Econ.2002.Utzinger.tmih.Economic Payoffs of Integrated Malaria Control in Zambian Copper Belt Between 1930 and 1950

8/14/2019 Econ.2002.Utzinger.tmih.Economic Payoffs of Integrated Malaria Control in Zambian Copper Belt Between 1930 a

1/21


2/21

than 100 years ago (Ross 1911; Watson 1921; MacDonald1950; Winslow 1951). However, Paul Russell stated asearly as 1959: questions about economic and socialimpact of malaria are frequently asked but accurate andauthoritative answers are difcult, indeed for the most partimpossible to formulate (cited in Packard 1997). Thecomplexity of this causal link and the lack of good-qualityepidemiological data precluded detailed appraisals of themacroeconomic impact of malaria. In this paper, weattempt to provide an in-depth response to Russellschallenge by drawing upon a substantial body of experi-mental and observational evidence accumulated acrossmultiple settings and by bringing together the epidemiol-ogy and control of malaria with the economic and socialsciences.

Recent efforts have been undertaken to estimate theextent of a malaria-related negative effect on economicgrowth. Using aggregated national health statistics andcontrolling for potential confounding factors, it was foundthat highly endemic countries (more than 50% of thepopulation living at risk of becoming infected withPlasmodium falciparum ) had average income levels thatwere one-third of those in non-malarious countries. Cross-country regressions over 25 years commencing in 1965conrmed these ndings; the annual growth rate of grossdomestic product (GDP) in countries of intense malariawas 1.3% lower than in those countries with less malaria(Gallup & Sachs 2001; Sachs & Malaney 2002). Using acomparable approach but a different data source andregression over a shorter time period, the estimated annualgrowth rate reduction attributable to malaria was consid-erably lower, namely 0.25% (McCarthy et al . 2000). Witha different approach using four case studies in differentepidemiological settings and extrapolating to all of sub-Saharan Africa the annual economic burden of malaria was estimated to be equivalent to 0.61.0% of theGDP (Shepard et al . 1991). Although the extent to whichmalaria hinders economic development will be furtherdebated and rened, the most recent estimates by Gallupand Sachs (2001) surprised the development communityand called for signicant up-scaling of malaria controlefforts (Sachs 2001; Vogel 2001; Sachs & Malaney 2002).It is important to note, however, that these associationalanalyses fail to assess the causal relationships and under-lying mechanisms of how malaria inhibits economicdevelopment.

A considerable amount of work in different ecological,epidemiological and socio-cultural settings has tried toassess the economic impact of malaria on the householdlevel (Sharma et al . 1990; Ettling & Shepard 1991;Sauerborn et al . 1991; Shepard et al . 1991; Mills 1993a;Guiguemde et al . 1994; Asenso-Okyere & Dyator 1997;

Attanayake et al . 2000). This approach provides high-quality local data with detailed information on direct(prevention and treatment) and indirect (lost production)economic costs related to clinical malaria attacks andpremature death. Most of these studies were conductedin rural communities mainly engaged in subsistenceagricultural activities.

A notable missing element is appraisal and discussion of the economic impact of malaria on the industrial andservice sectors of sub-Saharan Africa. These sectors areincreasing in importance and are likely to become thebackbone of many developing economies. Here, we presenta comprehensive analysis of integrated malaria controlprogrammes implemented in four copper mining commu-nities of former Northern Rhodesia that were sustained fortwo decades during the British colonial period. Weestimate direct and indirect costs of malaria control andassess the consequences for deaths, malaria attacks,DALYs and work shift losses averted. We argue thatconsistently dramatic decreases of these outcome andimpact measures across multiple settings, together withdocumentation of event sequencing (i.e. establishing tem-poral plausibility), build a compelling body of evidence insupport of a chain of steps linking malaria control tocopper output. Our ndings are, in fact, used to show theexplicit chain of relationships that connect malaria control per se with copper extraction productivity, revenuereceived, and impact on national income. Finally, wediscuss analogous situations in contemporary settingswhere implementing and sustaining integrated malariacontrol will be a crucial component of further social andeconomic development. The central point is that thehistorical experience in Northern Rhodesia has manyfeatures that are relevant for both policy makers and theresearch community today.

Materials and methods

Study area

The work reported here focuses on four copper miningcommunities of former Northern Rhodesia (now Zambia)during the English colonial period between 1929 and 1949.Copper had been discovered in Zambia more than100 years ago, but it was only in 1909 that the economicexploitation of this natural resource began, following thecompletion of a railway into the copperbelt. Until the mid1920s, outcrop copper deposits at Bwana Mkubwa andRoan Antelope were the most important mining sites(Mitchell 1961). However, at the time, Zambias copperproduction was insignicant compared with neighbouringDemocratic Republic of Congo (formerly Belgian Congo),

Tropical Medicine and International Health volume 7 no 8 pp 657677 august 2002

J. Utzinger et al . Economics of malaria in the Zambian copperbelt

658 2002 Blackwell Science Ltd


3/21

which extracted rich surface deposits, and was by far theleading producer in Africa.

After 1923, there was a sharp increase in the worlddemand for copper, accompanied by rising copper prices onthe world market. Many companies were more willing toinvest capital in extensive scientically based prospectingmethods and systematic boring technology (Mitchell 1961;Parpart 1983). This led to the discovery of rich copper bedsin Zambia (Mendelsohn 1961; Fleischer et al . 1976).Subsequently, four copper mining companies were inau-gurated; and they signicantly expanded their productionbetween 1929 and 1936 (Watson 1953; Parpart 1983).They are situated on the southern slope of the watershedbetween the Zambezi and Zaire drainage systems,approximately 1300 m above sea level. In late 1929, theRoan Antelope mine, located near the town of Luanshya inthe Ndola Rural district, was the rst to begin majoroperations. It became the second most important site interms of the number of employees. Mufulira, situated61 km north of the Roan Antelope mine and in closeproximity to the border of Democratic Republic of Congo,began operating in early 1930. Nkana-Kitwe, located27 km south of Mufulira in the town of Kitwe, wasestablished in February 1923. However, the mine onlybecame fully operational 8 years later. It rapidly grew andbecame the largest of the four companies. Nchanga, thesmallest of the four mining sites, is situated 45 km west of Mufulira and was opened in 1936 (Figure 1). Details of thegeology and morphology of the mining areas, as well as theestimated ore reserves of the Zambian copperbelt in generaland these four mining sites in particular, are providedelsewhere (Mendelsohn 1961; Fleischer et al . 1976).

Copper mining population

For the Roan Antelope mine, detailed population censusdata are available when copper production was launched in1929/1930, as well as 10 and 20 years later (Watson 1953).Based on these data, annual population estimates wereextrapolated, assuming constant growth rates for each of the two 10-year periods (see Utzinger et al . 2001). ForMufulira, census data are also available for 1940 and adecade later, however, no such data were collected at theonset of copper extraction (Watson 1953). As total popu-lation growth rates for the period 194050 were remark-ably similar for Roan Antelope and Mufulira, we assumedthat the annual growth rate estimated for Roan Antelopebetween 1930 and 1940 also occurred in Mufulira.

In the absence of detailed census statistics for Nkana-Kitwe and Nchanga, we estimated annual population sizesutilizing aggregated hospital in-patients records, which areavailable for all four mining communities for the period

194046 (Watson 1953). Specically, we estimated thepopulation size at Nkana-Kitwe by multiplying the knownpopulation size at Roan Antelope by the ratio of numbersof in-patients over 7 years at Nkana-Kitwe to the numberof in-patients over 7 years at Roan Antelope. We estimatedthe population size at Nchanga by the same method byusing number of in-patients over 7 years at Nchanga in thenumerator of the above in-patient ratio. That this methodof estimation has some validity is attested to by the factthat when we applied it to Mufulira, where we know thepopulation size in 1940 and 1949, the estimated and actualsizes approximate each other (e.g. known population sizeat Mufulira in 1940: 18 229; estimated population sizebased on hospital in-patient data: 19 898).

The total annual number of employees at all four miningsites, stratied by Africans and Europeans, is available from

-

Roan Antelope

Kilometers

0 10 20Copper mines

Zambia

N

Western

Northwestern

Southern

CentralLusaka

Luapula

Eastern

Northern

Kilometers0 100

Study areaRailway

Nchanga

Mufulira

NkanaKitwe

RoanAntelope

N

RailwayRoad

Copperbelt

Figure 1 Map of the Zambian copperbelt, indicating the locationsof Roan Antelope, Mufulira, Nkana-Kitwe and Nchanga, the fourmajor copper mining sites in the 1930s and 1940s.



2002 Blackwell Science Ltd 659


4/21

1932 onwards (Parpart 1983). In 1930 and 1931, weestimated the total work force and the proportion of Africans to Europeans by assuming the same employmentrate and the same ratio of Africans to Europeans as in 1932.

