a Occupational and Environmental Health Research Unit, Department of Public Health and Primary Health Care, Anzio Rd, University of Cape Town, South Africa. E-mail: ll{at}anat.uct.ac.za
b Menzies School of Health Research and Flinders University Northern Territory Clinical School, Darwin, Australia.
Abstract
Background Surveillance is a critical public health tool for the control of pesticide poisoning. However, surveillance activities in developing countries are bedevilled by multiple problems, and inferences made from review of flawed data may lead to mistaken policy decisions.
Methods Results of intensified surveillance from an intervention project in the Western Cape Province of South Africa were compared to the pattern of poisonings reported in routine notifications to the health authorities for a control farming district and in the study district over a 5-year period preceding the study. Intensified surveillance data results were also contrasted with policy approaches based on routine notifications and on Regional Poison Centre reports.
Results Poisoning rates reported in the study area increased almost 10-fold during the intervention period. Compared to intensified surveillance, hospital and health authority sources greatly underestimate the proportion of cases due to occupational poisoning, and overestimate suicide as a proportional cause. In addition, the risks for women appear underestimated from routine notifications. Assumptions that a lack of awareness is responsible for most poisonings are not borne out by the empirical data when reporting is intensified.
Conclusions Current policy assumptions are faulty, may result in inappropriate blame being attributed to victims and, by relying on information as the main element of education, may shift responsibility onto the individual. Improvements in the surveillance system should aim to restructure the types of data collected, and facilitate intra-governmental and inter-sector collaboration. The culture of monitoring based on report writing must change to one of surveillance that leads to intervention.
Keywords Pesticide poisoning, surveillance, policy, gender, suicide, occupational poisoning, public health intervention
Accepted 20 July 2000
Pesticide poisoning is a major public health problem in developing countries15 particularly in settings of low education and poor regulatory frameworks.6,7 Pesticides usage in South Africa, both agricultural and non-agricultural, has increased substantially in the past decade8,9 and this country is the largest market for pesticides in sub-Saharan Africa.7
Expansion of the small-scale farming sector is a key government strategy for future economic growth in the country.1012 However, this strategy does not address the need for measures to prevent hazardous exposures in the most vulnerable of farming sectors.9
In this context, key to the control of pesticide-related morbidity and mortality is the need for accurate, timely and effective surveillance systems.13,14 Above all, for a surveillance system to be effective, it has to demonstrate usefulness of the data generatedappropriate action needs to follow reporting of the data.
This paper examines the monitoring and surveillance system in South Africa from three perspectives: (1) policy inferences in government publications and other documents; (2) review of previous research findings related to completeness of reporting; and (3) implications of the findings of an active surveillance project for current policy.
Surveillance for pesticide poisoning in South Africa
In South Africa, surveillance for pesticide poisoning occurs across a number of different information systems (Figure 1) involving at least three different government departments. The major sources of surveillance data lie in the health sector, relying on statutory notification by medical practitioners in terms of the Health Act.15 Follow-up investigation and remediation is the responsibility of environmental health officers employed at the local authority level. Other potentially important sources of data reside with the Department of Labour (work-related morbidity and mortality), and the Department of Home Affairs (cause of death data derived from death certificates). The typical pattern that emerges from such routine surveillance is that approximately 100200 cases of human pesticide poisoning are reported nationally, of which about 1020 are fatal.16 Government control activities are integrated at national level in an interdepartmental committee responsible for pesticide safety. In theory, this structure co-ordinates surveillance for the human health impacts of pesticides although most weight is attached to Department of Health surveillance data.9
|
An assessment from a major Poison Centre framed the critical causal pathway for pesticide poisoning as follows: These agents (pesticides) are freely available in shops and convenience stores, and there exist no special measures to inform the buyer of the dangers of the pesticide they will use. This can give rise to careless and negligent use of these agents. ... An education programme is urgently needed.18 (Authors' emphasis italicized.)
This conclusion implies that the source of the problem rests with the end-user, for which the solution lies in individualized strategies aimed at behaviour change. For example, in reflecting on notifications of pesticide poisoning since inception of the system, the Department of Health16 concluded that the urgent need for very specific training and education is reflected in the ignorance factor taking on such a dominant role .
