The losses to society from work-related accidents and illnesses are very large, but no society can afford to prevent all these losses. Because of the scarcity of resources, limited investments have to be targeted carefully to give the “biggest bang for the buck”. The mere costing of occupational ill-health does not facilitate the targeting of investment. Proper economic evaluation can be helpful if it is well designed and executed. The results of such evaluation can be used, with appropriate critical appraisal of the practice of evaluations, to inform investment choices. Economic evaluation will not and should not determine investment decisions. Such decisions will be the product of economic, political and social values. As Fuchs (1974) argued:

At the root of most of our major health problems are value choices. What kind of people are we? What kind of life do we want to lead? What kind of society do we want to build for our children and grandchildren? How much weight do we want to put on individual freedom? How much to equality? How much to material progress? How much to the realm of the spirit? How important is our own health to us? How important is our neighbour’s health to us? The answers we give to these questions, as well as the guidance we get from economics, will and should shape health care policy.

A decision to regulate the mining industry so that fewer workers are killed and maimed will, if successful, bring health benefits to the workforce. These benefits, however, have associated costs. In real-life situations, the increased costs of improving safety will drive up prices and reduce sales in competitive world markets and may induce in employers deviance from regulations. Such deviation may be condoned by labour unions and their members, who may prefer imperfect implementation of health and safety laws if it improves incomes and employment prospects.

The purpose of economic analysis in occupational health is to facilitate the identification of that level of safety investment which is efficient. Efficiency means that the costs of doing a little more (the marginal cost) to enhance safety equal the benefits (the marginal returns in terms of health and welfare enhancements result from risk reduction). The economic aspects of occupational health and safety are central to decision making at all levels: the shop floor, the firm, the industry and the society. To behave as if all workplace risks to workers’ health can be eradicated may be inefficient. Risks should be eradicated where it is cost effective. But some risks are rare and too costly to eradicate: they have to be tolerated and when these rare events damage workers’ welfare, they have to be accepted as unfortunate but efficient. There is an optimal level of occupational risk beyond which the costs of risk reduction exceed the benefits. Investment in safety beyond this point will generate safety benefits which should be bought only if society is prepared to act inefficiently. This is a social policy decision.

Types of Economic Analysis

Cost analysis

Cost analysis involves the identification, measurement and valuation of the resource consequences of occupational accidents and ill health. Such descriptions illuminate the size of the problem but do not inform decision makers about which of many competing interventions by all who govern and regulate the workplace environment is the most efficient.

A nice example of this is a British study of the costs to economy of work accidents and work related illness (Davies and Teasdale 1994). In the year 1990 there were 1.6 million accidents reported at work, and 2.2 million people suffered ill health which was caused or exacerbated by the working environment. As a result of the events, 20,000 people were obliged to give up work and 30 million working days were lost. The loss of income and welfare to the victims and their families was estimated to be £5.2 billion. The loss to employers was between £4.4 and £9.4 billion. The loss to society as a whole was £10.9 to £16.3 billion (see table 1). The authors of the British report noted that whilst the numbers of reported accidents and industrial diseases had fallen, the estimated costs were higher.

Table 1. The costs to the British economy of work accidents and work-related health (1990 £m)

Costs to individual victims and their families	Costs to their employers	Costs to society as a whole
Loss of income	(£m)	Extra production costs	(£m)	Lost output	(£m)
InjuryIllness	376579	InjuryIllness	336230	InjuryIllness	1,3651,908
	Damage and loss in accidents	Resource costs: Damage in accidents
	InjuryNon-injuryInsurance	15-1402,152-6,499505	InjuryNon-injuryInsured	15-1402,152-6,499430
	Medical treatment
				InjuryIllness	58-24458-219
	Administration/recruitment	Administration, etc.
		InjuryIllnessNon-injury	58-6979-212307-712	InjuryIllnessNon-injury	132-143163-296382-787
Loss of welfare	Loss of welfare
InjuryIllness	1,9072,398	Employer liabilityInsurance	750	InjuryIllness	1,9072,398
Total	5,260	Total	4,432-9,453	Total	10,968-16,336
Less: compensation from employers’ liability insurance	650
Net total	4,610

The costs were higher than those reported in previous studies because of revised techniques of welfare loss estimation and better sources of information. The central information ingredient in this sort of costing exercise is the epidemiology of work-related accidents and diseases. As in all other areas of social cost analysis (e.g., alcohol—see McDonnell and Maynard 1985) the measurement of the volume of events tends to be poor. Some (how many?) accidents are not reported. The link between disease and the workplace may be obvious in some cases (e.g., asbestos-related diseases) but uncertain in other cases (e.g., heart disease and occupational risk factors). Thus it is difficult to identify the volume of work-related events.

The costing of those events which are identified is also problematic. If work stress leads to alcoholism and dismissal from employment, how are the effects of these events on the family to be valued? If an accident at work causes pain for life, how is that to be valued? Many costs can be identified, some can be measured, but often a considerable proportion of costs which are measured and even quantified, cannot be valued.

Before too much effort is expended on the costing of work-related health events, it is essential to be certain about the purpose of such work and the value of great accuracy. The costing of accidents and work-related illness does not inform the decision-making about investments in prevention of such events because it tells managers nothing about the costs and benefits of doing a little bit more or a little bit less of that prevention activity. The cost of events related to occupational ill health can identify section losses (to the individual, the family and the employer) and the costs to society. Such work does not inform prevention activity. The relevant information for such choices can be derived only from economic evaluation.

The principles of economic evaluation

There are four types of economic evaluation: cost minimization analysis, cost benefit analysis, cost effectiveness analysis and cost utility analysis. The characteristics of these approaches are outlined in table 2.

