Rasmus Køster-Rasmussen 1,2, Maria L. Westergaard 2,3, Marie Brasholt 2, Richard Gutierrez 4, Erik Jørs 2,5,6, and Jane F.

Transcription

1 Scientific Solutions Mercury Pollution from Small-Scale Gold Mining Can Be Stopped by Implementing the Gravity-Borax Method A Two-Year Follow-Up Study from Two Mining Communities in the Philippines NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 2016, Vol. 25(4) ! The Author(s) 2015 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: / new.sagepub.com Rasmus Køster-Rasmussen 1,2, Maria L. Westergaard 2,3, Marie Brasholt 2, Richard Gutierrez 4, Erik Jørs 2,5,6, and Jane F. Thomsen 2,6,7 Abstract Mercury is used globally to extract gold in artisanal and small-scale gold mining. The mercury-free gravity-borax method for gold extraction was introduced in two mining communities using mercury in the provinces Kalinga and Camarines Norte. This article describes project activities and quantitative changes in mercury consumption and analyzes the implementation with diffusion of innovations theory. 1 The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, Denmark 2 The Danish NGO Diálogos, Odense, Denmark 3 Department of Neurology, Glostrup Hospital, Denmark 4 The Philippine NGO Ban Toxics!, Quezon City, Philippines 5 Department of Occupational and Environmental Medicine, Odense University Hospital, Denmark 6 ICOEPH, Danish Society of Occupational Medicine, Odense, Denmark 7 Department of Occupational and Environmental Medicine, Bispebjerg University Hospital, Copenhagen, Denmark Corresponding Author: Rasmus Køster-Rasmussen, The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, Copenhagen DK-1014, Denmark. glr478@sund.ku.dk

2 568 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) Activities included miner-to-miner training; seminars for health-care workers, school teachers, and children; and involvement of community leaders. Baseline (2011) and follow-up (2013) data were gathered on mining practices and knowledge about mercury toxicology. Most miners in Kalinga converted to the gravity-borax method, whereas only a few did so in Camarines Norte. Differences in the nature of the social systems impacted the success of the implementation, and involvement of the tribal organization facilitated the shift in Kalinga. In conclusion, the gravity-borax method is a doable alternative to mercury use in artisanal and small-scale gold mining, but support from the civil society is needed. Keywords mercury, gold, borax, small-scale gold mining, mining, environmental pollution, diffusion of innovation, technology transfer Introduction Artisanal and small-scale gold miners (ASGMs) worldwide use mercury to extract gold. 1 Whole-ore amalgamation is a technique that uses g of mercury to produce 1 g of gold. 2 Within the last eight years, it has become evident that this technique is more widely used than earlier anticipated, and artisanal small-scale gold mining (ASGM) is now considered the single largest contributor to global mercury pollution. The United Nations Environment Programme has estimated that ASGM contributes a 37% share of the anthropogenic emission of mercury to the atmosphere. 3 Whole-ore amalgamation is commonly practiced in the Philippines. 2 The ore is mixed with water and crushed in a rod mill (a rotating drum with metal rods inside), after which mercury is added and milled with the ore. The gold dissolves in the mercury and forms amalgam. After the milling, the content is poured into a tub where the mercury (holding the gold) sinks to the bottom. Excess mercury is recovered, and by blowtorching the remaining bit of amalgam, the mercury evaporates leaving behind the gold. In Benguet province, a group of miners have been using the mercury-free gravity-borax method (GBM) for gold extraction for decades. This method basically requires the same equipment as the amalgamation methods. However, after the rod milling, an ore concentrate holding the heavy minerals is produced by using a launder and a gold washing pan. The heavy mineral concentrate is mixed with borax powder. By blowtorching the mix, the borax melts and the gold sinks to the bottom. It has recently been demonstrated that under identical conditions, GBM yields more gold than the traditional amalgamation method. 4 Despite this advantage, GBM is not widely used outside Benguet. No earlier studies have documented that GBM can be implemented in an area using whole-ore amalgamation. Earlier technology transfer projects have promoted retorting of amalgam as a preventive technology in ASGM. 5,6