Routine surveillance also generates a picture that most cases of poisoning are predominantly male (67%) and that over 50% of all cases are suicide attempts.16,22 Occupational causes are therefore seen as relatively less important in the scale of causation: Contrary to popular belief, acute organophosphate poisoning as a result of exposure of farm labourers during the spraying season is relatively uncommon. In fact, surprisingly few serious acute pesticide poisonings occur as a result of exposure during farming activities. Those reported from farming areas usually fall within the household category, e.g. in cases where the pesticide has been brought into the house in an unmarked container.21 (Authors' emphasis italicized.)
Key policy inferences regarding pesticide poisoning of current surveillance procedures and information are summarized in Table 1.
|
Firstly, there is inadequate capture and inadequate sharing of data between databases. A review of deaths registered at mortuaries in rural towns in the Western Cape province from January 1990 to July 1992 found that 10% of work-related deaths were due to pesticides, of which none had been reported to the Department of Labour or to the Department of Health.23 All these deaths had been registered with the Department of Home Affairs. The study highlighted the lack of co-ordination of different information systems.
Secondly, under-notification of pesticide poisoning is repeatedly identified in the literature.16,22,24,25 Estimates of the percentage of hospitalized cases notified range from 10%24 to 20%,22 and the percentage of fatal cases at mortuaries notified is estimated at around 5%.16,25
The third problem, generic to health information systems throughout South Africa, relates to data quality. Too many data are collected, with little reference to relevance or to any prioritization of what is important data to collect.26,27 For example, the toxicological investigation form used by local environmental health officers contains the categories of suicide, accident, carelessness and other to categorize cause of the poisoning, and has over 50 variables for completion, few of which are ever used in reviewing public health action. This results in a situation where the data on which public health policy on pesticides is based may be potentially flawed, raising questions around how true are the inferences contained in Table 1 for South Africa.
The impact and findings of intensified surveillance
A community-based project to determine the impact of a public health intervention on poisoning notification in a rural district of South Africa suggests several deficiencies of current routine surveillance. The intervention included the provision of free cholinesterase testing to local GPs, and provision of information, training and educational materials to all local health care providers. The intervention28 and its impact on GP behaviour29 are described in more detail elsewhere. The annual mean rate of notification of cases per 100 000 population in the study area for the 5-year period 19871991 was 4.2 while the rate for the study year was almost 10-fold greater at 40.5 per 100 000. By comparison, the rate in all surrounding areas together was more or less unchanged (3.1 per 100 000 per year in the period 19871991, and 3.9 per 100 000 in the study year). The increase in the study area before and after the intervention and the difference between study area and control were both statistically significant.
The categories negligence, accident or ignorance were cited by environmental health officers as the main cause of poisoning in 42% of cases. Table 2 summarizes the circumstances of the poisoning cases detected in the intervention area. Poisonings directly related to occupation involved 86% of the cases and 43% of the poisoning events, whereas suicide involved 9% of the cases and 36% of the poisoning events. Two events were mass poisonings, affecting 20 and 24 people, respectively, and these occupational poisonings were also more likely to affect women.
|
|
Many of the poisonings in this study occurred under circumstances of normal use (cases 7,8,1012) or under circumstances over which subjects had effectively no control (case 14). In other cases, lack of awareness played a role only as the final part of a sequence of social factors. For example, family members of farm workers poisoned by pesticides taken home for domestic use (as in case 12) are subject to an environment of easy availability of chemicals, experience poor household sanitary conditions leading to domestic pests, and lack easy movement off their farms to acquire safer means of pest control.
Completeness of reporting was assessed from other databases in the Departments of Health and Labour. Review of the database of the regional Department of Health found one case had reached the Provincial Office and by-passed the study. Of greater concern was the fact that, 6 months after study completion, the Provincial Head Office had no records of nine events forwarded to it as part of the surveillance project. At the same time, the Department of Labour reporting system picked up none of the six poisoning cases classifiable as directly occupational. The only instance reported to the Department of Labour (personal communication, Mr R van der Merwe, 1998) was the case involving the consumption of contaminated alcohol which had received widespread press publicity at the time of the event.
