The avoidable costs of alcohol abuse in Australia and the potential benefits of effective policies to reduce the social costs of alcohol
Appendix A. Methodological issues
This appendix provides a more detailed consideration of the methodological issues involved in the estimation of the avoidable costs of substance abuse than is presented in the main body of this report. The exposition is largely based upon the International Guidelines for the Estimation of the Avoidable Costs of Substance Abuse (Collins et al., 2006).
A1. The nature of avoidable costsAs explained in the original International guidelines for estimating the costs of substance abuse published by the World Health Organization (Single et al., 2003), estimates of the costs of substance abuse constitute just one component in a range of potential economic information on substance abuse. The following table from this WHO publication presents a summary of the hierarchy of potential economic information on the costs of substance abuse, together with examples of their possible uses. This table puts avoidable costs in the context of the range of possible types of abuse cost studies.
Table 30. Substance abuse cost estimates and their policy uses
|Type of estimate||Interpretation of results||Example of policy use|
|Aggregate costs||Total external costs of substance abuse compared with the alternative situation of no substance abuse||Indication of the size of the substance abuse problem|
|Avoidable costs||Potential economic benefits from substance abuse harm-minimisation strategies||Determination of the appropriate level of resources to be devoted to harm-minimisation strategies|
|Costs incidence||The distribution of the external costs of substance abuse among various community groupings||Mobilisation of support from various groups (for example, the business community) for anti-substance abuse programs|
|Disaggregated costs||External costs of substance abuse disaggregated by categories||Economic evaluation (cost-benefit or cost-effectiveness analysis) of harm-minimisation programs|
|Budgetary impact||The impact of substance abuse on government expenditures and revenues||Assessment of the case for industry compensation payments to government as a result of abusive use of substances which the industry produces|
Top of Page
In estimating the aggregate costs of substance abuse, the actual substance abuse situation must be compared with some alternative hypothetical situation, known as the counterfactual situation. This counterfactual situation is usually a situation in which there has been no past or present abuse of the substance(s) in question. Thus the counterfactual situation is implicitly one in which the community would bear no substance abuse costs. A comparison of the actual situation with the counterfactual zero-abuse situation indicates the extra costs which abuse has imposed on the community.
However, the counterfactual situation is purely hypothetical and in almost all circumstances most unlikely to be realisable, even over long periods of time. Estimates of the aggregate costs of drug abuse comprise both avoidable and unavoidable costs. As Single et al. (2003) state:
- Unavoidable costs comprise the costs which are currently borne relating to drug abuse in the past, together with the costs incurred by the proportion of the population whose level of drug consumption will continue to involve costs. Avoidable costs are those costs which are amenable to public policy initiatives and behavioural changes.
Avoidable cost estimates indicate the benefits potentially available to the community as a result of directing public resources to the prevention or reduction of substance abuse. They provide valuable economic information on the basis of which a more efficient allocation of productive resources could be achieved.
The estimation of social costs involves, in principle, first the estimation of the relevant avoidable proportion of each of the cost categories and then applying these proportions to each of the relevant aggregate cost estimates. Since not all substance abuse costs fall on the government sector, governments are likely to show an interest in the proportions of budgetary costs which are avoidable. Information developed for estimating the avoidable proportions of aggregate costs can also be readily applied to the estimation of avoidable budgetary costs.
A2. Reasons for estimating avoidable costsThe Health Canada avoidable cost guidelines (Collins et al., 2006, section 1.3) identify the following four major reasons for estimating the avoidable costs of substance abuse.
Top of Page
Priority for substance abuse expenditures
- In most countries, the allocation of public funds between competing programs is substantially influenced by pubic servants trained in economics or finance, who are looking to maximize the social rates of return on public expenditures. Accordingly, as part of the decision-making process, they utilize information on economic evaluation of the proposed expenditures. While aggregate cost estimates indicate the economic impact of substance abuse, they do not indicate what proportions of these costs are avoidable, and over what period of time. It is possible that some forms of substance abuse are much more susceptible to prevention measures than others, and so may yield higher gross benefits. Thus, avoidable cost estimates give a better indication of the potential benefits of anti-abuse programs and policies, though they still do not indicate the costs, and the rates of return, of appropriate programs.
Appropriate targeting of specific problems
- Estimates of the avoidable costs of substance abuse, in conjunction with estimates of the aggregate costs, provide an extremely valuable information resource for policy analysis and design. This information is of two types:
- Cost estimation data. In the process of developing cost estimates, researchers generate virtually all the information necessary to value the benefits (that is, the reduction in social costs) of programs, once their physical outputs (in terms of, say, improved health outcomes, lower work absenteeism and reduced crime rates) have been determined.
- Information about the physical outcomes potentially available to specific programs. Essentially, the estimation of avoidable costs involves an examination of the extent to which it is possible to reduce substance abuse and, in the process, what policy measures should be implemented in order to achieve this reduction. As will be seen, this effectively involves an examination of best practice in a range of comparable jurisdictions. Thus the development of avoidable cost estimates will involve an international survey of substance abuse policies and outcomes, and the provision of the resulting information on a systematic basis.