Copper production and revenue

Data on annual global copper production and annualcopper extraction in Zambia and neighbouring DemocraticRepublic of Congo between 1926 and 1950 were obtainedfrom statistical yearbooks initially prepared by the Leagueof Nations and subsequently by the United Nations(League of Nations 1936; United Nations 1951). Theannual revenue of the total copper extracted in Zambiaduring the period 193049 was calculated based on theannual production and the average annual copper pricespaid on the world market in New York. Average annualcopper prices were obtained from commodity yearbookspublished by the Commodity Research Bureau(Commodity Research Bureau 1940, 1951). We convertedthese annual revenues into 1995 US$, based on thepurchasing power of the dollar, derived from the USconsumer price index (US Census Bureau 1966, 1999).Finally, we calculated the cumulative economic return, alsoin 1995 US$, for the total amount of copper extracted inZambia over the 20-year period, starting in 1930.

Detailed annual income and expenditure statistics forZambia before and during the colonial era are virtuallynon-existent, which is probably also the case for mostother African countries south of the Sahara. However, weidentied one study that estimated the national income andexpenditures for Zambia in 1938. The same concepts andtechniques were applied for Zambia (which was thenBritish colonial territory) as used in Britain (Deane 1948).The national income was dened as the aggregated netvalue of all goods and services produced within the countryboundaries for the year 1938. Subsistence activities wereincluded, employing a system of hypothetical prices, basedon common market prices adjusted for transportation coststo the market. The total taxable national income wasestimated as the sum of subgroups of agricultural, manu-facturing and industrial activities and services, which wereall recorded in British pounds. We converted these taxablenational incomes into 1995 US$, using British historicalstatistics (Mitchell 1988) and the US consumer price index(US Census Bureau 1966, 1999).

Integrated malaria control

Early on in the decision process to extract Zambian copperresources on a large scale, members of the Roan Antelopeand Mufulira mining boards realized that effective tropical

disease control measures would be crucial for sustaininghealthy labour forces. This would be the foundation forsound economic development (Watson 1953). Malaria wasknown to be of particular importance, placing a heavyburden on the native population as well as immigrants andexpatriates. The high endemicity of the disease wasconrmed when the rst Europeans arrived at RoanAntelope. Early health records revealed that in a singlemonth there were 105 malaria attacks per 1000 people,counting only those cases which were treated at themedical department of the mine (Watson 1953). Whenmining activities began, malaria parasite rates and spleenindices among children from neighbouring villages rangedbetween 50 and 60% (Rodger 1944). Today, more than70 years later, a considerable body of clinical, epidemio-logical and entomological data has been accumulated,conrming that malaria was, and continues to be, highlyendemic in this part of Zambia (Friis-Hansen & McCul-lough 1961; Wenlock 1978; Snow et al . 1999). In the1920s and 1930s, advice was sought from the RossInstitute in London, which sent a delegation of leadingmalariologists and tropical sanitary engineers to thedesignated mining sites. Members of the delegation hadpreviously achieved outstanding success at malaria controlin the Malay States (now Malaysia). They had primarilyused environmental management interventions targetingthe larval stages of malaria vectors (Watson 1921).

Following the delegations recommendations, watersupply, sanitary facilities and housing conditions weregreatly improved. Basic hospital amenities were put inplace and run by trained personnel. Initial antimalarialmeasures consisted of house screening and, for part of themining communities, the use of mosquito nets andadministration of quinine for prophylaxis and treatment.Ongoing surveillance and monitoring, however, showedthat these interventions alone were insufcient to sub-stantially reduce the sickness gures, and the incidence of malaria remained high. Nonetheless, rapid diagnosis andtreatment of infected individuals was maintainedthroughout programme implementation, alongside healtheducation (Watson 1953).

An important feature of malaria control thereforeconsisted of a package of environmental managementinterventions that were designed and readily adapted to thelocal ecological settings. Interventions were primarilyaimed at modication or destruction of larval habitats of Anopheles gambiae and A. funestus , the chief malariavectors identied at the start of the programmes through aseries of systematic entomological surveys. Briey, theyconsisted of vegetation clearance, swamp drainage andriver boundary modication (details in Utzinger et al .2001). An additional component, although less





5/21

environmentally sound, was regular application of oil to allopen water bodies. Environmental management interven-tions were launched in late 1929 and adaptively tuned andmaintained for two decades. It is remarkable that thesemeasures proved particularly successful against the larvalstages of A. gambiae , while modications of larval habitatsof A. funestus were more challenging. This is conrmed bythe consistently higher adult catches of the latter vectorspecies at several monitoring stations within the controlarea throughout the programmes (Watson 1953; Utzingeret al . 2001). Most importantly, malaria incidence ratesdecreased sharply. During the last 4 years of the pro-grammes, residual house spraying with dichlorodiphenyl-trichloroethane (DDT) became the nal, and widelyapplied, intervention tool. It is important to observe thatDDT was an addition to the previous package of inter-ventions, never a substitute for them.

Similar interventions built around environmental man-agement were implemented at Mufulira, Nkana-Kitwe andNchanga between 1930 and 1936. At Nchanga, the initialprogress of control measures was somewhat slower than inthe other mining sites, and interventions stagnated duringthe war years in the early 1940s.

Cost of integrated malaria control

Annual implementation and maintenance costs of malariacontrol measures at the Roan Antelope mine are availablefor the entire 20-year period (Watson 1953). For theMufulira mine, annual maintenance costs from 1933onwards are also available from the published programmebudget. As the extent of control areas, the package of interventions and the total maintenance costs for 193349were virtually identical for the Mufulira and Roan Ante-lope companies, we assumed that the capital investment in1930 and the maintenance costs for the rst 2 years wereequal at the two sites. In both settings, physical resourcesand unit prices were separately recorded for each inter-vention, and the accounting system remained xedthroughout the programme. Costs of integrated malariacontrol at the Nkana-Kitwe and Nchanga mining com-panies are more fragmentary than for the Roan Antelopeand Mufulira mines. Hence, several assumptions had to bemade. First, we assumed that capital investment mainlyemployed for the initial drainage, river boundary modi-cation and vegetation clearance was equal in all fourmining sites. Secondly, the cumulative maintenance cost atNkana-Kitwe and Nchanga for the years 1945 and 1949indicated that they were comparable with the cumulativemaintenance costs at Roan Antelope and Mufulira for thesame 2 years. Therefore, annual maintenance costs atNkana-Kitwe and Nchanga were estimated on the basis of

the annual maintenance costs at Roan Antelope andMufulira. Thirdly, in 1949, the costs for maintainingmalaria control interventions at Nkana-Kitwe were 8%higher than at Nchanga. We assumed that the annualmaintenance costs in the previous years were also 8%higher at Nkana-Kitwe than at Nchanga.

Direct treatment costs for clinical malaria episodes werealso estimated. As the detailed control programme budgetsnever specied any costs for malarial treatment (Watson1953), we assumed that the costs were directly borne bythe mining employees and their families. The total annualcosts were calculated by multiplying the number of malariaattacks at each mining site within 1 year with the meantreatment costs per malaria attack. The number of malariaattacks per year was estimated by multiplying the meanannual incidence rate with the estimated annual populationat risk of malaria. Detailed clinical records from thehospital at the Roan Antelope mine suggested that aneffective system of rapid diagnosis and treatment withquinine, following 910-day regimens, was in place(Rodger 1944). In the absence of the actual treatment costsfor a single malaria episode at the time of programmeimplementation, we used an estimated mean cost of US$2.22 (in 1995 US$). This value is the translation of US$1.87 (in 1987 US$), provided by Shepard et al . (1991), andis based on cost analyses carried out in four differentepidemiological settings across sub-Saharan Africa in the1980s.

Finally, we estimated total indirect costs due to workshift losses because of malaria. We estimated the number of work shifts lost per clinical malarial attack based on thetotal annual work shifts lost for a representative sample of the total labour force at the Roan Antelope mine for 194449. The total income lost because of malaria-related workabsenteeism was estimated for the active labour force ateach mine. Labour wages were stratied for Europeans andAfricans as they differed by more than an order of magnitude. Although data on wages for mining employeesduring the colonial era are scarce, we found historicalaccounts indicating that Europeans earned between 1.375(profession: miner) and 1.625 (platelayer) per day in 1930(Anonymous 1931). This corresponds to a daily wage of US$ 5565 (in 1995 US$), and we used an average of US$60 for further analyses. We also assumed that Europeans,mainly engaged in skilled labour, earned a monthly salaryand that they were also paid during illness episodes.

The average wage of an African employee for 30working shifts at Roan Antelope in 1930 was 1.05 forsurface and 1.65 for underground work (Parpart 1983).The average annual earning of an African employee in theyear 1946 was 30 (Arrighi 1973). These wages convert toUS$ 511802 per year (in 1995 US$). For further analyses





6/21

we used an average annual wage of US$ 700. There areclear indications that African mine workers were not paidwhen they suffered clinical malarial attacks and could notwork in the mines. However, the wages African employeeswere paid were considerably higher than what they wouldhave gained with subsistence farming.

These various cost estimates were all discounted to allowfor time preferences. We used a discount rate of 3%, whichis currently considered to be the standard rate in contem-porary cost-effectiveness analyses (Gold et al . 1996). Ourestimates facilitated calculation of annual and cumulativedirect and indirect costs for malaria control by the miningcompanies and the employees and their families for theentire 20-year period.