Implications for surveillance policy
Patterns of poisoning as reported in routine surveillance may seriously misinform the development of control policies. The intervention study28 suggested that routine surveillance misses about 90% of cases of pesticide poisoning. This degree of underreporting is consistent with results of other local studies that have suggested that between 5% and 20% of hospitalized cases are reported.16,22,24 Moreover, the study may underestimate real underreporting because it does not take into account milder cases that do not reach hospitals, and which are untreated or treated by primary care workers without referral. These results are similar to international findings with regard to underreporting.3,33,34
Underreporting appears worst for poisoning in occupational settings (Table 3). Presumably, suicide attempts with pesticides are more likely to achieve rapid hospitalization because of the associated distress and attention. Once hospitalized, the greater severity of illness may increase the chances of patient transfer to a tertiary service where the chances of notification may be increased. Evidence from Nicaragua suggests that the nature of the clinical presentation may well influence the likelihood of reporting of cases of pesticide poisoning.34 Hence, it is not surprising that the contribution of suicide as the major cause of pesticide poisoning may be overestimated in developing countries where surveillance generally underreports cases.
The relative significance of suicide in the incidence of pesticide poisoning is a point of international contention. Data from Asia3537 identify suicide as the main circumstance for poisoning with pesticides, while equivalent data from countries in the Americas3,38,39 point to occupational causes as the priority. There is also evidence from research in Costa Rica3 and Mexico (personal communication, Dr Luis Lujan, El Colegio de Michoacan, Mexico, 1998) that occupational poisonings with pesticides may be systematically misreported as suicide to avoid compensation claims.
In terms of preventive strategies, the implications of these differences in pattern are substantial. If suicide is the major cause of poisoning with pesticides, industry arguments for safe use can place the responsibility for poisoning on the individual. Notions of ignorance as commonly expressed by public health policy documents in South Africa, effectively relieve industry of accountability for the impacts of widespread pesticide use. Critics have argued that industry safe use campaigns have paradoxically increased health hazards for residents in developing countries by successfully promoting greater market penetration by pesticides.4,39
The use of such categories as negligence, ignorance and carelessness in surveillance nomenclature appear to reflect more the health system's attitude to patients (often blaming victims) than providing any helpful pointers to preventive strategies needed. While education is clearly a critical strategy required for prevention, as an isolated solution it does not address the full range of contributory factors, both social and behavioural, that result in pesticide-related morbidity and mortality.
Moreover, suicide with pesticides in the context of agricultural work may be considered an occupational event.40 For example, Schilling and Anderson41 have framed occupational disease in terms of a spectrum of work-relatedness, including conditions for which the workplace is responsible for increased access to workplace hazards. In this formulation, suicide amongst farm workers using pesticides would qualify clearly as an occupational disease.
The other critical policy implication to emerge from the data is that, rather than being a situation where two-thirds of pesticide poisoning cases involve men, the reverse may possibly hold true (Table 3). The impression that women are at lower risk of exposure, as is typically claimed in many areas of occupational epidemiology,42 may be erroneous. The involvement of women in piece-work43,44 and as seasonal labourers,45 and the unfavourable conditions of employment associated with such work,43,44 (e.g. less training and protective equipment) increases their risk of exposure. Increasing concern about an association between pesticide exposure and impacts on reproductive health not only of the mother4649 but in transgenerational effects on the offspring50,51 adds to the significance of pesticide poisoning in women. The cumulative impacts of long-term low-dose exposure are liable to even greater underestimation4 than is evident from this review.
Moreover, in settings where developing countries believe that their agricultural sector is critical to national development, economic objectives may place greater priority on expansion of agricultural production than on the social and health costs of pesticide usage.7 Policies relating to pesticides usually externalise these costs when weighing up the different control options.9,52,53 The absence of a coherent and accurate surveillance system for monitoring adverse health consequences of uncontrolled pesticide usage may therefore contribute to this externalization of cost, and facilitate policies that are both hazardous to human health as well as unsustainable in an ecological framework.