Table 2. Types of economic evaluation

	Costmeasurement	Outcome measurement: What?	Outcome measurement:How valued?
Cost minimization analysis	£	Assumed identical	None
Cost benefit analysis	£	All effects producedby the alternative	Pounds
Cost effectiveness analysis	£	Single common specific variable achieved to varying extents	Common units (e.g., life years)
Cost utility analysis	£	Effects of the competing therapies and achieved to differing levels	QALYs or DALYs

In cost minimization analysis (CMA) it is assumed that the outcome effect is identical in each of the alternatives being compared. Thus we may have two interventions to reduce the carcinogenic effects of a production process, and engineering and other data shows that the effects are identical in terms of exposure and cancer reductions. CMA can be used to cost the alternative strategies in order to identify the cheapest alternative.

Obviously the assumption of identical effects is strong and not likely to be met in most investment cases; for example, the effects of alternative safety strategies on the length and quality of workers’ lives will be unequal. In this case alternative evaluation methods have to be used.

The most ambitious of these methods is cost benefit analysis (CBA). This requires the analyst to identify, measure and value both the costs and the benefits of the alternative prevention strategies in terms of a common monetary measure. Valuing the costs of such investments can be difficult. However these problems tend to be slight compared to the monetary valuation of the benefits of such investments: how much is an injury avoided or life saved worth? As a consequence of such difficulties CBA has not been used extensively in the accident and health areas.

A more restricted form of economic evaluation, cost effectiveness analysis (CEA), has been used extensively in the health field. (CEA) was developed by the US military, whose analysts adopted the notorious measure of effect, “body count”, and sought then to identify which was the cheapest way of achieving a given enemy body count (i.e., what were the relative costs of artillery barrages, napalm bombing, infantry charges, tank advances and other “investments” in achieving a target mortality effect on the enemy).

Thus in CEA there is usually a simple, sector-specific effect measure, and the costs of achieving differing levels of reduction in, for instance, workplace events or workplace mortality can thus be computed.

The limitation of the CEA approach is that the effect measures may not be generalizable—that is, a measure used in one sector (e.g., reducing exposure to asbestos) may not be usable in another area (e.g., reducing electrical accident rates in the power distribution industry). Thus CEA may inform the decision making in a particular area but will not generate evaluative information to elucidate the costs and effects of investment choices across a wide range of prevention strategies.

Cost utility analysis (CUA) was devised to overcome this problem by using a generic effect measure, such as a quality adjusted life year (QALY) or disability adjusted life year (DALY) (see Williams 1974 and World Bank Report on Health 1993, for example). CUA techniques can be used to identify the cost/QALY effects of alternative strategies and such information can inform prevention investment strategies in a more comprehensive manner.

The use of techniques of economic evaluation in health care is well established, although their use in occupational medicine is more limited. Such techniques, given the difficulties of measuring and valuing both costs and benefits (e.g., QALYs), are useful, if not essential, in informing choices about prevention investment. It is extraordinary that they are used all too rarely and that, as a consequence, investment is determined “by guess and by God” rather than by careful measurement within an agreed analytical framework.

The Practice of Economic Evaluation

As in all other areas of scientific endeavour, there is variance between the principles of economic evaluation and its practice. Thus when using studies about the economic aspects of occupational accidents and diseases, it is essential to evaluate the evaluations with care! The criteria to judge the merit of economic evaluations have long been established (e.g., Drummond, Stoddart and Torrance 1987 and Maynard 1990). A pioneer in this work, Alan Williams, set out the following list of relevant issues over two decades ago (Williams 1974):

·         What precisely is the question which the study was trying to answer?

·         What is the question that it has actually answered?

·         What are the assumed objectives of the activity studied?

·         By what measures are these represented?

·         How are they weighted?

·         Do they enable us to tell whether the objectives are being attained?

·         What range of options was considered?

·         What other options might there have been?

·         Were they rejected, or not considered, for good reasons?

·         Would their inclusion have been likely to change the results?

·         Is anyone who has not been considered in the analysis likely to be affected?

·         If so why are they excluded?

·         Does the notion of cost go wider or deeper than the expenditure of the agency concerned?

·         If not, is it clear that these expenditures cover all the resources used and accurately represent their value if released for other uses?

·         If so, is the line drawn so as to include all potential beneficiaries and losers, and are resources costed at their value in their best alternative use?

·         Is the differential timing of the items in the streams of benefits and costs suitably taken care of (e.g., by discounting) and, if so, at what rate?

·         Where there is uncertainty, or there are known margins of error, is it made clear how sensitive the outcome is to these elements?

·         Are the results, on balance, good enough for the job at hand?

·         Has anyone else done better?

 There are several areas in economic evaluation where practice tends to be defective. For instance in the area of back pain, which causes major work-related illness losses to society, there is dispute about the competing treatments and their effects. The “old-fashioned” treatment for back pain was bed rest, but the preferred modern treatment is activity and exercise to dissipate the muscle strain which generates the pain (Klaber Moffett et al. 1995). Any economic evaluation has to build on clinical knowledge, and this is often uncertain. Thus without careful appraisal of the effectiveness knowledge base, modelling of the economic effects of alternative interventions may be biased and confusing for decision makers, as happens in the health care field (Freemantle and Maynard 1994).

High quality economic evaluations of prevention investments to reduce work-related illness and accidents are few in number. As in health care in general, the studies that are available are often of poor quality (Mason and Drummond 1995). Thus, buyer beware! Economic evaluations are essential but deficiencies in current practice are such that users of this science must be able to appraise critically the available knowledge base before committing society’s scarce resources.