Consequences of integrated malaria control

We assessed the consequences of the integrated malariacontrol programmes by calculating the number of averteddeaths, malaria attacks, DALYs and work shift losses andrespective costs. For Roan Antelope, malaria-specicmortality rates are available for Europeans at baseline(1929/30) and for the periods 193238 and 193843(Rodger 1944; Watson 1953). We assumed that the latterrates remained constant until the end of interventions in1949. In the absence of detailed malaria-specic mortalityrates for the other mining communities, we assumed deathrates equal to those at Roan Antelope, both at baseline andover the course of programme implementation. Deathsaverted were estimated by the reduction in the malaria-specic mortality rates (comparing the rates before theprogrammes started with those measured during theimplementation and maintenance phase) multiplied by thetotal person life years at risk.

The mean annual malaria incidence rates at the fourmining companies facilitated estimation of reduction inincidence rates during the course of programme imple-mentation. The data indicate that the initial package of integrated malaria control interventions reduced the base-line malaria incidence rate by 5075% in the rst 3 years.This rate was sustained until the mid-1940s. Indoorresidual spraying with DDT resulted in another sharpdecline in the annual malaria incidence rates (Watson1953; Utzinger et al . 2001). We estimated the number of malarial attacks averted by comparing incidence rates priorand after implementation of integrated malaria controlmeasures, multiplied by the total population life years atrisk. Thus, without interventions, it was assumed that thebaseline incidence rate would have remained the same forthe next 20 years. This was compared with what actuallyhappened based on the mean annual incidence rates duringthe course of programme implementation.

Estimation of DALYs averted followed the methodologypresented in the Global Burden of Disease study (Murray &Lopez 1996). We adapted this technique for the miningcommunities studied here, as described in detail elsewhere(Utzinger et al . 2001). In brief, we stratied the minepopulations into three age groups (04, 515, > 15 years),according to population percentages given by Snow et al .(1999) for communities living in areas of stable malariatransmission. We assumed equal mortality and malariaincidence rates for Africans and Europeans, which isjustied by careful analyses of the original data recordscollected at Roan Antelope (Watson 1953; Utzinger et al .2001). Furthermore, we assumed a life expectancy at birthof 50 years, employing a West African model life table(United Nations 1982). We used a discount rate of 3%,with no age-weights (Murray & Lopez 1996). Malaria-specic mortality and malaria incidence rates before andafter programme implementations were the bases for DALYcalculations. We used age-specic proportions according toSnow et al . (1999). In the absence of data on neurologicalsequelae and anaemia, we only considered clinical cases of malaria for disability calculations. This is justied, as themining companies paid great attention to rapid malariadiagnosis and treatment with quinine. At the time of programme implementation, this drug was highly effective.Thus, occurrence of neurological sequelae and anaemiawere likely very rare. We used setting-specic disabilitydurations for a single malaria episode. Finally, DALYs wereexpressed in 3-year incremental periods as this correspondsto the short duration of cost-effectiveness analyses of contemporary malaria control programmes and allowscomparison with other studies (Utzinger et al . 2001).

Sensitivity analyses

We performed a series of one-way sensitivity analyses forthose parameters where inherent uncertainties wereattached. We followed a similar analytical approach tothose used in recent evaluations of costs, consequences andnet cost-effectiveness of malaria control programmes(Aikins et al . 1998; Goodman et al . 2001). The parameterstaken into account in our sensitivity analyses were directtreatment costs, the indirect costs because of work shiftlosses, the discount rate and the package of interventions.Based on the existing literature examining the direct treat-ment costs fora singlemalaria attackacrossmultiplesettingsin sub-Saharan Africa, we used a range of US$ 1.005.00(Shepard et al . 1991; Asenso-Okyere & Dyator 1997). Forestimation of indirect costs because of malaria-related workabsenteeism, we considered minimum and maximum labourwages, stratied for African and European employees.Wages were derived from the historical literature pertaining





7/21

to the Northern Rhodesian copperbelt (Anonymous 1931;Arrighi 1973; Parpart 1983). We assessed the impact of different discount rates by either decreasing or augmentingthe standard rate of 3% by 1 percentage point. Finally, weestimated costs, consequences and net cost-effectivenessratios of the malaria control programmes between1930 and1945, prior to the use of DDT as an additional controlintervention.

Connection between malaria control, labour force sizeand macroeconomy

We developed the following approach to facilitate apprai-sal of the macroeconomic impact of integrated malariacontrol in copper mining communities of the NorthernRhodesian copperbelt. First, in order to exploit largecopper deposits and increasing annual productivity andrevenues, access to a large reservoir of cheap labour wasnecessary. Labour migration in the 1920s was nothingnew. In fact, Northern Rhodesia served as a reserve formines in neighbouring Belgian Congo and southernRhodesia (Parpart 1983; Ferguson 1999). Secondly, whencompanies decided to seriously invest in northern Rhodesiato signicantly increase copper production, the promotionof rapid and sustained in-migration became of centralimportance. Indeed, there is clear documentation that anunsuccessful attempt on malaria control at Roan Antelopemine prior to 1929 resulted in migrant workers abandon-ing the site (Watson 1953). Thirdly, it was necessary tokeep the number of work shifts lost at a low level.Fourthly, experience with the implementation and adaptivetuning of the malaria control measures revealed that theprogrammes displayed the desired outcomes within about3 years. In delineating a plausible counterfactual situationfor copper production in the absence of a sustainedeffective control programme, we used the actual controlprogramme results and associated in-migration and copperproduction data only through 1933. We then assumed thatwithout sustained and effective malaria control in place,there would have been only marginal in-migration there-after; hence, the total work force would remain relativelyconstant. Consequently, copper production over the nextseveral years would stay at the 1933 level. In theprogramme, effective malaria control led to diffusion of information among potential migrant workers about safeemployment opportunity at the mines. Mine workers hadeffective networks for information exchange about livingand working conditions and news circulated rapidly alongthe main labour routes (Parpart 1983). The subsequentlarge in-migration facilitated enhanced copper productiv-ity, and revenues continued to increase. We then comparedthe percentage of national income attributable to mining

(essentially derived from the four mines in this study) forthe year 1938 when malaria was successfully undercontrol with the estimated counterfactual revenues andtheir percentage contribution to the total national incomeunder the scenario of no effective malaria control.

Results

Copper mining communities

Annual population estimates for the four mining commu-nities for 193049 are presented in Table 1. When extensivecopper extraction started at Roan Antelope, the estimatedpopulation was 6067, consisting of 5000 Africans (82.4%)and 1067 Europeans (Watson 1953). At Mufulira, theinitial population was estimated at 4897. One year later,Nkana-Kitwe started operation with an estimated popula-tion size of 8559. Finally, copper extraction commenced atNchanga in 1936 with a population estimated at 5598. In1940, the estimated total population living on the fourmining sites was 77 872. Census data from Roan Antelopeand Mufulira revealed that 91.8% of them were Africans,indicating that the proportion of Europeans was halvedover the rst 10-year period of copper extraction. Onedecade later, the total population residing in the four miningsites further increased by 80%, reaching an estimated size of 140 368. This dramatic increase is mainly because of sustained in-migration (Parpart 1983; Ferguson 1999).Census data from the Roan Antelope and Mufulira mines atthe end of December 1949 revealed that the proportion of Europeans had increased from 8.2 to 9.9%.

The number of African and European employees and thetotal annual work force at all four mines from 1932onwards were derived from Parpart (1983) and are shownin Table 2. The initial estimated work force in 1932 was6465. Within 20 years it grew to 37 354, an increase by afactor 5.8. The mean proportion of employees to the totalestimated population was 32.3%. The percentage of European employees varied between 9.8 and 13.8%, withan estimated mean of 10.8%. As expected in miningcommunities, the majority of the population were adultmales. Census data at the Roan Antelope mine in 1949revealed a sex-ratio between male and female of 1.9. Theproportion of children under the age of 15 years wasconsiderably smaller than what would be expected at thattime for other communities in sub-Saharan Africa.

Copper extraction in Northern Rhodesiaduring the colonial period

In the second half of the 1920s, Northern Rhodesiacontributed less than 5% of the African, and about 0.3%





8/21

to the annual global copper extraction. At that time,Belgian Congo was by far the leading African producer,accounting for approximately 90% of the extraction onthe continent. Belgian Congo controlled more than 7% of the world copper ore market. After the discovery of largecopper deposits in Northern Rhodesia, and the decisionby the British government to exploit them on a largescale, three copper companies started operation betweenlate 1929 and 1931. Consequently, the annual copperextraction increased exponentially. While Northern Rho-desia produced 9100 metric tons in 1931, there was a7.5-fold increase to 69 000 metric tons in the followingyear. Virtually overnight, the country became the leadingcopper producer in Africa (Figure 2). Within the next8 years, and complemented by the inauguration of theNchanga mining company, there was a 3.9-fold increasein annual copper extraction to 266 600 metric tons in1940.

