Sustainability of pyrethroid-impregnated bednets for malaria control in Afghan communities.

Sustainability of pyrethroid-impregnated bednets for malaria control in Afghan communities* M. Rowland,1 S. Hewitt,2 N. Durrani,3 P. Saleh,4 M. Bouma,5 & E. Sondorp6 Between 1992 and 1995 a series of studies was undertaken to assess the long-term suitability of pyrethroid- impregnated bednets (PlBs) for malaria control in Afghan refugee communities in two villages in North-West Frontier Province, Pakistan. During 1992, 86% of bednet owners volunteered to have their bednets re- impregnated, and a further 15% of families purchased nets at two-thirds of cost price. From 1992 onwards, 27% of the villagers returned to Afghanistan, and annual house spraying campaigns were introduced to protect those still resident but sleeping without bednets. Within 3 years, these campaigns, together with PlBs, reduced the annual incidence of malaria by 87%, from 597 to 78 cases per 1000 population. Nevertheless, 65% of resident families continued to re-impregnate their nets annually with permethrin. To assess whether PlBs were still being used and were still protective, in view of these reduced transmission rates, we carried out a case-control study in 1994 on febrile or otherwise symptomatic patients presenting at village health centres. Comparison of the slide-positivity rates of PIB users and those without bednets showed that regular usage reduced the odds of contracting falciparum and vivax malaria to 0.22 (95% confidence interval (Cl): 0.09-0.55) and 0.31 (95% Cl: 0.19-0.51), respectively. There was no evidence of a sex- or age-bias in bednet use or in protective effect. The results indicate that a community-based PIB programme is an appropriate malaria control measure in areas where management or security problems make traditional house-spraying campaigns impossible. A relevant finding for those involved in the monitoring of bednet distribution projects is that the local coverage of bednets and the local impact on malaria, even when introduced to remote areas, can be estimated very cheaply by health centre microscopists who simply catalogue blood film diagnoses according to patients' bednet use practices. Introduction Field trials in many countries have demonstrated that use of pyrethroid-impregnated bednets (PIBs) is effective for preventing malaria (1). Even in coun- tries where there is no great tradition of bednet use, carefully conducted field trials, where appropriate attention is paid to health education before and dur- ing implementation, generally show that recipients use the nets with enthusiasm and to good effect. One such trial was conducted in Afghan refugee camps in 1991; the incidence of clinical parasitaemic episodes was reduced by 53% in an intervention group whose previous use of nets was only 2% (2). Nevertheless, a number of questions remain on whether use of PIBs * From: HealthNet International, P.O. Box 889, University Town, Peshawar, North West Frontier Province, Pakistan. Requests for reprints should be sent to Dr Rowland at this address. Coordinator, Malaria Control Programme. 2 Vector Control Manager. IParasitologist and Laboratory Manager. 4Bednet Project Manager. IFormerly: Coordinator, Malaria Control Programme. 6 Director. Reprint No. 5750 is a sustainable strategy for malaria control: Will recipients continue to use their nets in the long term in the absence of further encouragement? Will re- cipients actively participate in the re-impregnation process and thereby help to reduce recurrent im- plementation costs? Will neighbours buy bednets for themselves? Will the reduction in malaria be sus- tained? Will people continue to use nets if a second method of control - house spraying, for example is conducted simultaneously? This article describes a number of follow-up studies in the Afghan trial villages that attempt to answer such questions. Methods Distribution, re-impregnation and sale of bednets The two trial villages, Baghicha and Kagan, are situ- ated in Mardan district, North-West Frontier Prov- ince, Pakistan. Extended family groups of Afghan refugees live in houses constructed from mud and located within high-walled compounds. The underly- ing water table is higher than normal for the district, Bulletin of the World Health Organization, 1997, 75 (1): 23-29 © World Health Organization 1997 23 M. Rowland et al. and both villages are malarious. Refugees sleep out- doors during the summer and indoors from October. The populations in a 1991 census were 11220 (Baghicha) and 2735 (Kagan). In the PIB trial con- ducted in 1991, 20% of the village health centres' family registration cards were selected at random and divided into two groups. One group (173 fami- lies) was issued with imported, 100-denier, polyester bednets, while the other group (186 families) served as controls. The two groups