Introduction

The impetus for the development of the Hazard Scoring System for Pesticides has been the need for a nationally applicable system for measuring the reduction of risks resulting from pesticide use.

There are a number of reasons for choosing a hazard scoring approach rather than a risk scoring approach:

• failure of toxicological assessment processes to accurately identify all risks, such as the effects of long term low level exposure and effects on chemically sensitive people
• lack of data on exposure
• lack of data on the effects of adjuvants, contaminants and metabolites
• lack of data on cumulative impacts and interactions
• the assumption of positive dose-response relationships when this may not always be the case.

The result is a great deal of uncertainty which is not accurately reflected in risk terminology: the term 'risk assessment' tends to convey a degree of sophistication of knowledge which does not exist. It is preferable in my view to remain as close to the data as possible and reduce the levels of uncertainty as far as possible. The problems with accuracy in risk assessment are further compounded by the gulf between lay and expert assessment because of the different frameworks within which the assessments are being made.

The Hazard Scoring System

The main distinguishing feature of this system is the manner of its proposed use - that is as a means of progress evaluation at the core of a national pesticide risk reduction policy. This means that its final form should be developed, and its implementation should be undertaken, by the decision-making group involved in the pesticide policy, rather than by experts. The decision group should consist of equal representation of public interest groups and pesticide user groups, assisted by scientific experts and government policy personnel. ¹ The Hazard Scoring System should therefore be capable of meeting the needs of these two groups.

The system is based on simplicity in order to apply to the widest possible set of circumstances, in order to remain as close to accurate datasets as possible, and in order to be understood and supported by the public.

1. Choice of parameters

Whilst the ultimate choice of parameters would need to be made by the decision group, this model proposes parameters which I believe to reasonably reflect general societal concerns.

The Hazard Scoring System does not attempt to weigh environmental and human health hazards; it retains separate scores. The prime reason for this is that the relative importance of human health and environmental health should be a societal decision, not a scientific or grower's decision. Retaining the scores separately allows the community to track progress or lack of progress in both areas. If it is found that, for example, the hazard scores for human health are consistently driving downwards but at the expense of environmental hazard, or vice versa, then this can be clearly identified and the appropriately informed policy decision made.

There is a strong tendency amongst public interest groups for a greater concern about chronic health effects than about acute effects, whereas pesticide users are perhaps more concerned with the later. Often, as the acute hazard decreases chronic hazard increases (for example, compare azinphos with chlorothalonil) ² although this is not always so (eg glyphosate). There may perhaps therefore be merit in separating the parameters for acute and chronic subchronic/toxicity, so that each pesticide has a final score in 3 parts:

acute human hazard
chronic human hazard
environmental hazard.

Chemical sensitivity

Another major area of public concern with respect to human health is that of chemical sensitivity. I have searched the medical literature on chemical sensitivity to ascertain if it is possible to incorporate a parameter for this health effect in the Hazard Scoring System - without a great deal of success. Limited findings are presented below.

Onset of sensitivity:
There are indications that the initial onset of chemical sensitivity may be linked to acute and chronic toxicity. Dr William Rea (1992) notes that all pesticides can cause sensitivity, but those which may cause the greater problem are those which are extremely toxic, bioaccumulative and lipophilic chemicals. With respect to chlorinated pesticides he notes that the degree of sensitivity appears to be a function of the extent of chlorination on the molecule.

Rea (1992) links a number of specific groups of compounds to chemical sensitivity:

• Sulfur- and phosphate-containing pesticides can damage the vigor of cytochrome P-450, the catalytic enzyme which is the mainstay of the microsomal monooxygenase system for detoxification of xenobiotics. Failure of this enzyme can result in chemical sensitivity.
• As one of the mechanisms of chemically sensitivity is immunological, pesticides that are immuno-suppressive or immuno-deregulative can be a particular problem. Rea mentions specifically a number of organochlorines, carbaryl, malathion, chlorpyrifos, and parathion.

Triggers for the chemically sensitive:
Given the lack of understanding of the mechanisms of chemical sensitivity (in particular the effects of pesticides on biochemical processes) it would be unwise to assume that there is a direct correlation between effect on the chemically sensitive and toxicity. This point is illustrated by Rea's observation that ethylene bis dithiocarbamate fungicides have low mammalian toxicity, but that they exacerbate chemical sensitivity. Sulphur compounds, of low acute and chronic toxicity, appear to enhance chemical sensitivity according to Rea.

In addition the individuality of response to a chemical by the sensitive may mean that one pesticide that has a particularly drastic effect on one person because of its effect on a specific enzyme system already weakened (eg by lack of minerals, genetic predisposition, etc), may not so adversely affect a second person, who be virtue of their biochemical individuality is more sensitive to a different type of chemical.

A parameter for chemical sensitivity:
In conclusion it seems that it is not possible to incorporate a parameter for chemical sensitivity in a hazard scoring system. However it is important to ensure that what ever system is used reflects the potential for effect on the chemically sensitive. Rea (199) claims that "all pesticides can produce toxicity and sensitivity in humans". Certainly the experience of some New Zealand medical practitioners is that the low hazard herbicide glyphosate can cause quite severe problems at low level exposure for the chemically sensitive (Davies et al 1998). For this reason a scoring system based on zero is rejected in favour of a scoring system based on 1.