3 Køster-Rasmussen et al. 569 Unfortunately, this method is ineffective in areas using whole-ore amalgamation, since up to 90% of the mercury is lost to the tailings (the waste product consisting of fine sand) and only 10% is lost by blowtorching of amalgam. 2,7 Thus, only a small fraction of the mercury spill from whole-ore amalgamation can be recovered with a retort. Introduction of cyanidation techniques has also been suggested as a way to reduce or eliminate the use of mercury in ASGM. 8 Although most cyanide compounds decompose into nontoxic forms, cyanide itself is highly toxic and exchanging one toxic chemical with another may be regarded as a controversial solution. We present the results of a project in which GBM was introduced in two mining communities in the Philippines where whole-ore amalgamation was widely practiced. The main objective of the project was to reduce mercury pollution from small-scale gold mining by encouraging ASGMs to use a mercuryfree method. This article aims to (1) describe the changes in mining practice in two ASGM communities after two years of project implementation, (2) describe the activities undertaken to encourage the communities to adopt mercury-free gold mining, and (3) explore enabling and hindering factors in phasing out mercury use. Materials and Methods This two-year follow-up study was implemented in in two provinces in the northern Philippines: Kalinga and Camarines Norte, both located on the main island Luzon. In Kalinga, the project was focused in a remote mining site (Gaang) in the mountains, a four-hour hiking distance away from the road and the nearest village, whereas in Camarines Norte, the project was implemented in a community consisting of three neighboring villages (Labo, Jose Panganiban, and Paracale) accessible by road and with mining activities taking place directly in the villages. The two mining communities in Kalinga and Camarines Norte are described in Table 1. The two project sites were selected because they were known to be centers of small-scale gold mining using whole-ore amalgamation. During initial meetings, miner representatives and community leaders expressed interest in the project. Other larger mining communities, particularly in the southern Philippines, were also considered. However, the selected villages had not previously been involved in mercury pollution projects and were more accessible from the capital Metro Manila and Benguet where the miner-trainers were based. The local peace and order situation was also taken into consideration in the selection of project sites. All ASGMs in the project sites were invited to the training sessions in GBM. The project s goal was to get 90% of ASGMs in the selected areas to shift to GBM and to improve the miners level of knowledge about mercury toxicology. The project also aimed for 80% of school children in the villages to be aware of the dangers of mercury to health and environment and for 80% of the

4 570 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) Table 1. Description of Two Project Sites in the Philippines Selected for Implementation of the Mercury-Free Gravity-Borax Method for Gold Extraction in Small-Scale Mining. Project site Kalinga Gaang Camarines Norte Labo, Paracale, and Jose Panganiban Total population in project sites (approximately) ,000 Estimated number of active ASGMs in project site a Estimated proportion of ASGMs who are women 40% 5% Milling stations in project sites, n Residential area for families? No Yes Existing organisation of ASGMs? Yes No Note. ASGMs ¼ artisanal small-scale gold miners a All residents in the remote site of Gaang were miners. In Camarines Norte, mining activities were carried out within the villages. health-care workers to have a substantial knowledge about prevention, diagnosis, and treatment of mercury poisoning. Furthermore, the project aimed to set up a mercury committee in each mining community that would meet regularly and discuss issues related to mercury use, and how to overcome challenges in shifting to mercury-free methods. The project activities were carried out by the Manila-based NGO Ban Toxics! (Bantox) assisted by the Benguet Federation of Small-Scale Miners. Bantox staff coordinated and managed the interventions. The Danish NGO Dia logos supervised and supported the project in cooperation with the Geological Survey of Denmark and Greenland (GEUS). The project was approved by the Department of Health in the Philippines and by the local city councils in the project areas. The Interventions The core activity of the project was hands-on miner-to-miner training in GBM. During the training sessions, the miners were informed about the occupational and environmental hazards of mercury. The training sessions took place in a local milling station similar to the conditions under which the ASGMs normally work in groups to refine their ore. One or two experienced ASGMs from Benguet undertook the training assisted by two to three Bantox staff members. A typical training program lasted one or two days and fifteen to twenty-five local ASGMs participated. The project provided no economic incentive to shift to the new method. Local health-care workers (doctors, nurses, and midwives) attended a two-day seminar carried out by two Danish doctors from Dia logos and a Filipino doctor who specialized in toxicology. In Kalinga, the seminar was attended by sixteen health-care workers, and in Camarines Norte the seminar

5 Køster-Rasmussen et al. 571 was attended by twenty-three. The teaching manual is available upon request to the corresponding author. The training of school teachers was conducted by Bantox staff. Seminars lasted two hours and covered mercury toxicology, the use of mercury in small-scale gold mining, GBM as a mercury-free alternative, and the need to pass on knowledge to fellow teachers and students. Bantox staff held several seminars in local schools in the project areas using narrative techniques to teach the children about the dangers of mercury and to encourage them to advocate for mercury-free mining. Numerous meetings between Bantox staff and local politicians and community leaders were held, and whenever possible attended by representatives from Dia logos or GEUS. Local political leaders were involved throughout the project, and there was continuous dialogue and interaction between Bantox, the local city council, and administration in both project areas. Various other activities were carried out to raise awareness about the mercury pollution, and a broader audience was targeted through exposure in local and national media as well as social media platforms. The aim of these activities was to encourage a broad community base to take action against mercury pollution, to secure local ownership of the process, and to increase the likelihood that the change would be sustainable. The overall project methodology with ample involvement of the civil society is in line with the Strategy for Danish Support to Civil Society in Developing Countries issued by the Ministry of Foreign Affairs in Denmark. 9,10 Outcome Measures and Data Collection The main indicator of success was a reduction in the miners use of mercury and adoption of the GBM or another mercury-free alternative to the amalgamation techniques. A baseline survey of ASGMs was conducted in September October 2011 using two questionnaires for structured interviews (see online Supplementary Material). Questionnaire 1 concerned mining practices and detailed information about the average monthly use of mercury. Mercury consumption was quantified as grams purchased per month in three different ways: (1) the miners one-month recall of individual purchase, (2) the miners one-month recall of mercury purchased by their milling station, and (3) recordings in the accounting books of the milling stations, where the monthly mercury consumption was calculated as the mean monthly purchase of mercury in the last year according to the accounting books. Questionnaire 2 concerned ASGMs knowledge and attitudes about the use of mercury. The miners were asked about exposure routes, means to avoid mercury exposure, and precautions taken when handling mercury (four, zero, and ten miners did not answer these categories of questions, respectively). Regarding motivation for changing method, the miners were asked to prioritize how important the following six factors were on a scale from 1 to 6 (1 being the most important): their income, the cost of extraction method, their own health,

