Leading Causes of Death and Poverty Census Tracts

This website provides leading cause of death statistics by a socioeconomic status (SES) indicator based on the US census tract poverty level. The site is intended to statistically illustrate disparities in mortality.

The method of using census tract data to analyze disparities is proposed and discussed at the Public Health Disparities Geocoding Project. The method equates tracts to large neighborhoods, and assumes that each person living in a given tract is approximately of the same economic class. The economic class is assumed to be correlated with the number of persons living below the poverty line in a tract. If 20% or more of the persons living in a tract are living below the poverty line, then that tract is considered a poverty area in US Census reports. Therefore, those living in tracts where 20% of the persons are below the poverty line, are considered to be the poorest economic class. Those living in tracts where less than 20% of the persons live below the poverty line are considered to have more economic resources--either by employment, community services or better access to care.

Poverty Classification and 2000 and 2010 Poverty Level Shift

The tracts were linked to two sources of census tract poverty data: For deaths occurring between 2000-2009, the 2000 US Census was used to calculate the percent of persons living below the poverty line. For deaths occurring, after 2009, the poverty level was estimated from the 2007-2011, American Community Survey (ACS). Death records from 2005-2010 were dually classified using 2000 and 2011 poverty level estimates. The dual classification enables the calculation of averages that include data on or after 2010

Groupings from 2000-2009 mortality data will exhibit different counts than 2010 inclusive data, because the two periods use different sources of poverty classification.

In 2000, about 3,338,516 (or about 33% of the state population) were living in census tracts with a poverty level of less than 5%. By 2010, this number had dropped to 1,663,752, or to about 17% of the state population. Other poverty level groups experienced similiar shifts in population. For example, in 2000, 14% of the Michigan population lived in the poorest census tracts, but by 2010, 26% lived in the poorest tracts.

Because of the poverty shift, and because two different classification schemes are used, 2000-2009 exhibit very large changes in population by poverty group as compared to the 2010 classification. Further, counts of death calculated between 2000-2009 are not comparible to counts occuring after 2009. Because age-adjusted rates are good indicators of the relative risk across regions, gender and socioeconomic status, these tables should be used when comparing data before and after the 2010 census.

Unclassified Group Quarter Populations

Certain census tracts do not have an associated poverty level. These tracts tend to be certain instituional or college populations with age groups between 15-34. Counts of county mortality and population are sometimes much lower than estimates of the total. The total male and female population on this site uses the latest version of the CDC's Bridge Population, not the combined poverty census tract populations.