Accordingly, avoidable cost estimation will facilitate both program analysis (through cost-benefit analysis) and design of appropriate programs likely to yield the best achievable social rates of return.
Identification of information gaps and research needs
- systematic analysis of avoidable costs is likely to highlight information gaps which when rectified, will facilitate improved policy design. This information will include answering such questions as:
- What substance abuse outcomes are potentially achievable?
- How best may these outcomes be achieved?
- Over what periods of time would these outcomes be achievable?
Since this information, in terms of both scientific analysis and policy analysis, will inevitably not be static, it will be necessary to review and update information concerning potential best practice on a regular basis.
Provision of baseline measures to determine the efficiency of drug policies and programs
- Once policy makers have at their disposal measures of best practice substance abuse outcomes, they will be able much more efficiently to evaluate their existing anti-abuse policies and their allocation of public expenditures to and between anti-abuse programs. There is every reason to believe that the provision of avoidable cost estimates should lead to better policy design.
A3. Approaches to the estimation of avoidable costsCollins et al. (2006) provide an extensive discussion of possible approaches to the estimation of the avoidable social costs of substance abuse. The following exposition constitutes a brief summary of the discussion in that publication.
It is important, from a public policy perspective, to be able to identify the proportion of the estimated aggregate social costs of substance abuse which is potentially susceptible to reduction, and the period over which such reduction could be achieved. This information will yield estimates of the value of avoidable costs and, accordingly, of the potential benefits of policies to reduce these costs. These types of estimates for tobacco have been made by the present authors for various Australian states for targeted reductions in the rates of smoking prevalence (see Collins & Lapsley, 1999, 2001, 2004, 2005, 2006 and 2008c). The next stage would be to identify the nature of the programs necessary to achieve these reductions and the costs of these programs. It then becomes possible to estimate the social rate of return accruing to programs designed to achieve the elimination of avoidable costs.
Top of Page
This complete process can be summarised as consisting of the following stages:
- estimate the aggregate social costs of the substance abuse under study;
- estimate the proportion of these aggregate costs which are avoidable;
- estimate the time period over which the potential reduction in aggregate costs could be achieved;
- determine the nature of the programs which would achieve this reduction and the costs of these programs; and
- calculate the social rate of return which would result from the elimination of the avoidable costs.
The avoidable cost concept explored here has been contrasted by Collins et al. (2006) with the economic concept of the optimal level of substance abuse.
- Economists argue that the optimal level of drug consumption is reached when the incremental cost to the community as a whole of achieving a given reduction in consumption is exactly matched by the incremental benefit to the community of that reduction. If the incremental benefit is greater than the incremental cost, achieving optimality would require a further reduction in consumption. If the cost exceeds the benefit, then consumption has been reduced to sub-optimal levels.
The concepts of avoidability and optimality can lead to quite different outcomes. It is perfectly possible that optimal levels of consumption may not be achievable. For example, optimal levels of tobacco consumption may be well below the levels which are achievable in a real world in which severe constraints exist on the public resources available to achieve reductions in smoking prevalence. On the other hand, it may be technically possible to reduce tobacco consumption below levels which economists judge to be optimal.
The estimation of the avoidable costs of substance abuse involves determination of the maximum reduction in substance abuse costs which effective policies can be expected to achieve. The lowest achievable level of substance abuse is termed the Feasible Minimum. The Health Canada avoidable cost guidelines identify four possible approaches to estimation of the Feasible Minimum. Which approach should be adopted for any particular type of cost is very much a pragmatic decision, taking into account data availability and the theoretical advantages and disadvantages of the various approaches.
Top of Page
The distributional approachThe Health Canada avoidable cost guidelines identify two different approaches to disease risk-factor attribution:
- The classical epidemiological approach. In this approach disease risk is estimated with respect to exposure and non-exposure of various populations to a single specific risk factor. In effect the counterfactual scenario question asked is ‘What would have happened if no exposure had occurred?’
- The distributional approach. This not only looks at distributional shifts at one time but also considers the future time dimension, thus being able to predict future developments. This approach asks the question ‘What would happen if factor distributions shifted to different counterfactual scenarios?’ Thus this approach, which is explained fully in the guidelines, conceptualizes the avoidable burden by specifying alternative scenarios for risk-factor distributions, including potential future distributions.
