Statement of John D. Graham, Professor of Policy and Decision Sciences, Harvard School of Public Health,
Before the National Transportation Safety Board, Effectiveness Panel, on March 19, 1997

Posted March 25, 1997.

Mr. Chairman, thank you for the opportunity to share with the safety community the methods, preliminary results, and limitations of our current study of the cost-effectiveness of driver-side and passenger-side airbags. I would like to begin by acknowledging the work of my co-authors, Kim Thompson, Maria Segui-Gomez, Sue Goldie, and Milton Weinstein. We plan to publish the study in a peer-reviewed scientific journal in the months ahead and thus we would appreciate any technical comments and criticisms by May 1st.

The study analyzes separately the cost-effectiveness of the driver-side and passenger-side air bag systems using a common methodology. The details of the methodology are described in the recent report of the Panel on Cost-Effectiveness in Health and Medicine, a group commissioned by the Department of Health and Human Services.

The analysis is conducted from the societal perspective, which means that all resource costs and safety consequences are considered, regardless of who in society incurs them. This perspective obviously differs from a particular corporation's perspective. We simulate the experience of a hypothetical cohort of 10 million new vehicles that are equipped with manual belts alone, driver-side airbags, or dual-front airbags.

Effectiveness is defined as the number of years of life saved by each technology, including an adjustment for quality of life due to changes in the frequency of nonfatal injuries. In the jargon of the field, this measure of effectiveness is called quality-adjusted life years (QALYs). We estimate net effectiveness, taking into account the fatalities and injuries prevented by airbags as well as the fatalities and injuries caused by airbags.

Costs include the resources involved in producing, maintaining, and replacing airbags as well as the resources associated with treating and rehabilitating crash-related injuries. The productivity losses from trauma are excluded in order to avoid double counting the benefits already captured in the effectiveness measure.

Data for the analysis were obtained from published reports by the National Highway Traffic Safety Administration, the Insurance Institute for Highway Safety, State Farm Insurance Company, the American Automobile Manufacturers Association, and peer-reviewed journal publications. A reference-case analysis uses base-case values for each input to the analysis, with sensitivity analysis conducted to define the nature and magnitude of the uncertainties.

The key summary statistic in a cost-effectiveness analysis is the cost-effectiveness ratio of one technology compared to a well-defined comparator. I report here the preliminary results of the base-case analysis. For manual safety belts, we estimate that the technology is cost-saving compared to no restraints, which means that the costs of the safety belts are more than offset by the savings in reduced hospitalization and rehabilitation expenses. For driver-side airbags, the net cost was about $70,000 per quality-adjusted life year saved compared to manual safety belts. For dual-front airbags, the net cost was $399,000 per QALY saved compared to driver-side airbags.

The sensitivity analysis shows that these results are sensitive to a variety of uncertain factors: the marginal cost of airbag systems, their effectiveness, baseline fatality rates, and the real discount rate. For example, if the effectiveness rate of airbags for persons 10 to 65 years of age were 17%, instead of 11%, the cost-effectiveness ratio improves from $70,000 to $43,000 per QALY saved with the driver-side bag, and from $399,000 to $127,000 for the passenger-side bag. However, the relative ranking of the drivers-side and passenger-side bag is not sensitive to these uncertainties.

It may seem surprising that the passenger-side airbag is less cost-effective. This difference is not because we have assumed that the passenger-side airbag is highly expensive. Indeed, we assign all of the sensor and monitoring costs to the driver-side bag, leaving only the incremental cost of the passenger side bag itself (estimated by NHTSA at $132 per vehicle). Nor did we assume that the passenger bag is less effective for adults; we used the same adult effectiveness rate for driver and passenger bags.

The passenger-side airbag was estimated to be less cost-effective for three reasons: (1) the number of drivers at risk of fatuity is about 2.6 times larger on the driver side because there are simply more drivers than passengers; (2) for this same reason, the number of drivers at risk of serious injury is also assumed to be 2.6 times larger on the passenger side; (3) the effectiveness of the passenger-side bag is assumed to be negative for children, meaning that passenger bags kill more children than they save. The children have an important effect on the analysis because they, on average, lose 75 years of life expectancy; the driver loses, on average, about 35 years of life expectancy.

Our tentative conclusion is that driver-side airbags save life years at a cost that is comparable to many medical and public health practices. Passenger-side airbags appear to save life years, on balance, but at a substantially higher cost per quality-adjusted life year.

These results should be interpreted with caution, for two reasons. First, the estimates of airbag effectiveness and risks to children are uncertain, and new data could alter the conclusions. Second, the cost-effectiveness of passenger-side bags could be improved considerably, from $399,000 per QALY saved to $104,000 per QALY, if all children were required to sit in the rear seat, as they are today in France and Germany. Moreover, new airbag designs may also prove to be more cost-effective than current designs, particularly if they can reduce the dangers to children.

Thank you very much for the opportunity to speak today, and I look forward to the questions and comments.

Links to Related Documents

RiskWorld news article titled "Harvard’s John Graham Releases Results of Cost-Benefit Analysis of Air Bag Safety"

Full text of John Graham’s March 17 testimony before the National Transportation Safety Board’s Supplemental Restraint Panel

Go to:

  • News department main menu

  • RiskWorld departments

  • Tec-Com Inc.

  • Copyright 1997 by Tec-Com Inc.