Inspectorates have tended to develop a culture in which the process of investigation and report writing becomes an end in itself. This may undermine attention to intervention in public health practice where surveillance is used to inform risk reduction strategies. The inadequacy of this focus on process rather than outcome is illustrated in cases 10 and 11 in this study, where one farm in the study had two notifications 6 weeks apart despite an inspection, recommendation and completion (and submission) of a toxicology report after the first event. A genuinely effective surveillance system should be able to support a culture of intervention rather than a culture of report writing.
Conclusion
This review shows that, on a number of criteria,13,14 surveillance for pesticide poisoning in South Africa is highly ineffective and may contribute to misinformed policy choices regarding pesticide use. These problems are fairly typical of surveillance in developing countries.4,6,7
Data collected for surveillance need to be pertinent and action-driven.54 Categories used in South Africa such as ignorance, carelessness and negligence are of little use for public health control. Surveillance data should provide a better understanding of the circumstances in which poisoning events occur to inform appropriate risk reduction strategies. For example, evidence in the intervention study that spray drift may be responsible for mass poisonings should be followed up by the introduction of specific legislation aimed at addressing re-entry exposures by restricting worker access to sprayed areas. That these regulations need to be formulated and implemented by a department (Labour) different to that which identifies the problem through surveillance (Health), illustrates the critical value of inter-sector collaboration in the Primary Health Care approach needed for pesticide safety.
The plethora of reporting systems across different government departments with little co-ordination or integration (Figure 1) appears to serve as an obstacle rather than facilitating effective surveillance. Departmental co-operation at local level is critically needed for the control of pesticide hazards, rather than reliance solely on a national non-statutory co-ordinating committee.9 Rationalization or, at the very least, sharing of inspection and monitoring functions across inspectorates can make a major contribution to improving surveillance and developing healthy and sustainable public policy with regard to pesticides in countries where the adverse impacts of pesticides are greatest.
Acknowledgments
The intervention study was supported by a grant from the Department of Health: Primary Health Care Research Fund. Financial support for the policy research was provided by the Danish Co-operation for Environment and Development (DANCED) and by the South African Medical Research Council. The authors wish to thank the staff of the Bree Rivier Regional Services Council, Mrs Costa and Mrs Airie from the Department of Health, Western Cape, the laboratories of the South African Institute of Medical Research and Dr's Penman and partners, and the medical practitioners and other health professionals of Worcester and surrounds for their co-operation and support.
References
1 Xue SZ. Health effects of pesticides: a review of epidemiologic research from the perspective of developing nations. Am J Ind Med 1987;12:26970.[ISI][Medline]
2 Jeyaratnam J. Acute pesticide poisoning: a major global health problem. World Health Stat Q 1990;43:12944.
3 Wesseling C, Castillo L, Elinder CG. Pesticide poisonings in Costa Rica. Scand J Work Environ Health 1993;19:22735.[ISI][Medline]
4 Wesseling C, McConnell R, Partanen T, Hogstedt C. Agricultural pesticide use in developing countries: health effects and research needs. Int J Health Serv 1997;27:273308.[ISI][Medline]
5 Ohaya-Mitoko G. Occupational Pesticide Exposure Among Kenyan Agricultural Workers. An Epidemiological and Public Health Perspective. PhD thesis, Wageningen University, Holland, 1997.
6 Forget G. Pesticides and the Third World. J Toxicol Environ Health 1991;32:1121.[ISI][Medline]
7 Dinham B. The Pesticide Hazard. A Global Health and Environmental Audit. Pesticides Trust, London: Zed Press, 1993.
8 London L. An overview of agrichemical hazards in the South African farming sector. S Afr Med J 1992;81:56064.[ISI][Medline]
9 Rother A, London L. Pesticide Health and Safety Policy Mechanisms in South Africa: The State of the Debate. Occupational and Environmental Health Research Unit. Working Paper No. 1. University of Cape Town, South Africa, 1998.