6 572 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) the health of family and other workers, pollution of water and soil, and how fast and easy the work is performed. Unfortunately, no valid baseline data are available from Questionnaire 2 from Kalinga due to incomplete instructions given to the interviewers there. Thus, changes in knowledge, attitudes and priorities could only be assessed for miners in Camarines Norte where 208 miners filled in Questionnaire 2 at baseline (which is twenty more than those who answered Questionnaire 1). The follow-up interviews of ASGMs were conducted in September November 2013, and the same questions were used as in the baseline survey. We attempted to reinterview the participants from the baseline study: the longitudinal follow-up. However, only seventy-four miners were still living in the areas and could be interviewed again (21%). In addition, we interviewed a randomly chosen sample of fifty to fifty-nine ASGMs from each village: the cross-sectional follow-up. One of the mining villages in Camarines Norte (Paracale) could not be visited during the cross-sectional follow-up period due to civil unrest in the area. All participating miners answered both Questionnaires 1 and 2. A specific questionnaire was developed to assess health-care workers knowledge about mercury toxicology at baseline, and we attempted to reinterview all participants with the same questionnaire at follow-up after two years (see online Supplementary Material). Focus group interviews of school children, school teachers, and other representatives from civil society were carried out by Bantox. Dia logos, GEUS, and Bantox developed the questionnaires and interview guides together. Bantox staff conducted the baseline and follow-up interviews by visiting milling stations, health clinics, schools, etc. The questionnaires were translated to the local languages and Bantox staff filled in the questionnaires in accordance with the miner s answers. The health-care staff questionnaire was in English and was filled in individually. Dia logos supervised the data collection process and the handling of the data. Bantox staff encoded the raw data in Microsoft Excel files, and Dia logos conducted the analyses. Diffusion of Innovations Theory To explore enabling and hindering factors in phasing out mercury by promoting the use of GBM, we used the theory about diffusion of innovations. 11 When implementing new technologies, the major factors influencing variations in adoption rates are the five main perceived attributes of innovation described by Rogers 11 : relative advantage, compatibility, complexity, trialability, and observability. A key assumption in planning the project was that a change of practice must be supported by stakeholders in the society surrounding the individual. Therefore, we included two other explanatory elements mentioned by Rogers, 11 namely the nature of the social system and type of innovationdecision. Besides the baseline and follow-up studies, the project was continuously monitored with monthly field trips by Bantox and four field visits by

7 Køster-Rasmussen et al. 573 Dia logos or GEUS and evaluated with quarterly reports and meetings between the project partners. The project was prospectively planned, but the diffusion of innovation analysis was retrospectively conducted and mainly based not only on unstructured field observations from the monitoring but also on results from the questionnaire surveys and focus group discussions. Results Whole-ore amalgamation was used by 94% 100% of ASGMs in the project areas before the interventions started (Table 2). The baseline purchase of mercury was more than ten times higher in Camarines Norte compared with Kalinga. Apparently 1.0 to 1.5 g of mercury was used to produce 1 g of gold in Kalinga, whereas approximately 10 g of mercury was used to produce 1 g of gold in Camarines Norte. Field observations by GEUS, Dia logos, and Bantox staff indicated that the low mercury use in Kalinga was to some degree explained by a relatively benign way of practicing whole-ore amalgamation with multiple remillings. In Camarines Norte, there was great variation in mercury consumption between milling stations and between villages. Labo reported low consumption, whereas Jose Panganiban and Paracale reported extremely high consumptions. Field observations indicated that these differences were probably due to different work practices. There were many compressor miners and other single-person operators in Labo; in contrast, larger groups of miners were operating larger milling station units in Jose Panganiban and Paracale. Changes in Mining Practice In total, 355 ASGMs participated in the baseline study. We succeeded in reinterviewing seventy-four of these miners after two years, i.e., the longitudinal follow-up study. In addition, 167 miners (who did not participate in the baseline survey) were interviewed for the follow-up study, i.e., the cross-sectional followup. Thus, 241 ASGMs were included in the follow-up. The main results in changes in mining practice are summarized in Table 2. At follow-up, only one miner in Kalinga reported using mercury. The rest had converted to GBM, and all participating miners reported that they were no longer buying mercury. In contrast, only one miner in Camarines Norte confirmed having converted to GBM. The rest were still using and purchasing mercury for gold extraction. Thus, a major reduction in mercury use, toward zero, was observed among the miners in Kalinga but not among their counterparts in Camarines Norte (Table 3). Changes in knowledge and attitudes were only assessed in Camarines Norte. Figure 1 shows the miners knowledge of mercury exposure routes, means to avoid exposure, and precautions at baseline and after two years. Overall, there