Northern Rhodesia was then the third most importantcopper ore producer worldwide, with a global share of more than 11% (Table 3). The United States and Chileranked rst and second, and Northern Rhodesia movedahead of Canada and Japan. During World War IIand particularly the early postwar years, the annualcopper production in Northern Rhodesia dropped con-siderably. Nevertheless, the country continued to supply

Population estimates

Year Roan Antelope Mufulira Nkana-Kitwe Nchanga Total population

1930 6067 4897 10 9641931 6919 5584 8559 21 0621932 7891 6369 9762 24 0221933 9000 7264 11 134 27 3981934 10 264 8284 12 697 31 2451935 11 706 9448 14 481 35 6351936 13 351 10 775 16 516 5598 46 2401937 15 226 12 289 18 836 6385 52 7361938 17 365 14 015 21 482 7282 60 1441939 19 804 15 984 24 499 8305 68 5921940 22 586 18 229 27 586 9471 77 8721941 24 100 19 330 29 814 10 106 83 3501942 25 717 20 497 31 814 10 784 88 8121943 27 441 21 735 33 947 11 507 94 6301944 29 281 23 048 36 223 12 279 100 831

1945 31 244 24 440 38 651 13 102 107 4371946 33 339 25 916 41 243 13 981 114 4791947 35 575 27 481 44 009 14 918 121 9831948 37 960 30 901 46 960 15 918 131 7391949 40 506 32 767 50 109 16 986 140 368

Table 1 Annual population estimates forRoan Antelope, Mufulira, Nkana-Kitweand Nchanga mines in the Zambiancopperbelt between 1930 and 1949

Table 2 Number (percentage) of African and European miningemployees and total work force engaged in copper extraction inZambia between 1932 and 1949 (source: Parpart 1983)

Mining employees

Year* Africans Europeans Total work force

1932 5572 (86.2) 893 (13.8) 64651933 7190 (87.5) 1026 (12.5) 82161934 13 808 (88.9) 1729 (11.1) 15 5371935 13 224 (88.3) 1758 (11.7) 14 9821936 11 957 (88.4) 1575 (11.6) 13 5321937 17 926 (89.8) 2037 (10.2) 19 9631938 20 358 (89.9) 2296 (10.1) 22 6541939 20 924 (88.9) 2609 (11.1) 23 5331940 24 382 (89.1) 2971 (10.9) 27 3531941 27 270 (89.8) 3098 (10.2) 30 3681942 30 425 (90.2) 3306 (9.8) 33 7311943 32 805 (90.2) 3566 (9.8) 36 3711944 30 470 (89.8) 3445 (10.2) 33 9151945 28 304 (89.6) 3272 (10.4) 31 5761946 27 832 (89.0) 3426 (11.0) 31 2581947 29 166 (88.8) 3681 (11.2) 32 8471948 30 932 (88.7) 3958 (11.3) 34 8901949 33 061 (88.5) 4293 (11.5) 37 354

* No reliable counts of mining employees are available for theyears 1930 and 1931, although there are some indications that thetotal work force might have been considerably higher than in1932.





9/21

approximately 911% of the global annual copperdemand.

Average annual copper prices uctuated signicantlybetween 1930 and 1949, as shown in Table 3. Shortly aftercopper production was intensied in Northern Rhodesia,

and in conjunction with a major global recession, theaverage price fell sharply from 28.90 US cents/kg in 1930to 12.50 US cents/kg 2 years later a reduction of morethan 50%. However, as a result of increasing annualextraction, the revenues rose steadily between 1931 and1937. Converted into 1995 US$, the annual monetaryvalue of the Northern Rhodesian copper extracted in 1937was more than half a billion US$. It rose to US$ 613million in 1940. Copper prices remained constantthroughout the war years and increased considerably in therst postwar years. The cumulative revenue between 1930and 1949 from copper was US$ 7.1 billion (in 1995 US$).

Analysis of the Northern Rhodesian income and expen-diture statistics for the year 1938 clearly revealed thatcopper production represented the dominant share of thecountrys economy. Converted into 1995 US$, the taxablenational income because of mining was 329 milliondollars, or 55% of the total taxable national income. Thisgure is somewhat smaller than the estimated annualmonetary value of copper extraction in Northern Rhodesiain the same year US$ 440.8 million (Table 3). The

C o p p e r p r o

d u c

t i o n

( 1 0 3

m e

t r i c t o n s

)

0

50

100

150

200

250

300

1926 1930 1934 1938 1942 1946 1950

Figure 2 Annual copper extraction in Zambia ( j ) andDemocratic Republic of Congo ( s ) during the colonial periodbetween 1926 and 1950 (source: League of Nations 1936; UnitedNations 1951).

Table 3 Annual global and Zambian copper production from 1930 to 1949 and economic revenues over this 20-year period expressed in1995 US$

Copper production(in 1000 metric tons)*

Revenue(in 1995 US$)

Year WorldZambia(% world)

Copper price(cents/kg)

Annual revenue(US$)

Purchasing power(1995 US$ 1.00) Annual Cumulative

1930 1530 7.7 (0.50) 28.90 2225 500 8.335 18 549 545 18 549 5451931 1330 9.1 (0.68) 18.17 1653 114 9.864 16 306 318 34 855 8631932 880 69.0 (7.84) 12.50 8625 145 11.103 95 764 980 130 620 8431933 990 105.9 (10.70) 15.76 16 693 071 10.920 182 288 332 312 909 1751934 1250 140.1 (11.21) 18.81 26 346 407 9.597 252 846 468 565 755 6431935 1450 145.8 (10.06) 19.31 28 157 615 8.990 253 136 956 818 892 5991936 1630 144.6 (8.87) 21.12 30 539 930 8.905 271 958 074 1 090 850 6731937 2230 211.5 (9.48) 29.26 61 875 045 8.335 515 728 497 1 606 579 1701938 1910 216.4 (11.33) 22.27 48 185 111 9.148 440 797 393 2 047 376 5631939 2030 215.1 (10.60) 24.41 52 495 522 9.331 489 835 720 2 537 212 2831940 2330 266.6 (11.44) 25.13 67 003 768 9.148 612 950 471 3 150 162 7541941 2470 232.0 (9.39) 26.17 60 711 776 8.232 499 779 337 3 649 942 0911942 2620 250.6 (9.56) 26.17 65 579 185 7.283 477 613 206 4 127 555 2971943 2590 255.0 (9.85) 26.17 66 730 615 6.979 465 712 965 4 593 268 262

1944 2420 224.4 (9.27) 26.17 58 722 942 6.914 406 010 418 4 999 278 6801945 2030 197.1 (9.71) 26.17 51 578 840 6.792 350 323 484 5 349 602 1641946 1680 185.2 (11.02) 30.69 56 834 815 5.942 337 712 469 5 687 314 6331947 2060 195.6 (9.50) 46.63 91 203 909 4.848 442 156 551 6 129 471 1841948 2120 217.0 (10.24) 48.94 106 205 483 4.478 475 588 154 6 605 059 3381949 2070 263.5 (12.73) 42.68 112 465 729 4.712 529 938 516 7 134 997 854

* Source: United Nations Statistical Yearbook in 1951 (United Nations 1951).Source: Commodity Yearbooks (Commodity Research Bureau 1940, 1951).

Conversion of annual revenues into 1995 US$ based on the purchasing power of the US dollar, derived from US consumerprice indices (US Census Bureau 1966, 1999).





10/21


11/21

costs averted. Because of the integrated control measures

being so effective in terms of reduction of incidence rateswithin the rst 35 years of programme implementationand maintaining them at these lower levels, the costsaverted gradually increased. Wide-scale application of DDT in the last 4 years of the malaria control programmeshad a dramatic effect on malaria incidence rates. This ledto a considerable increase in the total costs averted.Overall, the cumulative costs borne by mine employees andtheir families over the course of the malaria controlprogramme were US$ 337 525 (in 1995 US$). In the

absence of integrated control, these costs would have been

signicantly higher, namely US$ 1 134 147. The essentialpoint is that by implementing simultaneously a multiplicityof interventions that interfere with one or more compo-nents of the transmission system, there is a reduction in thedemand for any one tool (e.g. antimalarial drugs).

