Dr Janet Sikasote, Konkola Copper Mines Plc, PO Box 2000, Chingola, Zambia. Fax:+260 2 353822; E-mail: email@example.com
Malaria is endemic in the whole of Zambia and is the leading cause of morbidity and mortality. Prior to 1980, effective malaria control was achieved in the northern mining towns of Chingola and Chililabombwe by means of annual residual spraying programmes. In the 1970s, incidence rates wereaslow as 20/1000 p.a., but by 2000 had increased to 68/1000 p.a. in Chingola and to 158/1000p.a.in Chililabombwe. Konkola Copper Mines (KCM) initiated a malaria control programme in which all dwellings in the two towns and within a 10-km radius were sprayed with either dichlorodiphenyltrichloroethane or a synthetic pyrethroid (Icon by ZENECA or Deltamethrin by Aventis). Houses were sprayed in November and December 2000, at the start of the peak transmission period. There was a statistically significant reduction in malaria incidence recorded at KCM health facilities in the two towns, representing a protective incidence rate ratio of 0.65 (95% CI 0.44, 0.97) when comparing the post-spraying period with the corresponding period of the previous 2 years. This reduction followed a single round of house spraying during a year with higher rainfall than the preceding two and in an area where chloroquine was first-line treatment. This house-spraying programme is an example of private/public sector collaboration in malaria control.
Malaria accounts for about 1 million deaths in Africa annually and has slowed economic growth in African countries by up to 1.3% per year (WHO 2000). It is endemic in the whole of Zambia and is the leading cause of morbidity and mortality (TDRC 1991).
Prior to 1970, malaria in urban areas in Zambia was kept to a minimum because of an effective prevention and control programme. Malaria incidence rates in Zambia have nearly tripled over the past 24 years (1976–2000). In1976, the incidence rate was 121.5 cases per 1000 population per year; by 1999, the incidence rate had increased to 321.4 per 1000 per year (Anon 2000). Before 1980, effective malaria control was achieved in Chingola and Chililabombwe by means of annual residual spraying programmes. In the 1970s, incidence rates were as low as 20/1000/year. In 2000, the incidence rates had increased to 68/1000/year in Chingola and 158/1000/year in Chililabombwe (KCM 2000).
Konkola Copper Mines (KCM) embarked on a malaria control programme in October 2000. The main thrust of this programme was vector control by means of intradomiciliary residual insecticide spraying. We report here on the malaria incidence rate following the first round of house spraying in comparison with the incidence prior to initiation of this vector control programme. Although this represents a small first step, it highlights an initiative taken by the private sector to mitigate the effect of malaria on their economic activities. The bigger challenge that lies ahead is the incorporation of such fragmented efforts into the greater framework of national–regional efforts to roll back malaria in Africa.
Before implementation of the spraying programme, we collected baseline information on the parasite prevalence rate, on knowledge, attitudes and practices (KAP) of the community pertaining to malaria and on malaria incidence rates in the previous 6 years, based on blood slide positivity.
Intensive publicity preceded the intradomiciliary spraying campaign, which was conducted in November and December 2000. Spraying of the inner walls of houses was carried out using Hudson Expert spray pumps. The residual insecticides, dichlorodiphenyltrichloroethane (DDT) and the synthetic pyrethroids, λ-cyhalothrin and Deltamethrin, were applied at concentrations of 2, 0.006 and 0.012 g/m2, respectively. Spray operators were recruited from the community and trained in spraying techniques, emphasizing avoidance of environmental contamination by strict adherence to protocols for the cleaning of equipment as well as the storage, distribution and disposal of insecticides. Representatives from the Zambian Environmental Council audited the programme as independent observers. Because of the urgency to initiate the programme, it was not possible to obtain environmental baseline information on DDT.
The programme covered the districts of Chingola and Chililabombwe (Figure 1). All dwellings in the two towns and within a 10-km radius of the furthest built up council area were sprayed, a total geographical area of 2704 km2. The type of house determined which insecticide was applied: DDT was used in traditional houses, as porous walls do not easily absorb it. Synthetic pyrethroids, which could be sprayed on painted walls and curtains without staining, were used in urban style houses.
Fourteen days before the house spraying a convenience sample of 1276 blood samples was taken from the community at randomly selected, sentinel sites and tested for malaria parasites using Hrp-2 antigen tests (ICTTM). Informed consent was a prerequisite to inclusion in the study. All individuals found positive for Plasmodium parasites were treated with chloroquine.
Malaria cases were recorded at Konkola Mine health facilities, four clinics in Chililabombwe and three in Chingola and at the referral hospital in each town. All malaria cases were definitively diagnosed using Giemsa stained thick blood smears. Malaria case data were available from 1996, as was the employee population using these health facilities. Chloroquine was the first-line treatment and sulphadoxine/pyrimethamine the second-line treatment.
Incidence rates represent symptomatic individuals who report to a mine health facility. Approximate incidence rates per month were calculated by dividing the total number of incident cases for each month by five times the average number of employees for a particular year. This ignored any fluctuation of the population during the year and assumed that each employee has an average of four dependants who also make use of the mine health facilities. Between 1994 and 2001, the annual dependency ratio varied by a maximum of 11%.
