Background
The South African mining industry has been built on migrant labour from the rural areas of South Africa and neighbouring countries, a system which although much reduced in size persists today. The numbers as well as geographical origin of these miners have changed over time in response to the changing needs of industry and to political and economic shifts in the region [
1‐
4].
Prolonged or high exposure of miners to respirable silica dust released in the process of mining gold (and to a lesser extent coal) is the cause of silicosis, which now typically takes the form of a slowly developing lung fibrosis. Pulmonary tuberculosis is a frequent and potentially disabling complication of silica dust exposure and silicosis in this migrant population, as well as being the product of high rates of infection in congregate settings on the mines. Population tuberculosis risk is further elevated by a prevalence of HIV infection of the order of 25% [
5,
6] and a synergistic effect of HIV infection and silicosis on such risk [
7]. Consistent with these causal associations, studies of gold miners over the past two decades have confirmed high prevalences of silicosis of 13% to 25% in long service miners [
8,
9] and high annual tuberculosis disease rates of 2500 per 100,000 and above [
10]. Owing to the circular migration patterns of this workforce, high rates of tuberculosis among miners have translated into high tuberculosis rates in their home communities [
11‐
13]. A similar association with migrant status has been found for HIV infection [
14].
While employed miners on the larger mines have access to regular screening for tuberculosis and silicosis and to medical care, upon leaving the industry most of these ex-miners are lost to follow-up. However, their increased risk of tuberculosis due to silicosis and cumulative silica lung load is life-long [
15], while silicosis itself may appear or progress after leaving employment [
16]. Given the poorly developed health systems in the labour sending areas of Southern Africa, high rates of undiagnosed or incompletely treated tuberculosis are predictable, accompanied by poor access to the statutory compensation system for mining lung disease [
6,
17‐
19].
This triple epidemic of silicosis, tuberculosis and HIV infection among miners drawn from the rural areas of South Africa and neighbouring countries, particularly Lesotho, Mozambique, Swaziland and Botswana, has belatedly become a target for regionally directed international action. In 2012 the Ministers of Health of the Southern African Development Community (SADC) committed their countries to reducing the burden of mining related lung disease [
20]. Following a World Bank economic analysis of the benefits of treating and controlling tuberculosis among miners [
21], the Global Fund to Fight HIV, Tuberculosis and Malaria committed ZAR 500 million (US$ 36 million) to the screening and treatment for tuberculosis of 500,000 miners, ex-miners and residents of peri-mining communities [
22].
These initiatives have exposed the dearth of information on former miners – their whereabouts, occupational and other risk factors, health status, and changes in these phenomena over time. For example, while country or province of origin data on migrant miners have been reported [
3,
13,
21,
23], these data have not been disaggregated to district level. The first objective of this study was thus to make use of a large recruitment database going back four decades to provide a current and historical description of the population of mainly migrant miners and ex-miners from the South African mining industry. The focus was on the large commodity sectors of platinum group metals (hereafter “platinum”), coal, and particularly gold mining as the highest risk industry for lung disease. A second objective was to draw out implications for the health status of this population with specific reference to silicosis and tuberculosis.
Methods
The TEBA database
An electronic database of contracts of miners recruited into the industry between 1973 and 28 October, 2013, was obtained from TEBA, an organisation that in various forms has been the major recruiting agency for the mines for since 1901 [
2]. In 1977 its precursor organisation, the Witwatersrand Native Labour Association (WNLA), merged with the Native Recruiting Corporation (NRC) to form The Employment Bureau of Africa (TEBA). Owned for most of this period by the South Africa mining industry association, the Chamber of Mines, it became an independent private company, TEBA Ltd. (“TEBA” in this report) in 2005.
Personal communication from members of TEBA management constituted the source of information about the database in this and other sections. TEBA has a recruitment contract archive of 3 million person records. These records were held in hard copy until the 1970s. Digitisation of the archive commenced in the 1970s and electronic capture of the contracts of first time entrants to the industry was complete only for those entering from 1983 onwards, the year in which the mainframe computer was set up.