8 574 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) Table 2. Changes in Mining Practice Philippine Artisanal Small-Scale Gold Miners Use And Purchase of Mercury at Baseline (2011) and at Follow-Up (2013). Kalinga Camarines Norte Calculation Result Calculation Result The baseline study (Oct 2011) Miners participating, n Milling stations represented, n Mercury was used for gold extraction, % miners 33/ /98 94 a Gold production in the last month, mean g/ms 2354/ ,442/ Purchase of mercury in the last month Milling station, miners recall b, mean g/ms 2475/ ,318/ Milling station, accounted for in books c, mean g/ms 2335/ ,601/ Individually, miners recall, mean g/miner 1962/ ,274/ The longitudinal follow-up study (Sep Oct 2013) Miners participating, n Milling stations represented, n 8 16 Mercury was used for gold extraction, % miners 1/ /49 92 d Gold production in the last month, mean g/ms NR NR Purchase of mercury in the last month Milling station, miners recall b, mean g/ms 0/8 0 32,905/32 e 1028 Milling station, accounted for in books c, mean g/ms 0/3 0 15,181/21 e 723 Individually, miners recall, mean g/miner 0/ ,899/ The cross-sectional follow-up study (Nov 2013) f Miners participating, n Milling stations represented, n 11 7 Mercury was used for gold extraction, % miners 0/ / a Gold production in the last month, mean g/ms NR NR Purchase of mercury in the last month Milling station, miners recall b, mean g/ms 0/ ,025/31 e 1227 Milling station, accounted for in books c, mean g/ms 0/5 0 18,826/15 e 1255 Individually, miners recall, mean g/miner 0/ / Note. MS ¼ Milling station, NR ¼ Not reported. a The miners not using mercury did not specify what method they used for gold extraction. b The miners one-month recall of mercury purchased by the milling station How much mercury did your milling station purchase during the last month? c Mercury purchase as accounted for in the books of the milling stations. Monthly mercury consumption was calculated as the mean monthly purchase of mercury in the last year. d One miner confirmed having converted to the borax-gravity method, e Several miners reported from each milling station. f Paracale was not included in the cross-sectional follow-up as it was not safe to travel there due to civil unrest in the area.

9 Køster-Rasmussen et al. 575 Table 3. Estimated Mercury Consumption by Small-Scale Gold Miners in the Project Areas Before and After the Intervention. Baseline kg mercury/year Longitudinal follow-up kg mercury/year Cross-sectional follow-up kg mercury/year Miners recall a Accounted for in the books b Miners recall a Accounted for in the books b Miners recall a Accounted for in the books b Kalinga c Camarines Note. The yearly estimated consumption of mercury was extrapolated from the numbers given in Table 2. For example, in Kalinga 156 g/ms/month was accounted for in the books: 156 g/ms 165 MS 12 months ¼ 309 kg/year. a The estimate is based on the miners one-month recall of mercury purchased by the milling station. b The estimate is based on mercury purchase as accounted for in the books of the milling stations. c One miner (out of twenty-four) reported that he still used mercury, but he did not report his mercury consumption. was no pattern supporting improvement. Still, Figure 1(b) indicates some progress in knowledge of how to avoid mercury exposure. For instance, more miners at follow-up than miners at baseline knew that using a vapor-specific mask reduces exposure and that using a piece of cloth does not. Compared with baseline, miners seemed to have the same or better knowledge of how to avoid mercury exposure (Figure 1(b)); nevertheless, fewer reported taking precautions (Figure 1(c)). The results presented in Figure 1 are based on data from the longitudinal follow-up. The results from the cross-sectional follow-up did not differ substantially. In Camarines Norte, 50% (twenty-three of forty-six) of health-care workers who participated in the baseline study were reinterviewed at the follow-up study, and in Kalinga, the proportion was 25% (four of sixteen). The health-care workers attending the follow-up survey in Kalinga had major improvements in their understanding of mercury toxicology, whereas only a minor improvement was seen among their counterparts in Camarines Norte. The results are outlined in Table 4. A mercury committee was established in both project areas. Seminars for school teachers were attended by 269 teachers from Kalinga and 234 from Camarines Norte. Two films describing GBM were produced by the project. The ten-minute film Rather rich and healthy than poor and poisoned has been seen 415,000 times on YouTube (as of July 2015). 12 Diffusion of Innovations Analysis Relative advantage. Most miners in the two project areas accepted that mercury is toxic and that GBM is healthier than the amalgamation method. Borax is very

10 576 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) Figure 1. Knowledge, attitudes, and practices regarding mercury exposure among 45 small-scale gold miners from Camarines Norte with valid data from both baseline in 2011 (black bars) and at follow-up in 2013 (grey bars). Level and change in (a) knowledge of exposure routes, (b) perception of means to avoid exposure, and (c) precautions taken when handling mercury.

