Low-Level Environmental Radiation and U.S. Mortality
An approach to an association between environmental radiation and mortality rates is investigated for 61 U.S. cities for the years 1961 to 1967. Environmental radiation is measured by weapons fallout (90Sr, 131I, 137Cs , 140Ba , and 89sr) and by cosmic radiation. Other factors known to affect mortality rates are entered in the analysis, including air pollution and socioeconomic factors. Multiple regression techniques . are used to relate each mortality rate (total, infants under one year, infants under 28 days, and fetuses) to the radiation, air pollution, and socioeconomic variables. A series of specifications is developed in an attempt to account for important missing variables and other specification errors. In spite of·these specification errors, poor data, and theoretical 90SR and 137Cs . difficulties, the levels of Sr and Cs in milk are found to be associated with the mortality rates in a consistent, stable, and important fashion. Although these results do not prove that environmental radiation causes increased mortality they do indicate that it would be unreasonable to rule out the existence of such a relationship. The paper concludes with a call for the collection of more consistent, relevant, and comprehensive data to enable a better understanding and more precise estimation of this association.