TEBA predominantly recruited for the gold mining industry, with a much smaller role in platinum and coal. From the late 1980s, there was a switch away from TEBA recruiting channels as South African nationality workers began to apply directly to the mines or through their union for employment. These contracts were not entered into the TEBA archive. However, TEBA continued to record all foreign nationality workers, who had to meet immigration requirements.
In response, TEBA shifted the emphasis of its operations by the 2000s from recruitment to processing, by opening offices on the mines and taking over the first step in the process of engagement of all employees. This greatly increased the numbers entered into the TEBA record system, including contractors (defined below) and non-migrant supervisory level (mainly white) miners. The companies Amplats (platinum) and Amcoal also transferred their labour registration process to TEBA over time.
Analysis of the database
The database was converted by the researchers from one based on contracts [
N = 10,327,396] to one based on individual miners [
N = 1,644,264]. The process of organising, reducing and cleaning the original datafiles is described Additional file
1: Supplementary note 1. The variables available from the database are set out in Table
1.
Table 1
Variables from the TEBA database used in this study
Unique key | TEBA randomly assigned no. (Not personal ID number). | |
Date of birth | | Age |
Gender | | |
Racial ascription | Four category classification formalised under apartheid, and in continuing usage for some official purposes – Asian, black, coloured, white. | |
Origin | Country and district/region/province of office at which recruited or registered for a given contract. | Origin at first contract on the database |
Contract start date | | Number of contracts. Start date of first contract. Start dates of returning contract. |
Contract end date | | End date of first contract. End dates of returning contract. |
Contract duration (days) | Difference between contract end date and contract start date. | Cumulative service across all contracts. |
Contract duration (months) | Days or weeks rounded up or down to the nearest whole number of months, e.g. 35 days = 1 month. | Cumulative service across all contracts. |
Commodity sector | Gold, platinum group metals, coal, “other” (antimony, asbestos, chrome, copper, diamond, granite, iron, lead, lime, tin, vanadium)a
.
| Commodity sector of longest cumulative service for an individual. |
Occupational risk category | Underground, “surface risk”, “surface non-risk”. Surface definition based on exposure to hazardous airborne agents. | Occupational risk category of longest cumulative service for an individual |
Employment status | Mine employee versus “contractor”. See text for further definition. | Exclusively mine employee versus contractor “ever” on TEBA database |
With regard to employment status, although all miners have contracts, “contractor” is the term used on the TEBA database for miners who are employed by a subcontracting company or labour broker (“outsourcing” or “subcontracting”) rather than by the mining company itself, distinguishing them from “mine employees”. The database also has a “risk work” classification based on occupational respiratory hazards, mainly silica. All underground work and certain jobs on the surface (e.g. electricians, fitters, winding engine drivers) are categorised as risk work.
A total of 21% of miners had worked in more than one commodity sector and 36% in both surface and underground jobs. Where a unique assignment was needed for summary purposes, miners were assigned to the category in which they had accumulated the longest service. We elected not to use this method for employment status because of uncertainty about how many of the contractors registered on the TEBA database in the 2000s (when contracting became common) had unrecorded previous mine employee service. Employment status was thus dichotomised as exclusive mine employee only vs. “contractor ever”.
In order to assess the representativeness of the TEBA database with regard to industry employment, annual employment data were sought from sources other than TEBA. These sources and the comparisons made are described in Additional file
1: Supplementary note 2 and Additional file
2: Figure S1.
Frequencies, medians and ranges were calculated for variables of interest (listed in Table
1). Five-year intervals were used for comparison of changes over time. For the analysis of annualised cumulative service and secular trends, the database was truncated at 31 December, 2012.
For purposes of mapping country, province and district of origin, only the first contract of each miners was used. Because it was not possible to determine from the database whether contracts recorded after 1973 were new or returning, the numbers mapped as new entrants are most reliable from 1983 onwards (on the assumption of a low probability of a miner being on a returning contract with no evidence of contracts in the previous decade).