11 Køster-Rasmussen et al. 577 Table 4. Health-Care Workers Knowledge about Mercury Toxicology Before and After the Intervention Assessed as the Proportion of Participants Who Correctly Answered Questions on Mercury Exposure and Poisoning. Kalinga (n ¼ 4) Camarines Norte (n ¼ 19) Baseline Follow-up Baseline Follow-up Exposure Inhalation 100% 100% 84% 85% Fish 50% 100% 0 16% Transplacental a 25% 75% 5% 16% Symptoms Tremor 0 50% 0 5% One neuropsychiatric 25% 25% 10% 21% Multiple neuropsychiatric 0 50% 0 0 Developmental delay 25% 75% 0 5% Respiratory 0 50% 37% 58% Trained by the project 75% 74% Note. Only health-care workers that participated in both the baseline in 2011 and in follow-up study in 2013 are included in this table. None of the changes from baseline to follow-up were statistically significant (Chi 2 p > 0.05). a The question was phrased as in the womb, from mother to unborn child. cheap, whereas mercury is comparatively expensive. Mercury trade and possession is illegal, but this is only rarely enforced. Some miner leaders, especially in Kalinga, emphasized that eliminating the use of mercury would ease the way toward minahang bayan (people s mining or legal small-scale mining area). After gaining experience with the new method, miners in Kalinga recognized that the extraction rate was higher with the GBM, whereas in Camarines, the miners were never convinced about that. The major disadvantage of GBM is that it is slightly more labor intensive, as the miner has to reduce the crushed ore into a mineral concentrate by sluicing and panning. However, the total time consumption is approximately the same as with the whole-ore amalgamation method, as the ore has to be milled with mercury. Questionnaire 2 conveyed that the gold yield was more important for the miners than time or extra work. The miners highest priorities were their income and their health, whereas how fast the method was or if it caused pollution were of less importance. This pattern was only slightly changed at follow-up (Table 5). Compatibility. Among the Camarines miners, mercury was not perceived as a substantial problem before the intervention started, whereas in Kalinga it was. For instance, pregnant women and children were not allowed in the mining area

12 578 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) Table 5. Ranking of Priorities Regarding Gold Extraction Methods among Small-Scale Gold Miners in Camarines Norte. Priorities a Baseline (2011) Mean (SD) N ¼ 208 Follow-up (2013) Mean (SD) N ¼ 50 Income 2.8 (1.6) 2.5 (1.5) Cost of the method 4.5 (1.3) 4.5 (1.3) Own health 2.1 (1.1) 2.6 (1.4) Health of family 2.3 (1.2) 2.6 (1.0) Pollution 4.6 (1.2) 4.9 (1.4) How fast 4.7 (1.3) 3.8 (1.8) a Priorities are ranked 1 to 6, with 1 being the highest priority. in Kalinga because of the mercury usage. All miners in Kalinga are from the Banao tribe, and the tribe owns the mining area and the land downstream for many miles. In contrast, the miners in Camarines are not tribal but more ethnically diverse and many are migrant workers. The land is owned by the government and a patchwork of small and large private landowners. All the rivers drain directly into the sea. The miners in Kalinga cannot sell their tailings for further processing in a cyanide plant because of the remote location of the mining site. Therefore, the Kalinga miners did multiple remillings to improve the extraction rate, and GBM was perceived as a way to further improve yield. In Camarines, gold production and the turnover of ore were much larger compared with Kalinga, but the extraction rate was considerably lower. Complexity. The miners in both areas were familiar with the gravity element of the GBM, but washing pans were not used in the areas before the intervention started (in Kalinga, they had never seen a washing pan before). Borax was a known and readily available chemical in the mining areas. After amalgam burning, a bit of borax was often added and burned to make the surface of the gold shine. Many miners found it hard to believe that borax can be used as a gold catcher in the same way as mercury, until they had the technique demonstrated. Still, compared with the amalgamation method, the GBM was perceived as more complex. Trialability. First, using a washing pan and a launder had to be learned and then the smelting with borax. The miners who converted to GBM often had these processes demonstrated several times at the miner-to-miner training sessions, and often the trainers would visit the local milling facilities and supervise how the gold washing facilities could be improved. A drum contaminated with mercury has to be cleaned very carefully with chemicals or even replaced before

