price.

With support from the Rotary International, Keitchner-Conestoga club of Canada, bednets were purchased and a study was initiated to evaluate this method of financing (cost recovery) and distributing bednets through the community. This was also one of the requirements by Rotary International. The intention, however, is not to eventually ask communities to shoulder the burden of malaria control (when house spraying is currently provided at no cost), but rather to have answers should such a suggestion be raised by policy makers.

Initial attempts to use local shop owners to sell nets to the community failed partly because of cost. Nets were purchased locally at R80 per net and then re-sold to communities at a subsidised cost of R65 per net (out of which R10 was commission and R5 to cover the cost of the insecticide).  Alternative sources of procuring good-quality nets at a cheaper price were explored through the World Health Organization (Pretoria) and sourced at a cost of R37 per net. These nets and those bought locally are currently sold through local community structures (bednet committees) - headed by Indunas (councillors to the chief of the tribal area).  Eight such committees have been set up and operate a bednet account.  Money from the sale of these nets (R45/net) is deposited with the MRC who audit the process, and used for procuring more nets towards increasing coverage. Staff of the DoH and the MRC oversee the operation and advise the bednet committees. Coverage of nets is being monitored through a GIS platform which will allow an assessment of the impact of this strategy on malaria reduction.

Policy implications and the way forward
The study has clearly shown that insecticide-treated bednets are:

• effective for malaria control,
• easier to implement,
• cost less than house spraying and
• have been well accepted by local communities.

On the other hand, however, we do not as yet have all the answers with regard to the long-term sustainability of this control intervention. Moreover, there is fear among policy makers of losing the gains made with house spraying. Weighing the advantages of the bednet strategy over house spraying for malaria control, we recommend that:

1. With the support of the Department of Health, we need to scale up bednet programmes, i.e. move from trials/projects to large programmes in KwaZulu-Natal as well as in the other two provinces where malaria is a problem. This will allow further assessment of sustainability by and acceptability to local communities. With the Malaria Health Information System in place in all three provinces, this will enable us to determine the efficacy of bednets under programme settings, which is different from the trial/research environment. 

2. Resistance of Anopheles funestus (one of the malaria vectors to synthetic pyrethroids) has been reported recently from the study area. The degree and geographical spread of this resistance and its effect on malaria transmission is still unknown. Use of insecticide-treated bednets, however, is considered one of the methods to manage insecticide resistance in mosquitoes. It is therefore imperative that this study also addresses the problem of resistance in South Africa.

3. We recommend that implementation of a bednet programme should be through the public sector (Malaria Control Programmes) for both purchase, distribution and organising of bednet re-treatments. This seems to be the most appropriate method of implementation since an infrastructure is in place. The best results in the world (China) with bednets have been in areas as is the case in South Africa with such an infrastructure and with a previous long history of house spraying.7

4. Arguments supporting cost recovery of bednets, however good they may sound, are indeed no justification to ask communities to pay for nets while in actual fact they are protecting the people and development south of the malaria high-risk areas.  Moreover, no one has ever had to pay for house spraying all these years. Malaria control is essential to protect the economic and industrial development of large parts of South Africa.2

References

1. Lengeler C. Insecticide treated bednets and curtains for malaria control (Cochrane Review). In: The Cochrane Library, Issue 3, Oxford, UK: Update Software, 1998.                  Sharp  BL, le Sueur D. Malaria in South Africa - the past, the present and selected implications for the future. SAMJ 1996; 86 (1): 83-89. le Sueur D, Ngxongo S, Sharp B, Martin C, Fraser C, Teuschner M, Tollman  S, Green C, Tsoka J, Solarsh G, Mnzava AP.  Towards a Spatial Rural Information System. Durban: Health Systems Trust/Medical Research Council, 1997. Lengeler C, Cattani J, de Savigny D, eds. Net gain: a new method of preventing malaria deaths. Ottawa: International Development Research Centre; and Geneva: WHO, 1996.   Magesa SM, Wilkes TJ, Mnzava AEP, Njunwa KJ, Myamba J, Phillip M, Hill N, Lines JD,  Curtis CF. Trial of pyrethroid treated bednets in an area of Tanzania holoendemic for malaria. Part 2: Effects on the malaria vector populations. Acta Tropica 1991; 49: 97-108.
2. Carrol D, ed.  International conference on bednets and other insecticide-treated materials for the prevention of malaria (Proceedings Report). Washington, DC: USAID, 1997.
3. Cheng Huai Lu, Yang Wen, Kang Wuanmim, Liu Chongyi.  Large-scale spraying of bednets to control mosquito vectors and malaria in Sichuan, China. Bull WHO 1995; 73 (3): 321-328.