Interpreting the data
What does the targeting lead outreach map tell us?
The targeted lead outreach map combines information on testing rates and lead exposure risk in order to identify where and how these two factors overlap in geographic space.
The targeted lead outreach map can be used to prioritize areas for increased testing for lead, for example, by identifying census tracts with high exposure risk and low testing rates.
Tracts that fall in the highest risk category are of particular concern because they all have highest levels of risk of lead exposure based on housing, poverty, and population estimates for children under age 6. However, tracts that also have relatively low lead testing rates among children under age six represent potential targets for greater resources and lead testing.
What does the map of lead exposure risk tell us?
The lead exposure risk index can be used to identify areas of high and low risk of childhood lead exposure, based on the presence of known risk factors.
The lead exposure risk index is a neighborhood-level (census tract) estimate of risk of childhood lead exposure based on three factors: housing age, poverty level, and population of children younger than age six.
These data do not tell us for sure where lead poisoning cases are occurring. Rather, they show where the greatest potential for lead exposure exists based on risk factors that are well-established in peer-reviewed scientific literature.
What does the map of lead testing rates tell us?
The lead testing rate map can be used to assess the relative levels of lead testing among children younger than age six in different census tracts throughout Colorado.
- The lead testing rate is the total number of children younger than age six living in a given census tract and tested for lead between 2013 and 2017 expressed as a percentage of the estimated total population of children younger than age six living in the same tract. The rate is then categorized into high, moderate, or low compared to the rest of the state.
Limitations of the data
Limitations of lead exposure risk data:
- Accuracy of the exposure risk estimate depends on the accuracy of the demographic estimates obtained from the U.S. Census. All input data is taken from the U.S. Census American Community Survey’s (ACS) 2013-2017 five-year population estimates for census tracts in Colorado. We have flagged census tracts as ‘reliable’ if the population estimate for children under age 6 has a coefficient of variation less than 40.
Limitations of lead testing rate data:
- Accurate testing rates depend on the completeness of CDPHE’s blood lead testing registry. It is possible that some lead testing of children younger than age six occurs and goes unreported to CDPHE, even though reporting this information is required according to the Colorado State Board of Health.
- Accuracy of testing rates also depends on the accuracy and completeness of address information associated with the testing record. Not all tests reported to CDPHE have enough address information to be able to assign them to a census tract.
- Statewide, approximately 80% of tests reported to CDPHE are able to be geocoded to at least the census tract level. The remaining tested individuals have test records that contain only enough information to assign the individual to either their zip code, county, or state of residence.
- Accuracy of testing rates also depends on accuracy of the population estimates used to calculate the rate. Population estimates from the American Community Survey are subject to varying degrees of sampling variability, meaning the population estimate is less reliable for census tracts with small overall populations. We flagged census tracts as ‘reliable’ if the population estimate for children under age 6 has a coefficient of variation less than 40.
- Caution should be used when interpreting any map of lead testing rates in Colorado.
Targeted screening map calculations:
- The lead screening tool map was created by dividing the testing rate and exposure risk scores into high, medium, and low categories. The categorization was determined based on deciles.
- The high category contains the top 30% of the distribution, the low category contains the bottom 30% of the distribution, and the medium category contains the middle 40% of the distribution.
- Crossing the high, medium, and low categories for both testing rates and exposure risk results in nine categories: High Risk/Low Testing, High Risk/Medium Testing, High Risk/High Testing, Medium Risk/Low Testing, Medium Risk/Medium Testing, Medium Risk/High Testing, Low Risk/Low Testing, Low Risk/Medium Testing, Low Risk/High Testing.
Lead exposure risk index calculations:
- The lead exposure risk index is calculated using three different types of U.S. Census demographic data: 1) estimates of the number of housing units built in different time periods, 2) estimates of the overall poverty rate, and 3) estimates of the population of children under age 6. All input data is taken from the U.S. Census American Community Survey’s (ACS) 2013-2017 five-year population estimates for census tracts in Colorado (B17001, B25034, & S1701 data files).
- The methodology is an adaptation of a method developed by the Washington State Department of Health. More details about this methodology can be found here: https://github.com/voxmedia/data-projects/blob/master/vox-lead-exposure
Testing rate numerator data:
- We obtained the numerator for the lead testing rate calculation by first geocoding to their census tract of residence all unique children who were tested for lead and whose tests were reported to CDPHE during each year between 2013 and 2017.
Testing rate denominator data:
- We obtained denominator data from the U.S. Census American Community Survey’s (ACS) 2013-2017 five-year population estimates for census tracts in Colorado. The estimate for children under age 6 comes from the B17024 data file, which we downloaded from the American Fact Finder website.