13 Køster-Rasmussen et al. 579 GBM can be successfully tested (otherwise, the yield is low because mercury left in cracks and corners absorbs the gold the first couple of times the drum is used without mercury). Thus, for a miner, it is very inconvenient to try the new method on his own, if the milling station is using mercury. Observability. When using the new method, the miners observed immediately that they did not have to buy mercury, but it could not be recognized right away that the extraction rate with GBM can be higher as it takes some practice to refine the method. Also, the gold content of the ore varies considerably from sack to sack, and many miners preferred the safe amalgamation method with rich ores and were more likely to experiment with GBM with a less rich ore. Neither short- nor long-term health benefits were immediately recognizable. Nature of the social system. All miners in Kalinga belong to the Banao tribe and most of them recognize the board of the Banao Bodong Association as an authority. The leader of the Association was also the mayor of Balbalan (the largest town in the area with police station, health clinic, and schools). The Banao Bodong Association was able to make decisions regarding mining practices and to enforce these by means of their marshals (who are also ASGMs and some of them became miner-trainers in GBM). In contrast, most inhabitant in three project villages in Camarines Norte did not belong to any tribe. There, the city council included landowners and milling station owners with limited interests in acknowledging that they had polluted the area for decades, and decisions or directions from the council were often changed or not enforced. In general, the political situation in Camarines Norte was more unstable. In both project areas, the miners were directly affected by project posters, pamphlets, and mass media and indirectly by project activities educating the health staff, teachers, school children, and the local government unit. In Kalinga, women participated in the mining activities, whereas in Camarines they did not (Table 1). Type of innovation-decision. In Kalinga, the milling station units are relatively small. Most have two drums, compared with typically twenty to thirty drums per milling station in Camarines. In Kalinga, a milling station is run by a group of five to fifteen miners who share their earnings. In Camarines, the milling stations are organized in multiple ways and many are based on leasing agreements per hour, day, or week. After the first miner-to-miner trainings in Kalinga, two or three milling stations converted, which was an optional decision and at the same time a collective decision among the group of miners running the milling station. More milling stations followed, and after being convinced that the new method worked, the Banao Bodong Association decided that all milling stations in the area should convert over the next year. Thus, the success in Kalinga was eventually an authority innovation-decision. In Camarines, the decision to change would have to be taken by a land owner or a milling station

14 580 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) owner, and the change would require a simultaneous training of all the miners using the milling facilities (which was offered by the project) and a strict enforcement of a zero-mercury policy afterwards also among miners leasing the facility. Two test facilities were built in Camarines, but they were both contaminated with mercury after a short time. Thus, the innovation-decision remained optional in Camarines. Discussion We confirmed that the consumption of mercury in small-scale mining was massive especially in Camarines Norte. We demonstrated that it is possible to implement the GBM in an area using whole-ore amalgamation outside Benguet. In Kalinga, the mercury consumption dropped to near zero. However, in Camarines Norte, only a few miners converted to GBM and the mercury consumption was not substantially reduced. Perceived Attributes of the Innovation The relative advantages and disadvantages of GBM were largely the same in the two project areas. However, the miner leaders in Kalinga were more aware that the use of mercury in the future could turn into a legal problem. The GBM was more compatible with the existing cultural values in the mining community in Kalinga compared with Camarines, where the idea that mercury is toxic was new to many miners and stakeholders. Relative to the more heterogeneous population in Camarines, the Kalinga miners had a strong interest in not polluting the land as it was owned by their own tribe. The miners from Kalinga seemed more concerned about getting higher extraction rates than the Camarines miners who were able to sell their tailings. In both project areas, the miners were familiar with borax and the gravity aspect of GBM. The new method was probably equally complex for the miners in the two areas. In terms of trialability, it was a major hindering factor that the GBM only works well in a facility not using mercury. Even though this condition was the same in both project areas, the smaller milling station units in Kalinga made it easier for the miners to designate a milling station as mercury-free. The advantage of GBM was often not easy to observe right away, but observability did not differ between the project areas. The diffusion of innovations analysis also shed light on differences in the nature of the social systems in the two project areas and indicated that the striking difference in project impact could be partly explained by the diverging degree of organization of the miners and the power of having a legal body that was able to enforce decisions. In Kalinga, the Banao Bodong Association supported the project, encouraged its members to shift to GBM, and eventually played a vital role in the civil control of mercury use in the area.

15 Køster-Rasmussen et al. 581 The different participation rate of women reflects some of the cultural variation between the two project areas and the participation of women could well be an additional factor in favor of the change observed in Kalinga. Miners, tribal leaders, and health-care staff were taught that mercury exposure to fertile women any time before childbirth was especially harmful. Since mercury is of greater direct concern to women, the diffusion network may have been particularly effective in reaching women. 13 In Kalinga, young women make up a considerable part of the ASGM workforce, whereas in Camarines Norte, the women are not involved much in the mining activities. Still, in Camarines Norte, the mining takes place inside the villages among the women and their children, and the mercury pollution should be an incentive for change for them as well. Mertens et al. 14 have described how women and men may be more concerned with areas of expertise that are commonly regarded as specific to their sex. 13,14 Furthermore, men and women tend to discuss health issues preferentially with individuals of the same sex, and thus behavior can be expected to be more strongly influenced by members of the same sex. 13,14 The health-care staff consists primarily of women, and in Kalinga, the health-care workers reached a fair level of knowledge regarding the pollution. Perhaps the female miners in Kalinga were more influenced by the health-care workers than their male colleagues in both Camarines Norte and Kalinga. Thus, the different impact of the interventions in the two project areas may in part be explained by the fact that women are less often involved in mining activities in Camarines Norte and by the poor understanding of mercury toxicology by the health-care workers in Camarines Norte. In line with the nature of the social systems, the type of innovationdecision differed between the project areas. In Camarines Norte, there was no existing miners organization or any other appropriate civil society organization to support the transformation of the optional innovation-decision into an authority innovation-decision like in Kalinga. Changes in Knowledge and Attitudes among Miners and in the Civil Society There was no consistent trend toward improvement of the miners knowledge, attitude, or practices in Camarines Norte. This was probably because the training sessions highlighted manual skills transfer in the GBM, and not enough time was spent on emphasizing occupational hazards and preventive measures. This is, however, an important aspect as studies repeatedly have shown elevated levels of mercury in miners and residents. 15 In contrast, health-care workers in the project areas were trained specifically in these issues, but even though they appeared to have increased their knowledge (at least in Kalinga) this did not necessarily spread to the miners. Although there were no convincing reductions in mercury use in Camarines Norte, the local authorities and political leaders

