A prospective epidemiologic cohort study of 6,000 residentially stable and non-smoking Seventh-day Adventists (SDA) in California was conducted to evaluate long-term cumulative levels of ambient nitrogen dioxide (NO2) in association with several chronic diseases. These diseases included respiratory symptoms, cancer, myocardial infarction (MI), and all natural causes mortality. Cumulative ambient concentrations of NO2 were estimated for each study subject using monthly interpolations from fixed site monitoring stations and applying these estimates to the monthly residence and work place zip code histories of study participants. In addition, a personal NO2 exposure study on a randomly selected sample of 650 people in southern California was conducted to predict total personal NO2 exposure using household and lifestyle characteristics and ambient NO2 concentrations. It was found that good predictability could be obtained (correlation coefficient between predicted and observed values = 0.79) from a model predicting personal NO2. The resulting regression equations from the personal NO2 exposure study were applied to the epidemiologic study cohort to adjust ambient concentrations of NO2.