The GIS software package ArcGIS 10.1 by ESRI was used. The mapping was disaggregated at the district level, and included miners from South Africa and neighbouring countries. During the geocoding step, district names associated with each miner (as recorded at the point at which they were recruited – not necessarily resident - for the first time) were assigned a latitude-longitude coordinate relating to their geographic location on the earth’s surface.
These coordinates were then integrated into a map interface, creating a spatial data layer of all miners recruited for each year since 1973. This provided a visual indication of the distribution of recruitment over time, broken down into 5-year periods, and aggregated at the district (or regional or country level for countries for which district subdivisions were not available).
Discussion
This report is to our knowledge the first published attempt to map the recruitment origin of this workforce, and particularly the black migrant workforce, to the district level, as well providing a detailed description of occupational and demographic characteristics. The main trends are a decline in gold mining employment and rise in platinum employment, a decline in foreign recruitment, a shift in point of recruitment northwards and westwards and an increase in the number of female miners. There was also a substantial increase in the employment of contractors. These findings and their health implications are discussed below.
Commodity sector
During the first three decades covered in this analysis, gold mining employment dominated the TEBA database. Striking feature features are the increase in mining employment from the mid-1970s through the 1980s (Additional file
4: Figure S2), a time of a rising gold price, and the decline in new recruits after 1988 at a time of retrenchment following a major miners’ strike in 1987. The finding that this decline was not matched by a corresponding fall in total employment may be partly due to re-engagement of many of the retrenched workers [
24] with preferential re-employment of previously employed miners. However, a long-term fall in total mining employment becomes apparent from the late 1990s, driven by a decline in gold mining employment (Table
3, Additional file
3: Table S2 and Additional file
2: Figure S1).
This decline has important implications for public health outside the formal employment sector. Gold mining has been the main generator of occupational lung disease in the form of silicosis and related tuberculosis [
13]. As the ratio of ex-gold miners, including many with long service, to working gold miners rises, so does the “externalisation” of disease from the industry to community settings [
9]. Continued shedding of jobs is forecast for the gold mining industry in response to economic and political difficulties.
Employment numbers in the platinum mining industry passed those of gold in 2006 during platinum’s phase of rapid expansion [
25]. While platinum mining in South Africa carries a very low risk of silicosis, about a half of its miners have worked in other commodity sectors [
26], thus carrying their risk factors, including lung silica load and silicosis from gold mining, into the platinum sector. However, the platinum industry is also currently in economic crisis owing to rising costs and a fall in demand for the metal, which is likely to further reduce the employment opportunities for former gold miners.
Duration of exposure to silica dust
The available exposure variables most relevant to silicosis and related tuberculosis are time spent in underground or surface risk work, particularly on gold mines. The large majority of miners work underground or have mixed underground/surface exposure. While the median gold miner entered employment at 25 years of age for a service duration of 6.5 years, over the period covered by this database more than 220,000 gold miners accumulated service of 15 years or more. While a figure of 15–20 years is frequently used as a “threshold” for the appearance of radiological silicosis under conditions of dust exposure over the past half century, compensation data from the 1970s to 1990s suggest that as many as a half of black gold miners with silicosis claims had accumulated less the 15 years of service [
27].
In addition, many miners have silicosis that is not detectable on chest x-ray [
28]. Autopsy data of gold miners dying of non-natural causes (presumed to be unrelated to lung disease risk) show that 4% of black miners with less than 10 years of service and 13% with 10 to 14 years of service have autopsy evidence of silicosis in their lungs [
29]. These two low duration categories make up the majority of gold miners coming to autopsy from non-natural causes. Thus even relatively short service ex-gold miners may remain at risk of later progression of silicosis to the radiological stage, although information on the extent of this “long latency appearance” of silicosis is lacking.
With regard to tuberculosis risk, there is evidence of a low silicosis threshold for the increase in tuberculosis risk, with an elevated relative risk observable in the lowest category of both autopsy silicosis (“negligible”) [15) and radiological silicosis [7] (“possible”, i.e. International Labour Organisation category 0/1) [
7,
30].