Figure 2 shows the mean monthly incidence rate after spraying as well as the mean for the 2 years prior to spraying. To determine any association between incidence rates and spraying of DDT, mean incidence rates were calculated for the period after spraying (January 2001–June 2001), and for the corresponding months of the previous 6 years. Mean incidence rates for the corresponding months during the 2 years before spraying were also calculated for comparison, as it was believed that these are more comparable, given the long-term secular effects of drug resistance and HIV infection. A Student's two-sample t-test was performed to assess the difference in mean incidence before and after spraying. However, this test assumes independent observations for the calculation of standard errors, which is unlikely given the serial nature of the data from the same sites. The incidence rate was therefore also modelled using generalized estimating equations. This method allows for serial correlation in the data and adjusts the SEs of the incidence rate ratio (comparing incidence rates pre- and post-spraying) accordingly.
Rainfall data for the two towns were available from the Mines Survey Department as monthly totals from 1990 to date.
Pre-intervention parasite prevalence study
Forty-six of the 1276 blood samples were malaria positive, giving an overall parasite prevalence rate of 3.6% (95% CI = 2.7–4.8). Parasite prevalence rate in under-fives was 5.0% (13/260) and 3.3% (33/1003) in over-fives. There was no significant difference between the two groups (P < 0.2).
Knowledge, attitude and practices (KAP) survey
A total of 171 questionnaires were completed and the findings were as follows: 83.6% (143) of the respondents were female and 16.4% (28) were male. The mean age was 35.5 years with a range of 17–72 years. Eighty-two per cent of the respondents could define malaria appropriately. Forty-three per cent thought drinking bad water causes malaria. Twenty-seven per cent claimed to use mosquito nets. Forty-three per cent had suffered from a malaria-like illness in the 6 months preceding the survey.
Residual spraying programme
Ninety-nine per cent (31 463) of the dwellings within the project area were sprayed, 20.5% (6438) with a pyrethroid and the remaining 79.5% (25 025) with DDT.
Figure 2 shows comparative incidence rates pre- and post-spraying. The reduction in transmission is more clearly reflected by the line representing the mean of cases reported in the preceding 2 years. Monthly incidence over the whole 7-year period varied from 1.3 to 18.9 per 1000 per month, with a median of 7.5 per 1000 per month (interquartile range 3.6–12.0). The mean incidence rate for the 6 months after house spraying was 8.3 cases per 1000 per month. The mean incidence rate during the corresponding 6 months of the 6 years before spraying (1995–2000) was 11.4 compared with a mean of 13.0 per 1000 per month for the corresponding months during the two preceding years (1999 and 2000). This represents a protective incidence rate ratio of 0.65 (95% CI 0.44, 0.97), comparing the post-spraying period with corresponding months of the previous 2 years. Annual rainfall ranged from 864 to 1775 mm from 1990 to July 2001 at Chililabombwe and from 984 to 1743 mm at Chingola.
The malaria incidence data from Chililabombwe and Chingola show tremendous interannual variation from 1994 to the present, with a distinct malaria transmission season. Malaria cases generally increase in December, peak in February and May and decline rapidly thereafter. The malaria transmission season follows the very distinct rainfall pattern for the area with rains starting in October/November and stopping in April, with the rainfall peak extending from December to March. House spraying should therefore ideally be carried out during October and November, preceding the peak seasonal increase in transmission.
The fact that there was no significant difference in the parasite prevalence ratio of the two age categories (under- and over-fives) suggests a non-immune population and an area of low transmission. There was a statistically significant reduction in incidence of malaria in the two towns after spraying in comparison with the previous two seasons. This is particularly significant considering that the rainfall in the 2000/2001 season was the second highest recorded in 10 years and 40% and 20% higher than the preceding 2 years in the two towns. The success of the programme can be attributed to a number of factors, which include
•High spray coverage (approaching 100%);
•co-operation from the community because of intensive information preceding the spraying campaign and the involvement of community members in the spraying;
•use of insecticides with proven efficacy.
The KAP survey indicates the need for further health education on malaria and its prevention. The 43% (cf. confirmed incidence) report of malaria type symptoms indicates the weakness of self reported clinical diagnosis and/or distorted perception of time periods by individuals.
The statistically significant reduction in malaria incidence following one house spraying round compares well with the results of similar programmes in the 1950s and 1960s (Fontaine et al. 1978; Bradley 1991; Curtis & Lines 2000) for which data is available. Based on these data, further reductions in malaria incidence following the next round of spraying can be expected. It should be borne in mind that chloroquine was the first-line treatment in the study area and that Plasmodium falciparum is chloroquine-resistant there, indicating the efficacy of malaria control by house spraying in the presence of a partially effective drug. The overall unit cost of the residual spraying programme was US$ 6 per dwelling which is cheaper than providing insecticide-treated bed nets to families which, on average, consist of five or more people.
The malaria control programme initiated and reported on here is an example of how collaboration between the private and public sector can benefit both the community and business. All structures were sprayed in the two towns, irrespective of whether the inhabitants were employed by KCM or not.
We are indebted to the following for their kind support and advice, without which this project would not have been possible: Konkola Copper Mines, who funded the project, The Ministry of Health, Central Board of Health, The Malaria Technical Commission, Medical Research Council of South Africa, Department of Health Mpumalanga Province, South Africa, District Health Management Teams and Municipal Councils of both Chililabombwe and Chingola, National Malaria Control Centre, Zambia, Zambia Information Services, DAPP Child Aid and Tropical Disease Research Centre, Zambia. Dr D le Sueur for critical comments on the manuscript.