10 African National Congress. The Reconstruction and Development Programme. A Policy Framework. Johannesburg: Umanyano Publications, 1994.
11 African National Congress. Agricultural Policy. Johannesburg: Westro Reproductions, 1994.
12 Agricultural News. Cotton farming for small-scale or emergent farmers. Agric News 1997;13:11.
13 Centers for Disease Control. Guidelines for evaluating surveillance systems. MMWR 1988;37(Suppl.S-5):[inclusive page numbers].
14 Campbell BB. Health Management Information Systems in Lower Income Countries. An Analysis of System Design, Implementation and Utilisation in Ghana and Nepal. PhD thesis, University of Amsterdam, Amsterdam, 1997.
15 Department of National Health and Population Development. Health Act. Act 63/77. Pretoria: Government Printers, 1977.
16 Department of Health. Pesticidal poisoning in South Africa, 1980 1994. Epidemiol Comm 1995;22:11238.
17 Department of National Health and Population Development. Improved notification can help prevent pesticide poisoning. SALUS 1992;15(4):1011.
18 Van der Merwe PJ, Hundt HKL, Bekker M, van der Merwe JC. Epidemiologiese studie van vergiftigings in kinders in Bloemfontein en omgeweing. (Epidemiological study of poisoning in Bloemfontein and surrounding areas.) S Afr Med J 1988;74:22022 (In Afrikaans).[ISI][Medline]
19 Roberts JC, Leary PM, Mann MD, Glasstone M. The pattern of childhood poisoning in the Western Cape. S Afr Med J 1990;78:2224.[ISI][Medline]
20 Van der Merwe PJ, Botha JJD. Epidemiologiese studie van vergiftigings in kinders onder 18 jaar in Bloemfontein en omgeweing. (Epidemiological study of poisoning in children under 18 years in Bloemfontein and surrounding areas.) S Afr Med J 1991;79:25355 (In Afrikaans).[ISI][Medline]
21 Muller GJ, Hoffman BA, Lambrecht JH. Drug and poison information the Tygerberg experience. S Afr Med J 1993;83:39599.[ISI][Medline]
22 London L, Ehrlich R, Rafudien S, Krige F, Vurgarellis P. Notification of pesticide poisoning in the Western Cape 19871991. S Afr Med J 1994; 84:26972.[ISI][Medline]
23 Schierhout GH, Midgley A, Myers JE. Occupational fatality under-reporting in rural areas of the Western Cape Province, South Africa. Safety Sci 1997;25:11322.[ISI]
24 Emanuel K. Poisoned Pay: Farmworkers and the South African Pesticide Industry. Johannesburg: Group for Environmental Monitoring and The Pesticide Trust, 1992.
25 Coetzee GJ. The Epidemiology of Pesticide Mortality in the Western Cape. Department of Community Health, Unpublished thesis. University of Cape Town, 1981.
26 Bradshaw D, Mbobo L. Informatics support. In: South African Health Rev, Ch. 9, pp.12330. Durban: Health Systems Trust, 1995.
27 Braa J, Heywood A, Shung King M. District Level Information Systems: Two cases from South Africa. Meth Inform Med 1997;36:11521.[ISI][Medline]
28 Bailie R, London L. An elusive target: enhanced surveillance for pesticide poisoning in the Western Cape. S Afr Med J 1998;88: 110509.[ISI][Medline]
29 London L, Bailie R. Knowledge, attitudes and practices of doctors in the rural Western Cape with regard to notification of pesticide poisoning. S Afr Fam Pr 1999;20:11720.
30 Coye MJ, Fenske R. Agricultural workers. In: Levy BS (ed.). Occupational Health. Recognising and Preventing Work-related Disease. Boston: Little, Brown and Co., 1988.
31 London L. The dop system, alcohol abuse and social control amongst farm workers in South Africa: a public health challenge. Soc Sci Med 1999;48:140714.[ISI][Medline]
32 Scully P. Liquor and labour in the Western Cape, 18701900. In: Crush J, Ambler C (eds). Liquor and Labour in Southern Africa. Athens: Ohio University Press, 1992, Ch. 2, pp.5677.
