Air pollution and COVID-19 mortality in the United States: strengths and limitations of an ecological regression design