Table 1: Proposed Indicators for a Hazard Scoring System

Human Health Acute	Chronic	Environmental
LD50 oral/dermal/inhalation*	Subchronic/chronic NOELs	Earthworms
Skin and eye irritation	Endocrine disruption	Beneficial insects
Sensitisation	Immunotoxicity	Bees
	Neurotoxicity	Aqu. plant/fish/invertes.
	Carcinogenicity	Birds
	Mutagenicity	Non-target invertes.
	Reproductive effects	Soil micro-organisms
Bioconcentration - Kow
		Persistence in soils - DT50
		Leaching potential - GUS
		Volatility - Henry's constant

* the highest of these values

2. Scoring

As with the selection of parameters, it is important in selecting a scoring method to be cognizant of public concerns and pesticide user needs. The method therefore needs to be one that is on the one hand scientifically accurate and justifiable, and on the other, relatively easily understood, trusted and supported. An algebraic approach is rejected primarily because the system is measuring hazard not risk.

The HSS uses a simple step function, with a range of 1-10. Extending the range from the commonly used 1 - 5 or 0 - 5 lessens the threshold effect of two very closely related values being assigned significantly differing scores.

The scores are summed to provide three separate hazard scores for each pesticide. These can be scaled to reflect the number of parameters in each category to avoid any artificial weighting. For an individual pesticide, each of the category scores is multiplied by its total use volume to provide a total hazard figure for each pesticide for the whole country, in the three separate categories of acute and chronic health hazards and environmental hazard.

3. Dealing with Data Gaps

Funtowicz and Ravetz (1992) refer to the importance of anecdotal evidence and statistics that are gathered - what they refer to as "extended facts" - in the public policy arena, an amalgamating of science and public experience in response "to the changing needs of humanity". Such an approach could be successfully used in dealing with data gaps in the Hazard Scoring System, and the following procedure is suggested:

1. Information from the 'grey' literature and anecdotal evidence is collated.
2. The literature is reviewed by the decision group, aided by independent toxicological assessment.
3. Scores are allocated on the basis of the following decision rules:
   1. where there is no evidence of adverse effect a low median (ie median of lowest 25th percentile) is applied;
   2. where there is limited evidence of effect the median is applied;
   3. where there is considerable concern and/or evidence the upper 75th percentile median is applied.

This approach has some similarities to that of Pease et al (1996) who incorporated incident data and monitoring results into their model. Such information could also be incorporated into the HSS. However the approach presented here has advantages for countries where actual data on health effects and environmental contamination is lacking.

References:
Davies, A, Bellingham M., Watts, M.A. 1998. Weed Management Policy for Auckland City. Auckland City Council.
Funtowicz, S. O., Ravetz, J.R. 1992. Three types of risk assessment and the emergence of post-normal science. In Krimsky. S., Golding, D. (eds). Social Theories of Risk. Praeger Publishers, Westport, Connecticut. pp.251-273.
Pease, W.S., Liebman, J., Landy, D., Albright, D. 1996. Pesticide Use in California: Strategies for Reducing Environmental Health Impacts. California Policy Seminar, University of California, Berkeley, California.
Rea, W.J. 1992. Chemical Sensitivity. Lewis Publishers, Boca Raton, USA.
Watts, M.A., MacFarlane, R. 1997. Reducing Reliance: a Review of Pesticide Reduction Initiatives. Pesticide Action Network Asia and the Pacific, Penang.
Watts, M.A. 1997. Proposal For A Pesticide Risk Reduction Policy For New Zealand. Proc. 50th N.Z. Plant Protection Conf. 1997:498-505.

¹ The basing of a decision group such as this on democratic principles would require the exclusion of groups with conflict of interest - such as those whose prime motive is profit, the growth of which might conflict with the reduction in risk. Back to text.

²On a sample ranking scale of 0-10, azinphos scored 7 for acute and 2 for cancer, whereas chlorothalonil scored 1 for acute and 6 for cancer. Data from Pease et al (1996) and EXTOXNET. Back to text.

---

Ms. Meriel Watts
Address:
Soil and Health Association of NZ Inc.
P.O. Box 46-076
Herne Bay
Auckland
New Zealand
Phone: +64 9 378 8244
Fax: +64 9 378 8244
E-Mail: meriel@ccu1.auckland.ac.nz

Ms. Watts has a Bachelor of Agricultural Science and a Masters in economics, environmental law and applied entomology and is currently completing her PhD in pesticide policy.

She is currently work with the Soil and Health Association of NZ (an organic growers organisation). She is an advocate for community and environmental groups on issues relating to pesticides on a number of government and industry committees including NZ's Pesticide Board (national registering authority), National Spray Drift Advisory Group, NZ Apple and Pear Marketing Board's Integrated Fruit Production Committee, Tussock Moth Science Advisory Group, etc. She also represents NZ and the South Pacific environmental and community groups on the Steering Council of Pesticide Action Network Asia and Pacific. She attended the OECD Pesticides Forum Risk Indicators Workshop in Copenhagen in 1997, as observer for New Zealand's Ministry of Agriculture.

---