Based on a representative sample of the work force at theRoan Antelope mine, and the total number of work shiftslost by these employees during 194449, the meanduration of work shifts lost could be estimated. Onaverage, an employee suffering from a clinical malaria

Table 5 Direct treatment costs borne by the mining employees and their families with and without malaria control measures. All costswere discounted by 3% and are expressed in 1995 US$

Direct treatment costs

Without malaria control With malaria control Costs averted

YearRoanAntelope Mufulira

Nkana-Kitwe Nchanga

RoanAntelope Mufulira

Nkana-Kitwe Nchanga

RoanAntelope Mufulira

Nkana-Kitwe Nchanga

1930 6923 5588 6923 5588 0 0 1931 7665 6186 8683 5786 4778 8683 1879 1408 0 1932 8487 6850 9616 4310 3745 8164 4178 3105 1452 1933 9398 7585 10 648 2468 2420 7456 6930 5165 3192 1934 10 406 8399 11 788 2733 2680 6474 7673 5719 5314 1935 11 522 9300 13 053 3026 2967 7169 8496 6332 5884 1936 12 759 10 297 16 257 6388 3351 3285 7939 6388 9408 7012 8318 01937 14 127 11 402 16 005 7074 3710 3638 8790 5436 10 417 7764 7214 16381938 15 642 12 624 17 721 7832 4108 4028 9733 4190 11 534 8596 7988 36421939 17 319 13 978 19 621 8672 4549 4460 10 776 2632 12 770 9518 8844 6040

1940 19 177 15 478 21 449 9602 5037 4938 11 781 2914 14 140 10 539 9669 66881941 19 866 15 934 22 507 9947 5218 5084 12 362 3019 14 648 10 850 10 145 69281942 20 583 16 405 23 317 10 306 5406 5234 12 806 3128 15 177 11 171 10 510 71781943 21 324 16 890 24 157 10 676 5601 5389 13 268 3240 15 723 11 501 10 889 74361944 22 089 17 387 25 027 11 061 5286 6630 5969 3357 16 803 10 757 19 057 77041945 22 885 17 901 25 924 11 458 5476 6826 6183 3478 17 409 11 075 19 741 79801946 23 708 18 429 26 859 11 871 969 1112 1352 693 22 739 17 318 25 507 11 1781947 24 559 18 972 27 825 12 297 1003 1144 1400 718 23 556 17 827 26 425 11 5791948 25 443 20 712 28 824 12 740 1040 1249 1451 744 24 404 19 463 27 373 11 9961949 26 360 21 323 29 862 13 199 1077 1286 1503 770 25 283 20 037 28 359 12 429

Total 340 242 271 641 379 141 143 124 77 077 76 482 143 259 40 707 263 165 195 158 235 882 102 417

Table 6 Number of work shifts lost in relation to the mean annual malaria incidence rate for a representative sample of employees at theRoan Antelope mine between 1944 and 1949

Year No. of employees Malaria incidence (per 1000) Malaria attacks Work shifts lost Days lost per malaria attack

1944 2513 105.45 265.0 1580 6.01945 2483 140.50 348.9 2030 5.81946 2466 39.60 97.7 530 5.41947 2633 15.38 40.5 280 6.91948 2971 13.72 40.8 250 6.11949 3366 16.04 54.0 270 5.0

Total 16 432 846.8 4940 5.8





12/21

T a

b l e 7 I n d i r e c t c o s t s b o r n e b y t h e w o r k e r s a n d t h e i r f a m i l i e s ( f o r A f r i c a n e m p l o y e e s ) a n d t h e m i n i n g c o m p a n i e s ( f o r E u r o p e a n e m p l o y e e s ) f o r w o r k t i m e l o s t a s a r e s u l t o f

c l i n i c a l m a l a r i a a t t a c k s . A l l c o s t s w e r e d i s c o u n t e d b y 3 % a n d a r e e x p r e s s e d i n 1 9 9 5 U S $