The Arcadian NormalThe Health Canada avoidable cost guidelines (Section 3.6) describe this approach as follows:
- A second type of approach to estimating the Feasible Minimum is by estimating what has become known as the Arcadian Normal. Pioneering work in this area was done by Armstrong (1990). His work was expressed in terms of preventable mortality and morbidity but it is reasonable to extend the concepts embodied in his work to other costs such as the property costs resulting from drug-attributable crime or smoking-attributable fires. He talks of assuming “the existence of some level of disease that might reasonably be achieved if only we knew all that might reasonably be known about the causes of the disease in question and could apply them in practical programmes in the community. There is no simple way of identifying this level of disease but it may be assumed to be less than the lowest level of disease that obtains in some group of genetically similar populations”. Armstrong terms this level of disease the ‘Arcadian Normal’ “because it represents the nearest approach that we can make to harmony between humankind and its environment. Arcadia, in ancient Greece, was a region renowned for the contented pastoral simplicity of its people”.
Armstrong’s approach is to compare the most recently available age standardised mortality rates for a range of causes in a group of countries with genetically similar populations and with similar living standards. He takes the Arcadian Normal to be the lowest age-standardised mortality rate for each cause of death in the 20 European countries he studied and from these he estimates the proportions of potentially preventable mortality in Australia.
- The Health Canada avoidable cost guidelines consider the potential issues with the use of the Arcadian Normal in some detail. Probably the major issue is that it is based upon disease outcomes and so is not exposure-specific. Disease rates will be closely related to the rates of exposure to relevant risk factors (for example, binge drinking) but exposure may only lead to disease after significant time lags. Thus, Arcadian Normals based on current disease outcomes are, at least partly, a function of past exposure levels. If current exposure levels are declining, the use of the Arcadian Normal will almost certainly overestimate the Feasible Minimum.
Exposure-based comparatorsThe two approaches discussed above impose substantial data requirements. In addition, the Arcadian Normal is based upon current disease outcomes rather than upon current or potential risk exposure. The latter approach is the preferred approach of the avoidable cost guidelines.
A considerable amount of information on alcohol consumption is available for Australia, both nationally and state-by-state, and internationally. In principle, these data can be used to estimate Feasible Minimum exposure rates and, by extension, avoidable costs.
Top of Page
Evidence on the effectiveness of interventionsThe avoidable cost guidelines make the following points:
- In estimating avoidable costs it may at times be useful also to utilise research evidence on the effectiveness of interventions designed to reduce or alleviate the effects of substance abuse. Where such evidence exists, comparison between existing substance abuse policies and available interventions shown to be quantifiably effective may indicate the extent to which aggregate costs are avoidable.
There will be circumstances in which exposure-based measures, while relevant, are insufficient on their own to indicate avoidable proportions. Drink-driving is a case in point. While reductions in overall levels of alcohol consumption are likely to lead to a lower prevalence of drink-driving, regulatory enforcement using, for example, low maximum blood alcohol levels, extensive random breath testing and severe penalties has been shown to significantly reduce the prevalence of drink-driving. The most effective interventions often require both reduced exposure and effective enforcement of a regulatory environment.
Which method to adopt?Of the four methods discussed above, the issue of which method(s) to adopt in relation to a particular category of attributable cost is very much a pragmatic issue. It depends upon data availability and upon information on the effectiveness of relevant public policies.
A4. Time lags associated with policies to reduce substance abuseEstimation of avoidable costs implies the calculation of a Feasible Minimum below which real costs cannot be reduced. However, this minimum may only be attainable after an extended period of time—that is, with an extended lead time. The lead times involved in reaching the Feasible Minimum are of three types:
- policy implementation lead times, since policies cannot be instantly designed, legislated and implemented;
- lead times between the reduction in exposure to risk and the consequent reduction in the adverse consequences of exposure; and
- lead times between the reduction in exposure to risk and an adjustment in the population structure which fully reflects the reduction in premature mortality.
The present study attempts, within the constraints imposed by the available data, to estimate the value of the potential reduction in the social costs of alcohol abuse in Australia (that is, the avoidable costs of alcohol). The estimates presented are expressed in 2004/05 dollars. This should not, however, be taken to imply that, had a set of appropriate policies been implemented prior to that year, all these benefits would actually have accrued in 2004/05. Some benefits (relating to acute problems such as drink-driving and alcohol-attributable violence) would accrue very quickly. Other benefits (relating to chronic conditions such as pancreatitis and liver cirrhosis) would be subject to substantial accrual lags. It has unfortunately proved difficult to estimate these lags (see Rehm et al., forthcoming, for a review of some of the relevant literature). Our estimates are expressed in 2004/05 dollars but these benefits could not have fully accrued in 2004/05.
Top of Page
The issue of the actual time profile of benefits becomes very important once attempts are made to undertake cost-benefit analysis of particular polices. In cost-benefit analysis, the future time streams of both policy benefits and implementation costs are, for the purposes of comparison, converted into dollar figures for a common single year, by use of an appropriate social discount rate. Cost-effectiveness analysis simplifies the process by comparing the discounted costs of policies all aimed at producing the same objective or output (for example, a given reduction in accidents resulting from drink-driving). However, this simplification inevitably results in a reduced usefulness for the research results, since only policies with the same output can be compared.