Summary of Meeting Paper

The 1996 Annual Meeting of the Society for Risk Analysis-Europe

The Cost-Effectiveness of Lifesaving Interventions in Sweden. A. L. Joakim, J. Ramsberg and Lennart Sjöberg, Center for Risk Research, Stockholm School of Economics, Sweden

INTRODUCTION

What is the value of a human life? One way of answering this question is simply to state that all lives have an equal and infinite value. In a philosophical sense that may be the correct answer and for an individual it is probably true. Unfortunately, this answer does not give much guidance when dealing with the practical problems of resource allocation, since resources are always limited. Another way of approaching the question is to use implied life values. Implied life value is defined as the estimated cost of statistical lives saved. It measures the cost-effectiveness of interventions aimed at reducing the probability of premature death.

METHOD

We have searched for publicly available economic analyses of lifesaving interventions in Sweden in scientific literature, government reports and reports from non-government organizations. The criteria for including an analysis were that it should concern lifesaving interventions in Sweden and report cost per life saved, cost per life-year saved or sufficient information to calculate one of these measures. Our definition of a lifesaving intervention resembles that of Tengs et al. [4]; A lifesaving intervention is defined as any measure directed towards changing the behavior and/or technology of individuals or organizations, where reducing the probability of premature death in a population is a primary goal or an actual effect of the intervention. To improve the comparability of the interventions, certain criteria have been used to standardize the estimates.

RESULTS

Implied life values, together with references, are presented for 157 lifesaving interventions in Sweden. Several of these interventions produce net savings for society. The most expensive interventions cost several hundred million SEK, or more, per life saved. This large variation in cost-effectiveness can be seen in figure 1. The number of interventions we describe is 157. The mean cost per life saved is SEK 259.5 million and the median is SEK 4.7 million. However, in the data there is one extreme outlier of SEK 35,000 million and if that observation is dropped, the mean cost per life saved is SEK 36.8 million. The variation in cost per life saved is large, which can be seen in Figure 1.

Figure 1. Distribution of cost per life saved.

Implied life values have been collected or calculated for nine different categories of lifesaving: medicine, radiation protection, road safety, life style risks, electrical safety, accidents, pollutants in the environment and crime. These categories are admittedly somewhat ad hoc. Medicine includes by far the largest number of observations, whereas crime and accidents contain only one observation each. The number of observations and mean and median cost per life saved for the different categories are presented in Table I.

Table I: Number of observations, mean and median for the nine categories.
All costs in million 1994 SEK.

EXPLAINING THE VARIATION

Perhaps more interesting than to just show the variation would be to explain it. An important question is then whether the large variation in society's investments in life-saving interventions reflects the public opinion or not. There is some empirical evidence that this is not the case; The public seems to support a variation with a factor two or three, but not more [2], [1].

Once the public opinion on the matter of investments in lifesaving has been established, it is crucial to investigate the mechanisms by which public opinion influence decisions. Sjöberg found in a study of politicians that the risk reduction they desired was more strongly accounted for by their beliefs about the public's risk perception than by their own risk perception [3]. This finding strongly supports the idea that it is important to study public opinion on investments in lifesaving interventions.

A plausible hypothesis about investment in lifesaving interventions is that the size of the investment in absolute numbers can influence the decision to invest. The larger the investment is, the more cost-effective it has to be. We have performed a preliminary analysis of this hypothesis on our data, but could find no support for it. That is however not a reason to completely reject it; a more thorough study is needed.

LIMITATIONS

It has been our ambition to make the figures concerning the cost-effectiveness of the interventions as comparable to one another as possible. The costs and effects of the different interventions were originally calculated by different authors, in many cases using different methods. With this compilation, we have enhanced the comparability of the interventions, but the reader should be aware that large dissimilarities in the calculation of implied life values still remain. For instance, for some interventions it is the marginal cost per life saved that is presented, and for others it is the average cost per life saved. The original data on which the original analyses were based of course put a limit on the accuracy of our results as well.

An important qualification is that the level of implementation varies from fully implemented interventions to partially implemented interventions, to interventions that have only been proposed or proposed and rejected.

Both the somewhat uncertain quality of the data and the various levels of implementation restrain the possibility to draw conclusions about investment in life saving interventions in Sweden. Furthermore, it is unlikely that these interventions constitute a representative sample of all life-saving interventions, yet another reason to be cautious when drawing conclusions from these results. Despite these reservations, it is nevertheless possible to conclude that implied life values vary greatly both within and between different sectors of the Swedish society, much in the same manner as in current US analyses [4].

REFERENCES

1]. Cropper, M. L., & Subramanian, U. (1994). Public choices between life-saving programs: How important are qualitative factors versus lives saved?. Unpublished report. The World Bank, Washington, D.C.

2]. Mendeloff, J. M., & Kaplan, R. M. (1989). Are large differences in "lifesaving" costs justified? A psychometric study of the relative value placed on preventing death. Risk Analysis, 9, 349-363.

3]. Sjöberg, L. (In press). Risk perception by politicians and the public. Stockholm: Center for Risk Research, Stockholm School of Economics.

4]. Tengs, T., & et al. (1995). Five-hundred life-saving interventions and their cost-effectiveness. Risk Analysis, 15, 369-391.