D i r e c t t r e a t m e n t c o s t s

W i t h o u t m a l a r i a c o n t r o l

W i t h m a l a r i a c o n t r o l

C o s t s a v e r t e d

Y e a r

R o a n

A n t e l o p e

M u f u l i r a

N k a n a -

K i t w e

N c h a n g a

R o a n

A n t e l o p e

M u f u l i r a

N k a n a -

K i t w e

N c h a n g a

R o a n

A n t e l o p e

M u f u l i r a

N k a n a -

K i t w e

N c h a n g a

1 9 3 0

5 0 1 3 1

4 0 4 6 4

5 0 1 3 1

4 0 4 6 4

0

0

1 9 3 1

5 5 5 0 8

4 4 7 9 8

6 2 8 8 0

4 1 9 0 1

3 4 6 0 0

6 2 8 8 0

1 3 6 0 7

1 0 1 9 7

0

1 9 3 2

6 1 4 6 0

4 9 6 0 6

6 9 6 3 1

3 1 2 0 8

2 7 1 1 9

5 9 1 1 8

3 0 2 5 2

2 2 4 8 7

1 0 5 1 3

1 9 3 3

7 0 2 0 2

5 6 6 6 1

7 9 5 3 7

1 8 4 3 8

1 8 0 7 9

5 5 6 9 3

5 1 7 6 4

3 8 5 8 2

2 3 8 4 4

1 9 3 4

1 1 8 3 0 2

9 5 4 8 1

1 3 4 0 1 1

3 1 0 7 1

3 0 4 6 5

7 3 6 0 3

8 7 2 3 0

6 5 0 1 6

6 0 4 0 8

1 9 3 5

1 1 5 1 7 1

9 2 9 5 5

1 3 0 4 7 7

3 0 2 4 9

2 9 6 5 9

7 1 6 6 2

8 4 9 2 2

6 3 2 9 6

5 8 8 1 5

1 9 3 6

8 8 2 3 7

7 1 2 1 2

9 9 9 5 9

4 4 1 7 6

2 3 1 7 5

2 2 7 2 1

5 4 9 0 1

4 4 1 7 6

6 5 0 6 2

4 8 4 9 1

4 5 0 5 8

0

1 9 3 7

1 1 4 8 1 1

9 2 6 6 5

1 3 0 0 7 1

5 7 4 8 9

3 0 1 5 5

2 9 5 6 6

7 1 4 3 9

4 4 1 8 0

8 4 6 5 6

6 3 0 9 8

5 8 6 3 1

1 3 3 1 0

1 9 3 8

1 2 5 8 8 0

1 0 1 5 9 6

1 4 2 6 1 2

6 3 0 3 2

3 3 0 6 2

3 2 4 1 6

7 8 3 2 8

3 3 7 2 4

9 2 8 1 8

6 9 1 8 0

6 4 2 8 5

2 9 3 0 9

1 9 3 9

1 3 5 4 7 1

1 0 9 3 4 0

1 5 3 4 7 5

6 7 8 3 4

3 5 5 8 1

3 4 8 8 7

8 4 2 9 4

2 0 5 8 8

9 9 8 9 0

7 4 4 5 3

6 9 1 8 1

4 7 2 4 6

1 9 4 0

1 5 1 2 9 6

1 2 2 1 1 0

1 6 9 2 2 1

7 5 7 5 5

3 9 7 3 7

3 8 9 6 1

9 2 9 4 2

2 2 9 9 2

1 1 1 5 5 9

8 3 1 4 9

7 6 2 7 9

5 2 7 6 3

1 9 4 1

1 5 5 3 7 0

1 2 4 6 1 9

1 7 6 0 2 7

7 7 7 9 7

4 0 8 0 7

3 9 7 6 2

9 6 6 8 0

2 3 6 1 2

1 1 4 5 6 3

8 4 8 5 7

7 9 3 4 7

5 4 1 8 5

1 9 4 2

1 6 3 0 8 9

1 2 9 9 8 6

1 8 4 7 5 2

8 1 6 5 9

4 2 8 3 5

4 1 4 7 4

1 0 1 4 7 2

2 4 7 8 4

1 2 0 2 5 4

8 8 5 1 2

8 3 2 8 0

5 6 8 7 5

1 9 4 3

1 7 1 0 2 6

1 3 5 4 6 4

1 9 3 7 4 8

8 5 6 2 9

4 4 9 1 9

4 3 2 2 2

1 0 6 4 1 3

2 5 9 8 9

1 2 6 1 0 7

9 2 2 4 2

8 7 3 3 5

5 9 6 4 1

1 9 4 4

1 5 8 9 9 3

1 2 5 1 4 8

1 8 0 1 3 9

7 9 6 1 3

3 8 0 4 7

4 7 7 2 2

4 2 9 6 5

2 4 1 6 3

1 2 0 9 4 6

7 7 4 2 6

1 3 7 1 7 4

5 5 4 5 0

1 9 4 5

1 4 6 0 0 7

1 1 4 2 1 1

1 6 5 3 9 4

7 3 1 0 3

3 4 9 3 9

4 3 5 5 1

3 9 4 4 9

2 2 1 8 7

1 1 1 0 6 7

7 0 6 6 0

1 2 5 9 4 6

5 0 9 1 6

1 9 4 6

1 4 6 3 5 7

1 1 3 7 7 1

1 6 5 8 0 8

7 3 2 8 3

5 9 8 0

6 8 6 2

8 3 4 5

4 2 7 7

1 4 0 3 7 8

1 0 6 9 0 9

1 5 7 4 6 3

6 9 0 0 6

1 9 4 7

1 5 1 9 7 1

1 1 7 3 9 5

1 7 2 1 8 0

7 6 0 9 4

6 2 0 9

7 0 8 0

8 6 6 6

4 4 4 1

1 4 5 7 6 2

1 1 0 3 1 4

1 6 3 5 1 4

7 1 6 5 2

1 9 4 8

1 5 6 2 4 8

1 2 7 1 9 2

1 7 7 0 0 7

7 8 2 3 6

6 3 8 4

7 6 7 1

8 9 0 8

4 5 6 6

1 4 9 8 6 4

1 1 9 5 2 1

1 6 8 0 9 9

7 3 6 7 0

1 9 4 9

1 6 4 2 2 2

1 3 2 8 4 6

1 8 6 0 4 2

8 2 2 3 3

6 7 0 9

8 0 1 2

9 3 6 3

4 8 0 0

1 5 7 5 1 2

1 2 4 8 3 4

1 7 6 6 7 9

7 7 4 3 3

T o t a l

2 4 9 9 7 5 3

1 9 9 7

5 1 7

2 7 7 2 9 6 9

1 0 1 5 9 3 3

5 9 1 5 3 9

5 8 4 2 9 2

1 1 2 7 1 2 0 3 0 4 4 7 6

1 9 0 8 2 1 4

1 4 1 3 2 2 5

1 6 4 5 8 4 9

7 1 1 4 5 7





13/21

attack missed 5.8 working days (Table 6). During thisperiod there was no contribution to copper production.The indirect costs for mining employees are presented inTable 7. We assumed that the annum-specic proportionof African to European employees in the Northern Rho-desian copperbelt (Table 2) was equal for all four miningsites. Although there were approximately nine times moreAfrican employees, the total indirect costs borne byEuropean mine workers for productivity lost was almostthree-fold higher. This is explained by approximately 26-fold higher wages paid to Europeans. Analogous to whatwas observed for direct treatment costs, implementationand maintenance of environmental management interven-tions promptly decreased the amount of work shifts lost,hence the indirect costs averted gradually increased.

A direct consequence was an increase in the overallproductivity of the mining companies. Use of DDT towardthe end of the malaria control programmes had furtherprofound effects on the number of working days gained asa result of lower malaria incidence. The overall indirectcosts attributable to clinical malaria attacks and workshifts lost at the mines were US$ 2 607 427 (in 1995 US$).In comparison, we estimated that the total indirect costsbecause of work shifts lost would have been US$8 286 173, if no malaria control measures had beenimplemented and sustained. Therefore, successful integra-ted malaria control averted over US$ 5.6 million (Table 7).

Consequences of integrated malaria control

The baseline annual malaria incidence rate at the RoanAntelope mine was 514 per 1000. During the rst 3 years

of programme implementation the annual incidence ratewas reduced to 135 per 1000. It remained at this lowerlevel until the mid-1940s. Expansion of the control areaand regular application of DDT from 1946 onwardresulted in another sharp decline in the annual incidencerate to an average of 21 per 1000 (Table 8). Although thereis no baseline malaria incidence rate available for theMufulira mine, the annual incidence rates measuredbetween 1936 and 1949 showed a similar trend, suggestingthat the initial incidence rate at Mufulira was as high as theRoan Antelope rate. The baseline incidence rate in Nkana-Kitwe was slightly lower than at the Roan Antelope mine.However, the initial package of malaria control interven-tions was less successful than at the Roan Antelope andMufulira mines. This is probably attributable to the greaterdifculty of controlling malaria in an urban environmentwith even more human population movement (Ferguson1999). The average annual incidence rates during theperiod 193643 were considerably higher. Implementationof additional drainage work in 1944 halved the incidencerate within 1 year. Regular application of DDT to theinside walls of the resident houses further decreased theincidence rates to 23 per 1000. Similar reductions inannual malaria incidence rates were also observed at theNchanga mine (Table 8).

The integrated malaria control programmes thereforeaverted more than half a million clinical malaria attacksduring the 20 years of intervention (Table 9). Furthermore,the programmes averted the loss of an estimated 942 347work shifts.

At the Roan Antelope mine, the baseline annualmortality rate because of malaria was 10.3 per 1000. The

Table 8 Effect of integrated malariacontrol on mean annual malaria incidencerates in four mining communities of theZambian copperbelt during the colonialperiod

Year/period

Malaria incidence (per 1000 per annum) Baseline* 193643 1944/1945 194649

Roan Antelope 514 135 123 21Mufulira 514 164 196 31Nkana-Kitwe 457 251 109 23Nchanga 514 156 156 30

* Baseline malaria incidence rate for the Roan Antelope mine were derived from the year1929/1930 before implementation of malaria control measures. In the absence of baselinedata for the Mufulira and Nchanga mines, it is assumed that similar rates must haveoccurred here. The rst annual malaria incidence rate for Nkana-Kitwe is for the year 1935.The baseline incidence rate therefore is probably underestimated.

Additional drainage work was carried out in Nkana-Kitwe, which displayed promptsuccess, as the malaria incidence rate was more than halved shortly thereafter. From November 1945 onwards, residual house spraying with DDT became an additionalcontrol strategy, resulting in a sharp decline of malaria incidence rates. Mining operations at Nchanga were launched in 1936 but there are no malaria incidencerates available until 1944.





14/21

rate fell sharply to 0.5 per 1000 during the period 193238(Watson 1953) and was further reduced to 0.37 per 1000between 1938 and 1943 (Rodger 1944). We assumed thatthe malaria-specic mortality rate did not exceed 0.5 per1000 until the late 1940s and that malaria death rates inthe other three mining sites were similarly low. Hospitalin-patient records for Africans at Mufulira support thisassumption, as there was only one patient who died of malaria in the year 1949 from a total of 3421 in-patients(case fatality rate because of malaria: 0.3 per 1000). Overthe entire 20 years that control programmes were in place,14 121 deaths were averted (Table 9).

We estimated that the malaria control programmesaverted a total of 127 226 DALYs, as expressed in 3-yearincremental periods (Table 9). Finally, these estimates wereused to calculate the total costs per death and malariaattack averted, which were US$ 332.42 and 9.07. Theaverage costs per DALY averted in 3-year incrementalperiods were US$ 36.90 (Table 9).

Sensitivity analyses

The results of the one-way sensitivity analyses are sum-marized in Table 10. Assuming direct treatment costs of asingle clinical malaria attack of US$ 1.00 rather than 2.22(expressed in US$ 1995) had only a small effect on the netcost-effectiveness ratios, as they changed by less than 10%.On the other hand, introducing higher direct treatmentcosts of US$ 5.00, the net cost-effectiveness ratios changedby slightly more than 20%. When we incorporatedminimal or maximal daily wages for Europeans employeesof US$ 55 or 65, respectively, rather than the mean dailywage of US$ 60, this affected the net cost-effectivenessratios by less than 8%. Similarly, allowing for annual wageranges of Africans between US$ 511 and 802 rather than700, was accompanied by changes in the net cost-effect-iveness ratios of less than 8%.

Assuming that the integrated malaria control pro-grammes were only run from 1930 to 1945, and halted

Table 9 Summary of total person life years at risk, costs (direct and indirect), consequences, as well as gross and net cost-effectiveness of integrated malaria control in four copper mining communities of Zambia between 1930 and 1949 (all costs expressed in 1995 US$)

Copper mining site

Roan Antelope Mufulira Nkana-Kitwe Nchanga Total

Person life years at risk 425 342 339 253 518 322 156 622 1 439 539

Costs of integrated malaria control 2 952 284 2 793 946 2 894 699 2 528 542 11 169 472

Direct treatment costsWithout malaria control 340 242 271 641 379 141 143 124 1 134 147With malaria control 77 077 76 482 143 259 40 707 337 525Direct costs averted 263 165 195 158 235 882 102 417 796 622

Indirect costs because of work shifts lostWithout malaria control 2 499 753 1 997 517 2 772 969 1 015 933 8 286 173With malaria control 591 539 584 292 1 127 120 304 476 2 607 427Indirect costs averted 1 908 214 1413 225 1 645 849 711 457 5 678 745

Total net costs 780 905 1 185 563 1 012 968 1 714 668 4 694 104

Consequences of malaria controlDeaths averted 4172 3328 5084 1536 14 121Malaria attacks averted 173 688 129 028 155 107 59 461 517 284DALYs averted (in 3-year increments) 37 791 30 125 45 882 13 428 127 226

Gross cost-effectiveness ratiosGross cost per death averted 707.59 839.57 569.33 1645.81 790.99Gross cost per malaria attack averted 17.00 21.65 18.66 42.52 21.59Gross cost per DALY averted (in 3-year increments) 78.12 92.74 63.09 188.31 87.79

Net cost-effectiveness ratiosNet cost per death averted 187.16 356.26 199.23 1116.07 332.42Net cost per malaria attack averted 4.50 9.19 6.53 28.84 9.07

New cost per DALY averted (in 3-year increments) 20.66 39.35 22.08 127.70 36.90





15/21

before the additional application of DDT, this hadconsiderable implications for costs, consequences and netcost-effectiveness ratios. The total costs of control inter-ventions for these rst 16 years of implementation wereUS$ 9 915 185. The direct treatment costs averted and theindirect costs averted because of work shift losses were US$471 148 and 3 666 134, respectively. Therefore, the netcosts were increased by 23%, namely to US$ 5 777 903.Under this scenario, there were an estimated 9132 deathsand an estimated 266 272 malaria attacks averted. Con-sequently, the net costs per death averted were US$ 632.70and the corresponding net costs per malaria attack avertedwere US$ 21.70 (Table 10).