33 Mbakaya CFL, Ohaya-Mitoko GJA, Ngowi VAF et al. The status of pesticide usage in East Africa. Afr J Health Sci 1994;1:3741.[Medline]
34 Keifer M, McConnell R, Pacheco AF, Daniel W, Rosenstock L. Estimated underreported pesticide poisonings in Nicaragua. Am J Ind Med 1996;30:195201.[ISI][Medline]
35 Jeyaratnam J, De Alwis Senewiratne RS, Copplestone JF. Survey of pesticide poisoning in Sri Lanka. Bull World Health Organ 1982;60: 61519.[ISI][Medline]
36 Jeyaratnam J, Lun KC, Phoon WO. Survey of acute pesticide poisoning among agricultural workers in four Asian countries. Bull World Health Organ 1987;65:52127.[ISI][Medline]
37
Malik GM, Mubarik M, Romshoo GJ. Organophosphorus poisoning in the Kashmir Valley, 1994 to 1997 (letter). N Engl J Med 1998;338: 1078.
38 Cole D, McConnell R, Murray DL, Anton FP. Pesticide illness surveillance: the Nicaraguan experience. Pan Am Health Organ Bull 1988;22:119312.
39 Murray DL. Cultivating Crisis: The Human Cost of Pesticides in Latin America. Austin: University of Texas Press, 1994.
40 Parron T, Hernandex AF, Villenueva E. Increased risk of suicide with exposure to pesticides in an intensive agricultural area. A 12-year retrospective study. Forensic Sci Int 1996;79:5363.[ISI][Medline]
41 Schilling R, Andersson N. Occupational epidemiology in developing countries. J Occup Health SafetyAust NZ 1986;2:46878.
42 Messing K. Women's occupational health: a critical review and discussion of current issues. Women Health 1997;25(4):3968.[ISI][Medline]
43 Waldman PL. Here You Will Remain: Adolescent Experience on Farms in the Western Cape. MA Thesis, Department of Anthropology, University of Cape Town, Cape Town, South Africa, 1993.
44 Hill-Lanz S. Women on Farms. A Report on Women Farm Workers in the Western Cape. Pretoria: Lawyers for Human Rights, 1994.
45 Central Statistical Services. Statistical News Release P1101. Agricultural Survey. Pretoria: Central Statistical Services, 1991.
46 Restrepo M, Munoz N, Day NE, Parra JE, Romero L, Nguyen-Dinh X. Prevalence of adverse reproductive outcomes in a population occupationally exposed to pesticides in Columbia. Scand J Work Environ Health 1990;16:23238.[ISI][Medline]
47 Maroni M, Fait A. Health effects in man from long-term exposure to pesticides. A review of the 19751991 literature. Toxicology 1993;78: 1174.[Medline]
48 Sever LE, Arbuckle TE, Sweeny A. Reproductive and developmental effects of occupational pesticide exposure: the epidemiological evidence. Occup Med 1997;12:30525.[ISI][Medline]
49 Garcia A. Occupational exposure to pesticides and congenital malformations: a review of mechanisms, methods and results. Am J Ind Med 1998;33:23240.[ISI][Medline]
50 Colborn T, Myers JP, Dumanowski D. Our Stolen Future. London: Abacus Books, Little, Brown and Company, 1996.
51 Gray LE, Ostby J. Effects of pesticides and toxic substances on behavioural and morphological reproductive development: endocrine versus nonendocrine mechanisms. Toxicol Ind Health 1998;14: 15984.[ISI][Medline]
52 Pimental D et al. Environmental and economic costs of pesticide use. Bioscience 1992;42:75060.[ISI]
53 Agne S, Fleischer G, Jungbluth F, Waibel H. Guidelines for Pesticide Policy Studies. A Framework for Analysing Economic and Political Factors of Pesticide Use in Developing Countries. Publication Series No. 1. Pesticide Policy Project, Hanover, Germany, 1995.
54 Sandiford P. What can information systems do for Primary Health Care? An international perspective. Soc Sci Med 1992;34:107787.[ISI][Medline]