were similar with respect to variables that might otherwise confound the re- sults of the trial (2). The bednets were designed to fit around the wooden-framed rope beds (charpois) used by Afghans and were dipped in an aqueous solution of permethrin (25% emulsifiable concen- trate; Imperator, ICI, Karachi, Pakistan) to achieve an application rate of 0.5gm-2. A pre-trial census ensured that each family received sufficient nets to cover all members. In December 1991, at the end of the trial, bednets were also distributed to the 186 families belonging to the control group. Families were then told they could do what they liked with the nets but that if they kept them, the nets would be re- impregnated the following year. In April 1992 a free net re-impregnation service was set up, run by basic health workers at the two village health centres. Family-size bednets were sold for Rs 100 per net (US$ 3.25 at 1992 rates), two- thirds of the purchase price. The availability of these services was broadcast by loudspeaker from village mosques. To ensure accuracy in record keeping, we provided re-treatment services only to families who produced their health centre record card and number. Those families who had been issued with nets in 1991 but failed to appear with their nets in 1992 were visited at home. In 1993 and 1994 the re- impregnation service was provided but no further nets were put on sale. Protection from malaria among PIB users Each health centre employed a malaria supervisor to take blood smears and administer antimalarials and a microscopist to examine the slides. All out- patients suspected of having malaria were registered by name and family code number and given pre- sumptive treatment with chloroquine while awaiting the microscopy results. Those patients who were confirmed positive were treated with chloroquine (25mg per kg body weight), and cases of vivax ma- laria were given an additional 5-day course of primaquine (1.25mg per kg body weight). Health centre microscopists were monitored monthly, en- suring that incorrect diagnoses were below 3%. To determine the protective effect of PIBs, records of all patients from whom blood smears had been taken between June and December 1994 were extracted from the registers and matched with the code numbers of families known to possess bednets. Since not all the bednet owners necessarily used their nets regularly, the malaria supervisors were instructed to ask in passing during smear-taking whether or not the patient used a bednet. Because the malaria supervisors were not employed by HealthNet International, it was hoped that their enquiries would elicit a more truthful response from patients. A parasite prevalence survey was also con- ducted in October 1994, the month of peak malaria transmission. For this purpose, 196 families known to possess bednets were randomly selected from those who had participated in the trial in 1991. Also 230 families known not to possess nets were ran- domly selected from the health centres' records. Thick and thin smears were taken from each family member, stained with Giemsa, and examined "blind" by an experienced microscopist. Analysis of the data For the analysis of the passive case detection data, patients presenting at the health centres with symp- toms of malaria and who were slide positive were taken to be cases, distinguishing between falciparum and vivax malaria. Patients with symptoms of ma- laria but with a negative slide and who presented during the same period at the same health centres were taken as controls. By comparing cases with con- trols, we calculated odds ratios for protection by bednets. Data were analysed using Epi Info software (3). Discrete variables were compared using x2 tests, while odds ratios were calculated using Mantel- Haenszel stratified cross-tabulations and summary X2 tests. Village incidences Malaria cases from health centre records were com- piled to produce annual incidence and slide posi- tivity rates for 1991 to 1994. Results Distribution, re-impregnation and sale of bednets The data on distribution, re-treatment and sale of bednets are summarized in Table 1. The average number of nets issued in 1991 was sufficient for a family of 8.6 persons (average family size 7.8). The WHO Bulletin OMS. Vol 75 199724 Sustainability of pyrethroid-impregnated bednets Table 1: Ownership of bednets in the trial villages and attendance at the insecticide re-treatment sessions Original number of families No. of nets issued or purchased Mean no. of nets per family Double size Single size No. of families re-treating nets in 1992 No. of nets re-treated in 1992 No. of families still present in 1994 No. of families re-treating nets in 1994 No. of nets re-treated in 1994 Intervention families 173; 1ia 699 2.3 1.8 150 (87)b 605; 4.0' 125 80 (64) 313; 3.9 Control families 186; 10a 596 2.8 0.4 157 (84) 497; 