These figures have implications for any public health programme designed to detect tuberculosis and silicosis in ex-miners. The first is that while the number of gold miners on the TEBA database still alive who have worked over the past four decades and who have accumulated 15 or more of service is unknown, on an estimate of 60% still living (out of a denominator of 221,054 long service miners), this group is of the order of 133,000 individuals. This is a minimum, as it excludes the living fraction of miners who bypassed the database. The second implication is that even among those with shorter service histories there is an elevated lifelong risk of tuberculosis due to accumulated silica load and sub-radiological silicosis. Taken together with a high prevalence of HIV infection and background tuberculosis risk in their home communities, this group is likely to have a high lifelong annual risk of active tuberculosis.
Female miners
The rise in employment of women in the mining industry, particularly in risk work, is a recent phenomenon, linked to the Mining Charter of 2002, a negotiated code gazetted as a regulation in 2004, which set a target of 10% for participation by women in the industry [
31]. As women accumulate exposure in dusty areas, they are likely to be subject to similar rates of disease as men in those areas. There has not to our knowledge been any study of mining related disease among women, neither silicosis nor tuberculosis, and such research will be needed.
Place of recruitment
There is now a large literature on the oscillating migration of large numbers of miners recruited from surrounding countries and the rural Eastern Cape of South Africa, and on the profound socioeconomic impact of the migrant system on life in these regions [
4,
23,
32‐
34]. The trend over the past four decades has been for labour from surrounding countries to be replaced by South African nationals, with the proportion of foreign nationals declining from over half of all miners employed in the 1973–1978 period to a quarter in 2008–2012. The decline in the previously large number of Malawian recruits dates from 1974, with some reinstatement thereafter until complete cessation in the late 1980s. Reasons for these changes are varied, with pressures originating both in labour sending countries such as Malawi and Mozambique, and in South Africa, with changes in immigration laws and political and economic pressures on the industry to recruit domestically [
4,
23,
32,
34].
Within South Africa, the major shift of point of recruitment has been from the Eastern Cape to the Northwest province, particularly Rustenberg, and to other areas of mining or urban recruitment activity. This shift reflects the replacement of the gold by the platinum industry as the major employer of labour, the signing on by migrant miners directly at the mine rather than in their home district, and a policy of recruiting increasingly from surrounding rather than remote areas within South Africa.
Studies measuring the burden of silicosis and tuberculosis among former miners are sparse, limited almost entirely to former gold miners from specific districts in the Transkei [
35] and Botswana [
36] and to a cohort off laid off miners from Lesotho [
6,
8]. There are none to our knowledge from Swaziland nor Mozambique. The change in the geographic and sectoral concentrations of miners described above calls for community studies in these areas, where miners with both platinum and gold mining history now live. The information from such studies would inform education, screening, treatment and compensation services in these areas.
Contractor vs. mine employee status
The proportion of miners who had worked at least one contract as a contractor accelerated sharply as a proportion of those employed after 1998–2002, reaching over half of the total (and gold mining) workforce by 2008–2012. Contractors are a heterogeneous category, including miners hired via labour brokers, skilled workers employed primarily in surface work and those employed by established specialist companies contracted to undertake some stage of the mining process, such as shaft sinking [
23]. Because of this heterogeneity, it is difficult to state general hypotheses about the implications of contractor status for health and well-being. These implications are likely to be adverse where contractor status entails precarious employment in the sense of lower earnings, fewer benefits, and less access to routine medical surveillance and medical services than company employees [
23]. Whether contractor status translates into a higher risk of occupational lung disease is not known. This category needs to be disaggregated by precariousness of employment, occupational level and length of service as a contractor to answer this question.
Limitations
The TEBA database is not a complete representation of mining employment. For the period 1973–1983, it is clear that the database numbers are an undercount (Additional file
2: Figure S1). However, there is no reason to believe that demographic and occupational characteristics were greatly distorted by the undercount, an assumption supported by the consistency of these proportions with those of the immediately succeeding period.