16 582 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) were, and still are, supportive of the project. At the time of writing, local authorities in one of the villages, Jose Panganiban, requested Bantox to continue project activities in the area. Thus, the activities and the advocacy done so far may actually have paved the way for future changes. Borax Toxicology Borax (disodium tetraborate decahydrate or sodium tetraborate) is a salt of the element boron. The GBM involves the use of two to three spoonfuls of borax. During the gold extraction process, the mixture of heavy minerals (containing the gold) and borax is heated. The water contained in the borax evaporates and the boron smelts to a glass-like liquid and does not vaporize to any large extent. 4 Boron is found naturally in drinking water, fruits, and vegetables. It is used in production industries; for instance, in fabrication of glass and insulation material, and it is an additive to laundry detergents and fertilizers as certain plants need boron. Boron is marketed as a dietary supplement in some countries. It has been anticipated that borax is largely non-toxic except in very high doses. Studies of rodents such as mice, rats, and rabbits have shown adverse effects on spermatogenesis when administered by ingestion and in high doses. 16 However, a recent review conducted by the European Food Safety Authority summarized data from animal and human toxicology studies and concluded that borax does not raise concern for genotoxicity. 17 Still, these findings in animal studies have resulted in boron compounds being added to the EU candidate list of substances of very high concern. 18 This means that producers or importers of articles containing borates are required to notify the European Chemicals Agency and to provide customers with a safety data sheet. At the same time, it is allowed in the EU to use borates in cosmetic products such as eye and oral hygiene products. 17 In contrast to the animal studies, studies of boron miners exposed to inhalation of high boron concentrations have not been able to detect any adverse effects on sperm quality or fertility of the miners. 19 Compared with the exposure of boron miners, ASGMs using GBM have a very low exposure. There is no visible dust when handling borax and there is very little or no evaporation of boron when heated because boron vaporizes at a very high temperature. Borax does not penetrate the intact skin. 16 Thus, the likelihood of GBM causing harmful exposure in the workplace setting is very low. All together, the use of borax in ASGM is much less toxic than the use of mercury. Limitations The main limitation of this study is probably measurement error due to information bias. Data on gold production and mercury consumption were likely underreported. Especially at follow-up, we assume that our data on mercury

17 Køster-Rasmussen et al. 583 purchase were affected by social desirability bias, i.e., underreporting. The relatively low baseline consumption in Kalinga is probably also to some degree explained by underreporting, since Bantox staff began advocating for mercury-free mining in the area already six months prior to the baseline study. However, we do not think that the zero-result from the follow-up interviews in Kalinga was due to information bias since widespread implementation of GBM was confirmed by numerous field observations by Bantox, Dia logos, and GEUS. In contrast, the apparent reduction in mercury use in Camarines Norte is likely, to some degree, explained by social desirability bias. However, the decline was also in part explained by increased use of cyanide plants for processing the ore, and the fact that follow-up data unfortunately could not be gathered from Paracale village where ASGMs were known to use large quantities of mercury. The extrapolation of the quantitative data to yearly consumption as presented in Table 3 relied on the assumption that the participants in the questionnaire surveys were representative of the whole population in the mining communities. However, the participation rates could not be estimated for this study. At baseline, all miners were invited to participate in the project, but many were skeptical about the efficacy of the method. Small-scale mining with mercury is illegal in these areas, and there was some suspicion about the purpose of the data gathering. Therefore, it is difficult to know if the baseline sample was truly representative for the population. Regarding the longitudinal follow-up, only 21% of the baseline participants were identified and reinterviewed. The main explanation for this was migration of the miners. Thus, the participants in the longitudinal follow-up were probably a rather selected group of seasoned miners, whereas the participants in the crosssectional follow-up were randomly chosen and therefore more likely representative of the background population of miners in the two areas. We have no reason to believe that those interviewed were more or less knowledgeable or had a smaller or larger mercury use than miners not interviewed. Altogether, despite multiple sources of measurement error and bias, we regard it as beyond doubt that the level of mercury purchase dropped to almost zero in Kalinga. The reported improvement in health-care workers knowledge about mercury toxicology was based on relatively small numbers as only a few were available for reinterview. Thus, the participants at follow-up consisted of a selected group of health-care workers who were still around, and the results presented should be interpreted in that light. Still, we think that the trend described is valid, as the health-care workers in Kalinga were extraordinarily eager to improve the health of their tribe and to avoid a pollution disaster. For instance, two of the four health-care workers from Kalinga were able to correctly identify seven neuropsychiatric symptoms of mercury poisoning in the follow-up survey. A similar eagerness was absent among counterparts in Camarines Norte, which is rather

