Parallel analyses of individual and ecologic data on residential radon, cofactors, and lung cancer in Sweden

Divergent results from ecologic and analytic studies on residential radon and lung cancer have created uncertainty in terms of risk assessment. The authors performed concurrent analyses on individual and aggregated data from the nationwide case-control study of residential radon and lung cancer in Sweden. For data aggregated on the county level, the ecologic excess relative risk estimates per 100 Bqm(-3) residential radon concentration ranged from -0.03 (95% confidence interval (Cl) -0.21 to 0.15) to 0.00 (95% CI -0.21 to 0.21) with different adjustment for cofactors. For individual-level data, the average within-county excess relative risk estimates ranged from 0.07 (95% CI -0.01 to 0.15) to 0.11 (95% CI 0.01-0.27) with similar adjustment. Effect modification by differential county-level radon-smoking correlations appeared insufficient as an explanation for ecologic bias in the Swedish data. On the other hand, adjustment for latitude led to congruence between the two levels of analysis. The results confirm that ecologic studies may be misleading in studies of weak associations, even when major risk factors are accounted for. The large impact of latitude may be unique to Sweden and due to the correlation of latitude both with residential radon and other determinants of lung cancer risk.