Impact of integrated malaria control on the macroeconomy

Major mining activities at three of the four sites werelaunched in 1930 and 1931. This coincided with a globaleconomic recession. Large numbers of mining employeeslost their jobs in all parts of south-central Africa. In 1932,the estimated work force in the Northern Rhodesiancopperbelt was 6465. It increased by 27% to reach anestimated size of 8216 in the next year. Integrated controlmeasures had now been implemented for some 3 years anddramatically reduced the malaria-specic mortality andannual incidence rates. By this time, the annual copperproduction was 105 900 metric tons and the annualrevenues, expressed in 1995 US$, were US$ 182 million(Table 3). Toward the end of the recession, with copperprices sharply increasing, there was a need for largenumbers of workers. As effective malaria control measures

were sustained, there was signicant improvement inoverall living and working conditions. This was a keyincentive in the employment seeking behaviour of potentialworkers. As a result, in-migration continued on a verylarge scale. The total work force almost doubled from1933 to 1934 and by the year 1938 the size of the totalwork force was estimated at 22 654 (Table 2). Thesedramatic increases in the number of employees were instark contrast to neighbouring Belgian Congo and South-ern Rhodesia. Under the counterfactual assumption that noeffective malaria control would have been maintained,inhibiting further in-migration of cheap labour into thecopperbelt, we assume that productivity would haveremained unchanged after 1934. In 1938, for example, theannual revenues would have been US$ 216 million (in1995 US$), the increase simply being attributable to highercopper prices. In 1938, annual copper production hadreached a volume of 216 400 metric tons. Revenues, inturn, were as high as US$ 441 million (in 1995 US$). Thenational income statistics for the same year revealed that55% of the revenue was from mining. Thus, effectivemalaria control was a principle driving force behindNorthern Rhodesian economic development. Without it,the taxable national income would have been 28% lowerand the percentage of national income from miningactivities would have dropped to 37%.

Discussion

We have presented a comprehensive cost-effectivenessanalysis of copper extraction in four mining communities

Table 10 Summary of results from one-way sensitivity analyses (all costs expressed in US$ 1995)

Net cost-effectiveness ratios (% change)

Parametertested

Per deathaverted

Per malariaattack averted

Per DALY averted(in 3-year increments)

Base case 332.42 9.07 36.90

Direct treatment costsCosts to treat clinical malaria attack reduced from US$ 2.221.00 363.43 (9.3) 9.92 (9.4) 40.34 (9.3)Costs to treat clinical malaria attack augmented from US$ 2.225.00 261.78 (21.3) 7.15 (21.2) 29.05 (21.3)

Indirect costs because of work shifts lostDaily wage of European workers US$ 55 rather than 60 357.80 (7.6) 9.77 (7.7) 39.71 (7.6)Daily wage of European workers US$ 65 rather than 60 307.05 (7.6) 8.38 (7.6) 34.08 (7.6)Annual wage of African workers US$ 511 rather than 700 358.80 (7.9) 9.79 (7.9) 39.82 (7.9)Annual wage of African workers US$ 802 rather than 700 318.19 (4.3) 8.69 (4.2) 35.32 (4.3)

Discount rateAll costs discounted at 2% instead of 3% 321.68 (3.2) 8.78 (3.2) 35.70 (3.3)All costs discounted at 4% instead of 3% 341.00 (2.6) 9.31 (2.6) 37.85 (2.6)

Programme implementation prior to additional DDT application (193045) 632.70 (90.3) 21.70 (139.3) 71.83 (94.7)





16/21

of Zambia during the English colonial era by performingan empirical assessment of the impact of integrated malariacontrol. Complex as it was to connect malaria control tothe macroeconomy in the present case, we believe thatmaking this connection elsewhere is an order of magnitudemore difcult. Although the programmes were launchedmore than 70 years ago, and we had to make a series of assumptions, including plausible counterfactual scenarios,we have clearly demonstrated the tremendous impactintegrated malaria control programmes in copper miningcommunities of Northern Rhodesia had on the nationaleconomy. Therefore, our appraisal is a contribution to thechallenge which Paul Russell formulated more than40 years ago (Russell 1959, cited in Packard 1997).Drawing causal inference is a complex subject and in ourview it demands a exible, innovative and multifacetedapproach (Marini & Singer 1988). Here, a large body of experimental and observational evidence has been accu-mulated across multiple copper production sites over aperiod of 20 years. The sequence of events, starting withthe design and implementation of integrated malariacontrol, which fostered large-scale in-migration and suc-cessively led to exponentially growing copper extraction, issuggestive of strong temporal plausibility. Consistentobservations across the four settings that all outcome andimpact variables of the malaria control programmesexhibited the desired end points strengthens the plausibilityof our causal hypothesis. In addition, these observationswere in contrast to neighbouring copper producingcountries. Consequently, the association betweensuccessful malaria control and economic developmentcannot readily be attached to an alternative causalhypothesis.

It should be noted that disease control in the NorthernRhodesian copperbelt was not the overarching objective,but instead a necessity to facilitate economic development.From the companies perspectives, prot from copper wasthe primary driving force. In the rst instance there wasserious concern about the health of Europeans. Then therewas fear among potential mine workers about death frommalaria. Night shifts were a particular worry. Rumourspersisted among new employees who sought work andoften came from neighbouring Belgian Congo, SouthernRhodesia, and as far away as South Africa that they shouldonly buy one-way train tickets, as there was suspicion thatthey would never return (Watson 1953). However, bothrumours and fear disappeared shortly after integratedmalaria control was in place and proved successful(Watson 1953; Holleman & Biesheuvel 1973; Utzingeret al . 2001). It stimulated unprecedented in-migration(Watson 1953; Mitchell 1961; Parpart 1983), which, inturn, provided for labour substitution at a level that would

have ensured high revenues even with less successfulmalaria control programmes.

Our economic evaluation was carried out within themethodological framework proposed by Drummond andStoddart (1985), slightly modied by Mills (1993b) andadapted for the setting of copper mining communities inNorthern Rhodesia. Accordingly, costs and consequencesof malaria control were assessed to facilitate the cost-effectiveness analysis. As we had to rely on historicalrecords, it is important to note that record keeping duringthe British colonial period was excellent (Parpart 1983);hence, the present economic appraisal of a tropical diseasecontrol programme is comparatively thorough and unbi-ased. This kind of analysis, assessing the interrelationshipbetween disease control and economic performance anddevelopment by clarifying the underlying mechanisms,is virtually non-existent in the extant literature, becauseit has been viewed as almost impossible to conduct(Russell 1959, cited in Weisbrod et al . 1973; Packard1997).

On the cost side, the detailed programme budget for theRoan Antelope mine allowed appraisal of the total directcosts borne by the company, including the high initialcapital expenditures and annual maintenance costs of themalaria control interventions. Assuming equal capitalinvestments and estimating annual operational costs at theother three mining sites, also including a standard discountrate of 3%, facilitated estimation of the total costs thatwere spent on control measures over the entire 20 years of programme implementation. These costs allowed calcula-tion of gross cost-effectiveness ratios. Costs borne by thepatients and their families for curative therapies of clinicalmalaria attacks and indirect costs of work shifts lost werealso estimated and the costs averted as a result of integrated malaria control were included in the nalanalyses of net-cost effectiveness. Again, a series of assumptions had to be made. For direct treatment costs of a clinical malarial attack, we used a mean value of US$2.22 (in 1995 US$). It stems from four case studies indifferent countries of sub-Saharan Africa (Shepard et al .1991), approximately 50 years after the implementation of the control programmes presented here. Clearly, there areimportant differences in the range of treatment seekingoptions and behaviour, methods of diagnosis and malariadrugs used among mining communities in the rst half of the last century and the participants in those four casestudies. However, our analyses revealed that the overallcontribution of direct treatment costs was relatively smallcompared with other direct and indirect costs. This wasconrmed in the one-way sensitivity analysis, because morethan doubling the costs for a single malaria attack changedthe net-cost effectiveness ratios by only 21%.





17/21

The primary features of the control measures discussedhere were their consequences for malaria and work forceproductivity. In this regard, it is also important to note thatour estimated costs averted and the consequences becauseof the integrated malaria control programmes were prob-ably at the upper limit of the possible. With less successfulcontrol measures, far fewer employees would have soughtwork on these mines; hence, less person life years were atrisk of malaria. We found highly signicant reductions inmortality, incidence rates, DALYs and days of work lost.Furthermore, the programmes were more cost-effectivethan other, currently widely applied, malaria control tools(Goodman et al . 1999, 2001). For example, we estimatedgross and net costs per death averted of US$ 790.99 and332.42, respectively. These values were at the lower endwhen compared with other studies, normally using onlyone single intervention, with cost-effectiveness analysesperformed over short durations.