3.2 141 94 (67) 270; 2.9 Families sold nets Other families 271; 1i5a 615 _ 2.3 - 201 123 (61) 301; 2.4 83 173; 2.1 a Original families expressed as % of village. b Figures in parentheses are percentages. c Figures in italics are the mean number of nets per family. intervention families were issued with a mixture of single- and double-size nets, but owing to a lack of the single-size nets, the control families received mainly double-size nets. In 1992, a year after the PIB trial, a further 271 (15%) refugee families bought 615 double-size nets. Of these families, 40% pur- chased only one net each, and thus did not meet the entire needs of an average-size family (on average, families purchased 2.3 nets). In 1992 a total of 86% of the trial families re- sponded to the mosques' broadcasts and came to the health centres to have their bednets re-treated; the remaining 14% of trial families were later visited in their homes whereupon their nets were re-treated in sitfu. Between the summers of 1992 and 1994, a total of 27% of families from Baghicha and Kagan re- turned to Afghanistan. Of the trial families still resi- dent in 1994. the proportion responding to the broadcasts had fallen to 65%. The reduction in the rate of net re-treatment among families who had originally purchased nets was approximately the same as that among trial families who had originally been given bednets. Between 1992 and 1994. a further 83 village families acquired bednets and attended re-treatment sessions (this group accounted for 22% of all the families who had nets re-treated in 1994). For the group of 297 families who received or purchased bednets directly from HealthNet International and were still turning up annually to the re-treatment sessions, 92% of the nets originally issued could still be accounted for in 1994. Thus families who had acquired nets from sources other than HealthNet International had mostly obtained them from those families who were no longer responding to the mosques' broadcasts or from those who had re- turned to Afghanistan. since there were no other sources of polyester nets in the study area. Protection from malaria among PIB users The redistribution of bednets within the study com- munity between 1992 and 1994 meant that each re- maining family could be placed into one of the following categories: a) trial families from 1991 re- treating their nets in 1994 (174 families); b) trial families from 1991 not re-treating their nets in 1994 (92 families); c) non-trial families who had acquired nets after 1991, re-treating them in 1994 (206); d) non-trial families who had acquired nets after 1991 but did not re-treat them in 1994 (ca. 78 families); e) families who had not acquired nets (ca. 785 families). Table 2 shows the results of an analysis, with respect to these categories, of passive case detection (PCD) records from the Kagan and Baghicha health centres (these records were of patients showing clinical symptoms indicative of malaria, with the diagnosis verified positive or negative by microscopy). For families possessing nets (categories a)-d)), the Plasmodiluni falciparulm positivity rate among symp- tomatic outpatients at, for example, the Kagan health centre, was 0.7% (13/1826), whereas for fami- lies without nets (category e)) the rate was 1.2% (28/ 2247). Comparison of these data with those from the Baghicha clinic (which exhibited the same trend), showed that the overall odds of contracting falciparum malaria for families with bednets was 0.50 relative to those without nets (data stratified by vil- lage and age group). Not all the families with nets had sufficient for every member, nor did all families re-treat their nets in 1994; 0.50 is therefore an under- estimate of the true protective value of PIBs to users. A more realistic estimate is obtained if the families from the 1991 trial known to have re-treated their nets (category a) above) are compared with families who lacked nets; the P. falciparum rate among cat- egory a) families from Kagan. for example, was only WHO Bulletin OMS. Vol 75 1997 25 M. Rowland et al. Table 2: Distribution of malaria cases among Afghan refugee users or owners of permethrin-impregnated bednets and among those who did not use or own bednets, by passive case detection (PCD) at village health centres or from parasite prevalence surveys No. in Baghicha No. in Kagan +ve -ve +ve -ve Odds ratio x2 test P-value PCD at clinics:a P. falciparum P. vivax No. slides examined PCD at clinics:d P. falciparum P. vivax No. slides examined PCD at clinics:e P. falciparum P. vivax No. slides examined Parasite prevalence surveys:' P. falciparum P. vivax No. slides examined 12 (0.4)b 33 (1.0) 13 (0.7) 28 (1.2) 60 (2.1) 91 (2.8) 30 (1.6) 69 (3.1) 2777 3272 1 826 2247 3 (0.4) 33 (1.0) 5 (0.5) 28 (1.2) 14 (1.7) 91 (2.8) 12 (1.2) 69 (3.1) 826 3272 1004 2247 1 (0.2) 44 (0.8) 4 (0.3) 