After 1983 TEBA figures exceed or are very close to those from the two other sources used (Additional file
2: Figure S1). It is likely that the database provides a complete picture of black, and therefore migrant, gold mining employment during the 1980s. Because workers from other countries, whether contractors or mine employees, had to go through TEBA registration channels, the database is also likely to be a near complete representation of foreign migrant labour. Despite the observation by a TEBA official that many recruits from its traditional South African recruitment population bypassed TEBA in the 1990s, Additional file
2: Figure S1 does not show an undercount compared to government figures, which suggests that the latter may also be an undercount. The bypass factor would have diminished after 2003 with the shift in the TEBA modus operandi to registration from recruitment. The increase in the proportions of white workers and surface jobs on the database is likely to be an artefact of this shift.
With respect to platinum and coal, the TEBA database representativeness was incomplete from the beginning of the period of interest because of competing recruitment channels. Registration of platinum miners increased substantially after 2003, with a persistent but narrowing shortfall compared to government employment figures (Additional file
1 Supplementary note 2). TEBA has always been a minority recruiter of coal workers, who in South Africa have a lower risk of pneumoconiosis than gold miners [
37], although studies of the coal mining workforce remain scarce.
Finally, since miners with multiple contracts were assigned to area of registration at first contract, we could not identify miners who who signed up for their first contract or for later contracts at localities different from their home area. This size of this discrepancy is unknown but is likely to have increased over time for the reasons discussed earlier.
Conclusions
The objective of this report was to provide a descriptive rather than an explanatory analysis of changing recruitment patterns in the South African mining industry, with a focus on their implications for the overlapping domains of occupational and public health.
All miners and ex-miners from the South African industry are by statute entitled to medical examinations and monetary compensation for occupational disease. While miners who are actively employed usually have access to medical services for these diseases, including treatment, once they leave the industry the responsibility for their health surveillance passes from the industry mainly to the state. This system of post-employment health surveillance works very poorly [
18,
38,
39], and as a result many such ex-miners are lost to the health and compensation systems. For example, a survey of rural ex-miners found that 62% of those eligible had not been compensated, 35% partially compensated (i.e. not for disease progression) and only 2.5% fully compensated [
35]. This barrier to access is not only an individual and public health concern but reflects the failure of a century of legislation resulting from miner struggles to secure compensation for these diseases. Amelioration is thus a matter of social justice.
The detection and treatment of tuberculosis among miners, former miners and their communities are currently the target of a huge international effort. A 2011 modelling study suggested that 760,000 new cases of tuberculosis in sub-Saharan Africa every year may be linked to mining activity across the subcontinent [
12]. Control of the epidemic of tuberculosis, HIV and silicosis in miners and ex-miners in Southern Africa is therefore both an occupational health and a public health priority. This cannot be done without knowing where the miners or former miners and their families are.
Based on the region’s primary recruitment database, the study establishes an epidemiological and spatial framework that can be used for disease projection, costing and planning of programmes, including access to medical examinations and compensation, for a hard to reach and long neglected population. The findings should also assist actuarial costing of the compensation liability as part of current attempts to reform the compensation system.
Acknowledgements
We acknowledge TEBA Ltd. for making available an anonymised version of the database. This is an administrative database and was not set up for research purposes – we are therefore grateful to TEBA officials and former officials who provided valuable comment, corrections or estimates based on recall; specifically Dr. Graham Herbert, Kevin Cotterell, Nardo Neveling and Anneri Pieterse. Outside of TEBA, Peter Lewis provided valuable insights into mine employment practices and Dr. Martin Nicol supplied a summary of hard to find annual employment data from external sources. Rebekah Chang of the Clinton Health Access Initiative contributed to early analyses and summaries of the data. Thomas Slingsby and Nicholas Lindenberg of UCT GIS Unit did the work of matching districts used by TEBA to available software and produced the maps.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.