18 584 NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 25(4) unfortunate as the mercury burden of this area is much more severe than in Kalinga. Future Research Work GBM was successfully implemented in Kalinga, but whether GBM can be implemented in other mining communities remains to be tested. New ways to identify and involve powerful stakeholders in the social system needs to be explored. It was beyond the scope of our study to test other mercury-reducing technologies like retorting or cyanidation. Occupational hazards of borax when used for gold extraction should be further investigated. It would be very informative to compare the urine concentration of borax in miners using GBM with miners not using borax. Conclusions The consumption of mercury in small-scale gold mining is massive, but GBM is a doable alternative to the traditional use of mercury including whole-ore amalgamation techniques. The mercury-free GBM was adopted by almost all smallscale miners in the mining community Gaang in Kalinga during the two-year intervention period, whereas only few miners adopted GBM in Camarines Norte. Project activities included not only miner-to-miner training but also seminars for health-care workers, school teachers, school children, and involvement of community leaders and organizations to anchor the change in the civil society and secure sustainability. Differences in compatibility with existing ideas, cultural values, and social and organizational structures in the two project areas impacted the success of the implementation. Involvement of the tribal organization was an important factor enabling the shift in Kalinga. An alliance with a small-scale miner organization or another influential actor in the civil society appeared to be crucial for a successful implementation of the new technology. Thus, the emerging mercury pollution disaster from small-scale gold mining can be stopped by introducing GBM, but a successful implementation seems to depend on involvement of a miner organization or another appropriate civil society structure that can support and facilitate an optional innovation-decision or enforce an authority innovation-decision. Acknowledgments We thank Gil Veloria Jr., Abel Felix, Allan Camacho and the staff of Ban Toxics! for collecting data and implementing the intervention activities, often under very challenging circumstances. We would also like to thank Leoncio Na-Oy and his crew of miner trainers from Benguet. Furthermore, we thank Dr. Erle Castillo and his colleagues at the Philippine Society of Clinical and Occupational Toxicology (PSCOT) for their contribution in training and preparing teaching materials for health-care workers. The study

19 Køster-Rasmussen et al. 585 would not have been possible without help from representatives of the Banao Badong Association in Kalinga, the Philippine Department of Health, the city councils, school teachers, and health workers in the project areas, and of course from the miners themselves. Thanks to Tilde Mu ller for contributing to collecting the baseline data, and thanks to Sofie Jo Anna Helm Pedersen and Marius Brostrøm Kousgaard for assisting in applying the Diffusion of Innovation theory. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project was funded by the Civil Society Fund which is financed by the Danish Ministry of Foreign Affairs and administered by Civilsamfund i Udvikling (CISU, References 1. Veiga MM and Baker RF. Protocols for environmental and health assessment of mercury released by artisanal and small-scale gold miners. Vienna, Austria: GEF/UNDP/ UNIDO, 2004, p Perez E, Appel PWU and Koester-Rasmussen R. Training of small scale miners and their families in safe handling of mercury during extraction of gold in the Philippines. Improving access to social services: health services and income opportunities for small scale miners and their families. Report No: 2007/35. Denmark: Geological Survey of Denmark and Greenland (GEUS), 58 pp, geus_rep_ pdf (2007, accessed 10 September 2015). 3. Arctic Monitoring and Assessment Programme (AMAP)/The United Nations Environment Programme (UNEP). Technical background report for the global mercury assessment Oslo, Norway/Geneva, Switzerland: AMAP/UNEP, 2013, p Appel PWU and Na-Oy L. The Borax method of gold extraction for small-scale miners. Blacksmith Institute J Health Pollut 2012; 2: Jønsson JB, Charles E and Kalvig P. Toxic mercury versus appropriate technology: artisanal gold miners retort aversion. Resour Policy 2013; 38: Babut M, Sekyi R, Rambaud A, et al. Improving the environmental management of small-scale gold mining in Ghana: a case study of Dumasi. J Clean Prod 2003; 11: Cordya P, Veiga MM, Saliha I, et al. Mercury contamination from artisanal gold mining in Antioquia, Colombia: the world s highest per capita mercury pollution. Sci Total Environ 2011; : Veiga MM, Nunes D, Klein B, et al. Mill leaching: a viable substitute for mercury amalgamation in the artisanal gold mining sector? J Clean Prod 2009; 17: DANIDA. Policy for Danish support to civil society. Copenhagen, Denmark: Ministry of Foreign Affairs of Denmark, 2014, p.40.