37 (1.4) 2 (0.4) 149 (2.7) 16 (1.1) 80 (3.0) 488 5561 1337 2643 0 (0) 1 (0.01) 4 (0.01) 8 (0.02) 6 (0.06) 8 (0.06) 7 (0.02) 14 (0.04) 1081 1394 312 338 0.50; 0.31, 0.79C 0.65; 0.50, 0.85 0.39; 0.18, 0.79 0.48; 0.32, 0.73 8.3 0.004 9.6 0.002 6.5 0.011 11.9 0.001 0.22; 0.09, 0.55 11.8 0.001 0.31; 0.19, 0.51 21.5 0.001 0.47; 0.15, 1.55 0.97 0.326 0.68; 0.34, 1.38 0.89 0.345 a +ve, individuals from families who possessed bednets; -ve, individuals from families who did not possess bednets. bFigures in parentheses are %. c Figures in italics are 95% confidence intervals. d+ve, individuals from families who took part in the 1991 trial and who had their nets re-treated in 1994; -ve, individuals from families who did not possess nets. e +ve, individuals claiming to use bednets; -ve, individuals not claiming to use bednets. ' +ve, individuals from bednet-owning families; -ve, individuals from families not owning bednets. 0.5% (5/1004) and the overall odds ratio for both villages was reduced to 0.39. The parasite positivity rate for trial families who re-treated their nets (2.1%, 31/1447) was not signifi- cantly different from that for families who failed to do so (1.9%, 20/1035) (X2 test = 0.08. P = 0.77). This result is potentially important for determining the required re-treatment frequency. Not all members of families who re-treated their nets (category a)) necessarily used them regularly. To obtain an estimate of the odds ratio for regular users, we compared the parasite positivity rates for individuals who claimed to use nets against those who did not admit to using them. This gave a P. falciparium positivity rate of only 0.3% (4/1337) for bednet users from Kagan and, together with the re- sults from Baghicha, an odds ratio for contracting falciparum malaria of 0.22. PIBs also reduced the chances of becoming in- fected with vivax malaria, although to a lesser extent than for falciparum malaria (Table 2). Fig. 1 shows the PCD slide positivity rates among bednet users and non-users per month. The seasonal patterns are clear, with the peak for P. falciparuim malaria occurring slightly later than that for P. vivax malaria. As the year progressed and grew colder, the P. falciparum positivity rates tended to converge for PIB users and non-users, suggesting that some owners no longer used their nets regularly after they moved indoors. For P. vivax, however, this convergence was not observed. Parasite rates in the two villages were very low during the second half of 1994 because of house spraying campaigns carried out during July. Parasite prevalence surveys in October yielded very few posi- tive cases (1.5%, 48/3215) (Table 2). Although the odds ratios were not statistically significant for fami- lies that used bednets relative to those that did not, the values (P. falciparum (0.47); P. vivax (0.68)) and trends were similar to those obtained by PCD at the health centres. Census data revealed that 49.3% of the study population was male, whereas 58% of those who attended malaria clinics were female. The propor- tion of females who claimed to use bednets (15.4%, 1042/6761) was similar to that of males (17.4%, 846/4870). A bias towards use by males might be expected in families who bought only one or two nets, since not all females would necessarily be granted access to them in the male-dominant Afghan society. However, the proportions of males and fe- males who claimed to use the nets were very similar; WHO Bulletin OMS. Vol 75 199726 Sustainability of pyrethroid-impregnated bednets Fig. 1. Monthly malaria slide-positivity rates among users of permethrin-impregnated bednets and non- users presenting at the village health centres. a) Plas- modium falciparum; b) P. vivax. a) Fig. 2. Annual incidence of malaria in Baghicha and Kagan refugee villages, 1991-94. 500^_ S 8400 20 ti300 -1 200 0~~~~~~~~~M c 199 1992 1993 1991991 1992 1993 1994 Year P.vivax: P.falciparum: E Incidence * Incidence _ Positivity -*- Positivity WHO 96743 Jul Aug Sep Oct Nov Dec 7 Bednet users: 6 EII U .-- Yes No 1,5 - malathion; and in 1994 Baghicha was sprayed with malathion and Kagan with lambda-cyhalothrin. Discussion Ije12_ _ _ _The initial PIB trial conducted in 1991 was success- ful, with the incidence of malaria in the intervention group reaching only 47% of that in the control group 0 J (2). Our optimism continued during the following Jul Ajug Sep Oct Nov Dec year since 86% of trial families volunteered to have WH 6742 their nets re-impregnated and a further 15% of refugee families bought bednets. While these levels demonstrate the popularity of the new technique, they could also reflect the community's alarm at the for example, among the Kagan families who had rising trend in malaria, whose annual incidence had purchased nets, 65.1% (69/106) of females and risen to 597 cases per 1000; also, the study villages 67.5% (54/80) of males