A central feature of the programme design employed inthe Northern Rhodesian copperbelt was that severalinterventions were tuned to the local ecology and imple-mented simultaneously. Our analyses were based onhistorical controls, comparing the pre-programme stagewith that during programme implementation. At present,the most widely accepted method for assessing the effect-iveness of interventions in the health care eld is evidencederived from randomized controlled trials (Drummond &Stoddart 1985; Goodman et al . 1999). However, we haveargued recently that historical controls in the presentsettings are more appropriate particularly because of greatecological variations among the four mining communities(Utzinger et al . 2001). This inhibits comparison betweenthe different settings and makes the notion of a controlcommunity meaningless. Matching communities on eco-system structure is not practicable. Our present analysesclearly conrmed that there was great variability in thelocal ecology and epidemiology across the four miningcommunities, as can be seen, for example, by the differentmean annual malaria incidence rates.

Our study shows that the overall costs of implementingand sustaining integrated malaria control programmeswere minuscule when compared with the total revenuesfrom the mining sector in the Northern Rhodesiancopperbelt. Over the entire 20 years of programme im-plementation, a little more than US$ 11 million were spentfor malaria control measures. As the programmes werehighly successful, almost US$ 6.5 million in direct andindirect costs could be averted. Integrated malaria controltherefore contributed a share of approximately 0.07% of the total revenues of US$ 7.1 billion by the copper miningcompanies (all gures in 1995 US$). We argue that thecompanies initial capital investment mainly in environ-

mental management strategies, the maintenance and tuningof these interventions to the local settings and the integ-ration with additional malaria control measures were thekey determinants of social and economic development inNorthern Rhodesia. The underlying mechanism was thatsuccessful malaria control promoted unprecedented andessential in-migration and sustained healthy work forces. Itis important to note that wage labour was nothing newwhen copper exploitation commenced on a large scale inthe Northern Rhodesian copperbelt. Instead, it had per-sisted for at least two decades. Thousands of migrantworkers from this country had previously gained wages formine production work in neighbouring copper producingnations (Parpart 1983; Ferguson 1999). Initially, largenumbers of unskilled workers were required for mineconstruction. Early estimates suggest that by 1930 therewere nearly 30 000 workers; however, they only stayed forvery short periods of time. The work force rapidly declinedin parallel with the global economic recession at that time.From 1932 onward, there was a sharp increase in wagelabour in the Northern Rhodesian copperbelt. This wasdistinctly different from neighbouring Belgian Congo. Twoadditional features were characteristic of this period. First,there was a sharp increase in the length of employment.Secondly, the proportion of married couples grew rapidly,suggesting that mine workers were encouraged to bringtheir wives and children to the mining sites labourmigration was gradually replaced by permanent urbaniza-tion (Parpart 1983; Ferguson 1999). But above all, thedistinctively higher mining productivity compared withBelgian Congo, could not have been accomplished withoutsuccessful malaria control (Watson 1953; Holleman &Biesheuvel 1973).

It is important to emphasize that successful control wasachieved by a host of interventions. The key idea is that thisstrategy impacts themalariatransmission systemat multiplevulnerable points. Using several interventions exerted lesspressure on any one control measure (e.g. antimalarialdrugs) to do the complete control task. Thus, the demandfor drugs, for example, was reduced substantially beyondwhat it would have been if this had been the only controlmeasure. Applications of oil to open water bodies, andduring the last 4 years also residual spraying with DDT,were part of the interventions. Today, these tools areconsidered to be environmentally unsound. We did notadjust our costs of deaths, malarial attacks, DALYs anddays of work lost averted for the potential environmentaldamage caused by these control measures. However, we didperform some sensitivity analyses under the scenario thatcontrol programmes were only operated until 1945, prior tothe additional application of DDT. Although net cost-effectiveness ratiosdecreased considerably, the programmes





18/21


19/21

copper prots. To ensure steady revenue ows, themanagement boards of Northern Rhodesian miningcompanies carefully monitored the number of shifts lostbecause of malaria and implemented measures to keep asfew workers absent from the mines as possible. Clearly,there was a maximum tolerable level of absenteeism,beyond which copper extraction would not have beenprotable. Furthermore, a sound control strategy reducesindirect costs borne by mining employees because of reduced work shifts lost. High malaria incidence rates atthe Bwana Mkubwa mine prior to 1930 was the principalreason that copper production never really took off in thatperiod (Mitchell 1961). However, in the present case study,integrated malaria control was so effective reducing theincidence rates in the rst 3 years by 5075% that therewas never any shortage of labour to carry out copperextraction. Besides keeping malaria incidence rates as lowas possible, there were important seasonal variations intransmission and, hence, incidence. Monitoring monthlyrates showed that there was a distinct peak toward the endof the rainy season in March (Watson 1953; Utzinger et al .2001). The Northern Rhodesian mining companies weremost vulnerable during this time of the year. Going beyondthe context of mining companies, seasonal variation inmalaria transmission also applies to agricultural settings(Audibert 1986). Unfortunately, this feature of the malariaproblem has received insufcient attention in health policydiscussions, as well as in government planning andadministration (Chambers et al . 1979).

Malaria control was a critical ingredient for protablecopper extraction. Thereby, it became a leading factor foreconomic development of Northern Rhodesia. In the1950s, 1960s and early 1970s, copper production furtherincreased to reach a peak of 778 900 metric tons in the year1972 a share of 9.9% of the total annual global copperproduction (Commodity Research Bureau 1984). Thereaf-ter, copper production decreased gradually, and in 1998,Zambia produced a total of 315 000 metric tons of copper.Its global share had shrunk to 2.6% (Bridge CommodityResearch Bureau 2001). This amount of copper productionwas virtually the same as half a century ago. Recently, thecopper mines of post-colonial Zambia have been privatizedand there is great interest in, once again, increasing annualcopper production. Alongside these developments, there isrenewed interest in the question of how to control malariaat these copper mining sites. Importantly, copper extrac-tion in contemporary Zambia continues to be a majorpotential source of economic advancement for the country.The most recent economic indicators available for the year1998 revealed that mining activities still made a contribu-tion of 10.7% to the taxable income (Central StatisticalOfce 1999). Copper companies are currently aiming at a

signicant increase in copper production. Consequently, itis expected that the economic share of this natural resourceto the Zambian GDP might increase considerably over thenext several years. Integrated malaria control will be of pivotal importance to achieve the desired social andeconomic benets.

Besides the present example, there are a host of analogous situations in the industrial and service sectors of sub-Saharan Africa suggesting that effective malaria con-trol programmes will be a crucial element for much neededeconomic development. Implementation and maintenanceof integrated malaria control measures in agriculturalplantations and mining and oil drilling operations are of particular interest because of the high revenues on capitalinvestments. Integrated malaria control is likely to succeedin urban centres and in port cities of sub-Saharan Africa,where large populations are aggregated on relatively smallsurfaces. The essential point is that high populationdensities are characterized by large-scale drainage worksand freshwater pollution of remaining surface waters. Bothfeatures are negatively correlated with densities of anoph-eline mosquitoes. Effective vector control in these settingsmight have substantial positive impact on tourism (Sachs& Malaney 2002).

Finally, integrated malaria control approaches are likelyto have methodological and infrastructure spillovers toght other parasitic diseases, as well as HIV/AIDS andtuberculosis, so that the overall disease burden of poorcountries can be substantially reduced. Implementation of these programmes requires effective public health infra-structures. This is where investment is critically needed.Education of local public health engineers and agriculturalpersonnel who can carry out intersectoral and multicom-ponent intervention programmes is a pressing necessity.Substantial increases in global donor support leadingeconomists have forwarded gures of US$ 1020 billionper annum will be mandatory. It is encouraging that thisis nally entering the contemporary discussion of thedevelopment communities.

Acknowledgements

We are grateful to Jacqueline V. Druery, Lara J. Moore,Susan B. White and Bobray J. Bordelon from PrincetonUniversity for invaluable help screening the historicalliterature. Thanks are addressed to Patrick Gerland formapping support and two anonymous referees for aseries of excellent comments and suggestions. This workwas nancially supported by research grants from theCentre for Health and Wellbeing at Princeton University(J.U. and Y.T.) and the Swiss National Science Founda-tion (J.U.).





20/21

References

Aikins MK, Fox-Rushby J, DAlessandro U et al . (1998) TheGambian national impregnated bednet programme: costs, con-

sequences and net cost-effectiveness. Social Science and Medi-cine 46, 181191.

Anonymous (1931) Empire Copper the New Mines of Rhodesia.I. A Task for the Geologist. The Times, London.

Arrighi G (1973) Labor supplies in historical perspective: a studyof the proletarianization of the African peasantry in Rhodesia.In: Essays on the Political Economy of Africa (eds G Arrighi & JS Saul) Monthly Review Press, New York, pp. 180234.

Asenso-Okyere WK & Dyator JA (1997) Household cost of seeking malaria care: a retrospective study of two districts inGhana. Social Science and Medicine 45, 659667.

Attanayake N, Fox-Rushby J & Mills A (2000) Household costs of malaria morbidity: a study in Matale district Sri Lanka.Tropical Medicine and International Health 5, 595606.

Audibert M (1986) Agricultural non-wage production and healthstatus: a case study in a tropical milieu. Journal of Develop

Econ.2002.Utzinger.tmih.Economic Payoffs of Integrated Malaria Control in Zambian Copper Belt Between 1930 and 1950

Documents