claimed to use them. The had not been sprayed with insecticide for 2 years. results of the case-control study bear out these Between 1992 and 1994, however, annual spray claims; parasite rates were identical (1.9%) for campaigns with malathion and lambda-cyhalothrin female (18/972) and male (16/858) net users, while were renewed to protect the families that lacked for non-users the rates were 4.2% (179/4224) for bednets. Together with PIBs, this reduced the over- females and 5.3% (150/2805) for males. all incidence of malaria by 87%, to 78 cases per 1000 There was no evidence of bias in the use of per year (4). The lower bednet re-impregnation rate bednets by any particular age group (X2 = 12.7, df in 1994 (65%) may well have been a result of the 10, P = 0.24). successful integrated malaria control policy, which used a combination of PIB, house spraying, and Annual incidence of malaria case detection by village health workers. With the chances of contracting malaria being so much lower Between 1991 and 1994 the annual incidence of ma- in 1994, the reduction in re-impregnation rate was laria (PCD records) fell from 597 per 1000 to 78 per neither surprising nor unduly disappointing. 1000 population (Fig. 2). In 1991 no spraying opera- The protection conferred by PIBs was estimated tions were carried out, but in 1992 the two villages using case-control and prevalence survey methods were sprayed with malathion; in 1993 Baghicha was applied to various categories of bednet owners or sprayed with lambda-cyhalothrin and Kagan with users. Direct questioning of symptomatic outpatients WHO Bulletin OMS. Vol 75 1997 27 M. Rowland et al. indicated that regular users were almost five times less likely to contract falciparum malaria (odds ratio = 0.22) and three times less likely to contract vivax malaria (odds ratio = 0.31). These are greater reduc- tions in the incidence of clinical, parasitaemic epi- sodes than those obtained using PIBs in Africa (5, 6) and south-east Asia (7), and are more comparable to results from China (1), possibly because of the lower vectorial capacities of South Asian malaria vectors (8). PIBs appeared to offer less protection against vivax malaria, but this might have been an anomaly caused by relapses. An alternative method of estimating protective efficacy, based on health centre family code numbers and bednet distribution/sales records, produced lower values than those obtained by direct question- ing since not all members of families that owned bednets necessarily used them. The results of parasite prevalence surveys in randomly selected PIB and non-PIB families showed a similar trend to that from PCD records. But by 1994, parasite prevalences had been so reduced by annual house spraying campaigns that a significant difference between bednet owners and non-owners was not apparent. Nevertheless, the majority of bednet owners continued to use their nets to good protective effect against malaria. Indeed, the results of the case-control study indicate that a community- wide PIB implementation campaign could, in a sin- gle year, have the same effect as 2 years of house spraying. The study confirms that in areas such as Afghanistan, where the cost and level of organiza- tion required to conduct spraying campaigns would be prohibitive, a community-based PIB programme should be adopted as the norm. It may be significant that families who failed to re-treat their nets appeared to be no less protected than those who did so. There are two possible expla- nations; the bednet rather than the insecticide might be the crucial protective factor prevailing in the study region or residues of permethrin from previous treatments were still providing sufficient protection (9). Irrespective of the explanation, the finding is important, because the logistics required to achieve a high annual re-treatment rate in Afghanistan where PIBs are being sold widely - is not trivial. The possible protective efficacy of untreated nets warrants closer examination. Accurate microscopy by health centre techni- cians, combined with direct questioning of patients, gave an estimate of protective efficacy consistent with more complicated or expensive methods, such as prospective intervention trials (2) or prevalence surveys in intervention and control groups. Local coverage of PlBs and the local impact on malaria, even when introduced to remote areas, can therefore be estimated by health centre microscopists who catalogue blood film diagnoses according to patients' bednet use practices. Acknowledgements We are grateful to the European Commission for funding HealthNet International's Malaria Control Project for Afghans. Allan Schapira is thanked for his comments on the article. The United Nations High Commissioner for Refugees oversees all health care activities in Afghan refugee villages and gave support to our study. The work could not have been undertaken without the participation of staff of the Austrian Relief Committee who run the health centres in Baghicha and Kagan. Resume Viabilite de l'utilisation des moustiquaires impregnees de pyrethrinolide pour la lutte antipaludique dans des communautes afghanes Une intervention de lutte antipaludique r6alis6e dans deux villages de r6fugi6s afghans en 1991 a d6montr6 que les moustiquaires impregn6es de pyr6thrinoTde, distribu6es gratuitement a 20% des families, 6taient culturellement acceptables et regulierement utilisees, meme lorsque tres peu de personnes avaient d6ja utilis6 des moustiquaires. Par comparaison avec un groupe t6moin, l'in- tervention a r6duit l'incidence du paludisme a Plasmodium falciparum et P. vivax de 61% et 42% respectivement. Entre 1992 et 1995, une serie d'etudes de suivi ont 6te r6alis6es pour 6valuer la fiabilit6 a long terme de cette technique pour la lutte antipaludique dans des communaut6s de r6fugi6s afghans dans la Province de la Frontiere du Nord- Ouest, au Pakistan. En 1992, 86% des proprietaires de moustiquaires se sont portes volontaires pour faire r6impr6gner celles-ci, et 15% des familles des deux villages ont achet6 des moustiquaires aux deux-tiers du prix coOtant. A partir de 1992, 27% des habitants sont retourn6s en Afghanistan, et des campagnes annuelles de pulverisation d'in- secticide dans les habitations ont 6t6 realis6es pour proteger les familles rest6es dans les villages de refugies et qui n'utilisaient pas de moustiquaires. En trois ans, ces campagnes, associ6es a l'utilisation des moustiquaires impr6gnees, ont r6duit l'incidence annuelle du paludisme de 87% (de 597 cas a 78 cas pour 1000 habitants). Neanmoins, 65% des families continuaient a faire impr6gner leurs moustiquaires une fois par an avec de la permethrine. Pour rechercher si les WHO Bulletin OMS. Vol 75 199728 Sustainability of pyrethroid-impregnated bednets moustiquaires etaient encore utilis6es et con- servaient leur efficacit6 protectrice avec un taux de transmission aussi r6duit, nous avons effectu6 une 6tude cas-t6moins en 1994 sur les malades f6briles ou pr6sentant d'autres sympt6mes vus dans les centres de sante des villages. Une comparaison de la positivit6 des lames chez les utilisateurs et non-utilisateurs de moustiquaires a montr6 que l'utilisation r6guliere de celles-ci r6duisait le risque relatif de contracter un paludisme a falciparum a 0,22 (intervalle de confiance a 95%: 0,09-0,55 et un paludisme a vivax a 0,31 (IC 95%: 0,19-0,51). Aucun biais lie au sexe ou a l'age n'est apparu dans l'utilisation des moustiquaires ni dans l'effet protecteur. Ces resultats indiquent qu'un pro- gramme communautaire d'utilisation des mousti- quaires impr6gn6es est une mesure appropri6e de lutte antipaludique dans les regions ou des problemes de gestion ou de s6curit6 rendent impossibles les campagnes traditionnelles de pul- v6risation dans les habitations, et que la couverture de ces programmes et la protection qu'ils conferent peuvent etre 6valu6es de fagon simple et exacte par un microscopiste de poste sanitaire. References 1. Curtis CF. Control of disease vectors in the commu- nity. London, Wolfe Publishing, 1991. 2. Rowland M et al. Pyrethroid-impregnated bed nets for self protection from malaria in Afghan refugees. Trans- actions of the Royal Society of Tropical Medicine and Hygiene, 1996, 90: 357-361. 3. Dean AG et al. Epi Info, version 5: a word processing, database, and statistics program for epidemiology. Stone Mountain, GA, USD Inc, 1990. 4. Rowland M, Hewitt S, Durrani N. Malaria prevalence surveys in refugee camps sprayed with malathion and lambdacyhalothrin. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1994, 88: 378- 379. 5. Alonso PL et al. A malaria control trial using insecticide-treated bed nets and targeted chemo- prophylaxis in a rural area of The Gambia, West Africa 6: The impact of the intervention on mortality and morbidity from malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1993, 87(suppl. 2): 37-44. 6. Beach RF et al. Effectiveness of permethrin- impregnated bed nets and curtains for malaria control in a holoendemic area of western Kenya. American journal of tropical medicine and hygiene, 1993, 49: 290-300. 7. Luxemburger C et al. Permethrin-impregnated bed nets for the prevention of malaria in schoolchildren on the Thai-Burmese border. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1994, 88: 155-159. 8. Reisen WK, Boreham FL. Estimates of malaria vectorial capacity for Anopheles culicifacies and A. stephensi in rural Punjab province, Pakistan. Journal of medical entomology, 1982, 19: 98-103. 9. Miller JE et al. Village trial of bed nets impregnated with wash-resistant permethrin compared with other pyrethroid formulations. Medical and veterinary ento- mology, 1995, 9: 43-49. WHO Bulletin OMS. Vol 75 1997 29