Teen Birth Rates for Age Group 15–19 in the United States by County, 2003-2020

Teen Birth Rates for Age Group 15–19 in the United States by County, 2003-2020

These figures display estimated teen birth rates for age group 15–19 (expressed per 1,000 females aged 15–19) by county and year in a series of interactive maps, median and trend plots, and 95% Bayesian credible bands for each U.S. county and state for 2003–2020. The estimated teen birth rates for age group 15–19 were calculated using hierarchical Bayesian space–time models for the observed teen birth data from 2003 through 2020 for each county and year. More information on methods can be found in the Notes below the visualization.

The first three dashboards show heat maps of estimated teen birth rates for age group 15–19 by county and year. The first dashboard displays the continental United States; the second displays the Northeast Census region to provide both a more granular view and greater detail for counties in smaller states; and the third shows Alaska, Hawaii, and the District of Columbia (D.C.), which is shown separately to allow a more detailed view of D.C. The color scale indicates the magnitude of the estimated county teen birth rate, from lowest (light color) to highest (dark color). The county grid on the right shows the change in estimated teen birth rate by year using the same color scale as the map.

• Use the arrows or the slider to select a year. Click on any state to zoom into it on the map.
• Clicking on a state will update the list to show counties for that state. Selecting a county on the map will highlight that county in the grid.
• Click outside the state to remove the county highlight; click again to zoom back out to the U.S. map. Clicking on the gray “home” icon in the upper right corner of the map also resets the view.

The final dashboard shows trends in the median estimated teen birth rates for age group 15–19 (expressed per 1,000 females aged 15–19) for all counties by state, as well as the estimated teen birth rates for age group 15–19 and corresponding 95% Bayesian credible intervals for each county by year.

• The table shows the estimated teen birth rate for age group 15–19 (expressed per 1,000 females aged 15–19) and the corresponding 95% Bayesian credible intervals (upper and lower limit) by county for the selected state and year. Use the arrows or the slider to select a year.
• Click on a state in any of the maps to update the graphs and table to show trends for that state.
• Use the radio button to view trend lines for either the median estimated teen birth rates for age group 15–19 for all counties of the selected state, or the estimated teen birth rate and corresponding 95% Bayesian credible intervals for the selected county. In the “Credible Intervals” view, click on a county name in the table to update the graph to show the estimated teen birth rate for age group 15–19 and 95% Bayesian credible intervals for that county.

Download the data set in CSV format by clicking the “CSV Format” link. Additional file formats are available for download for each data set at https://data.cdc.gov.

Definitions:

• Estimated teen birth rate: Model-based estimates of teen birth rates for age group 15–19 (expressed per 1,000 females aged 15–19) for a specific county and year. Estimated county teen birth rates were obtained using the methods described elsewhere (1,2,3,4,5,6,7). These annual county-level teen birth estimates “borrow strength” across counties and years to generate accurate estimates where data are sparse due to small number of birth events and/or small population size (1,2,3,4,5,6,7). The inferential method uses information—including the estimated teen birth rates from neighboring counties across years—to provide a stable estimate of the county teen birth rate.
• Median teen birth rate: The middle value of the estimated teen birth rates for the age group 15–19 for counties in a state.
• Bayesian credible intervals: A range of values within which there is a 95% probability that the actual teen birth rate will fall, based on the observed teen births data and the model.

Notes:

Data on the number of live births for women aged 15–19 years were extracted from the National Center for Health Statistics’ (NCHS) National Vital Statistics System birth data files for 2003–2018 (8).

Population estimates were extracted from the files containing intercensal and postcensal bridged-race population estimates provided by NCHS. For each year, the July population estimates were used, with the exception of the year of the decennial census, 2010, for which the April estimates were used.

Hierarchical Bayesian space–time models were used to generate hierarchical Bayesian estimates of county teen birth rates for each year during 2003–2018 (1,2,3,4,5,6,7).

The Bayesian analogue of the frequentist confidence interval is defined as the Bayesian credible interval. A 100*(1-α)% Bayesian credible interval for an unknown parameter vector θ and observed data vector y is a subset C of parameter space Ф such that

1-α≤P({C│y})=∫p{θ │y}dθ,

where integration is performed over the set and is replaced by summation for discrete components of θ.  The probability that θ lies in C given the observed data y is at least (1- α) (9).

County borders in Alaska changed, and new counties were formed and others were merged, during 2003–2018. These changes were reflected in the population files but not in the natality files. For this reason, two counties in Alaska were collapsed so that the birth and population counts were comparable. Additionally, Kalawao County, a remote island county in Hawaii, recorded no births, and census estimates indicated a denominator of 0 (i.e., no females for ages from 15 through 19 years residing in the county from 2003 through 2018). For this reason, Kalawao County was removed from the analysis. Also, Bedford City, Virginia, was added to Bedford County in 2015 and no longer appears in the birth file in 2015. For consistency, Bedford City was merged with Bedford County, Virginia, for the entire 2003–2018 period. Final analysis was conducted on 3,137 counties for each year from 2003 through 2018. County boundaries are consistent with the vintage 2005–2007 bridged-race population file geographies (10).

Sources:

National Center for Health Statistics. Vital statistics data available online, Natality all-county files. Hyattsville, MD. Published annually.

For details about file release and access policy, see NCHS data release and access policy for micro-data and compressed vital statistics files, available from: https://www.cdc.gov/nchs/nvss/dvs_data_release.htm.

For natality public-use files, see vital statistics data available online, available from: https://www.cdc.gov/nchs/data_access/vitalstatsonline.htm.

National Center for Health Statistics. U.S. Census populations with bridged race categories. Estimated population data available. Postcensal and intercensal files. Hyattsville, MD. Released annually.

For population files, see U.S. Census populations with bridged race categories, available from:  https://www.cdc.gov/nchs/nvss/bridged_race.htm.

References:

1. Khan D, Rossen LM, Hamilton B, Dienes E, He Y, Wei R. Spatiotemporal trends in teen birth rates in the USA, 2003–2012. J R Stat Soc A 181(1):35–58. 2017. Available from: http://onlinelibrary.wiley.com/doi/10.1111/rssa.12266/abstractexternal icon.
2. Khan D, Rossen LM, Hamilton BE, He Y, Wei R, Dienes E. Hot spots, cluster detection and spatial outlier analysis of teen birth rates in the U.S., 2003–2012. Spat Spatiotemporal Epidemiol 21:67–75. 2017. Available from: http://www.sciencedirect.com/science/article/pii/S1877584516300442external icon.
3. Rue H, Martino S, Lindgren F. INLA: Functions which allow to perform a full Bayesian analysis of structured additive models using Integrated Nested Laplace Approximation. R package, version 0.0. 2009.
4. Rue H, Martino S, Chopin N. Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations.J R Stat Soc B 71(2):319–92.
5. Martins, T. G., Simpson, D., Lindgren, F., and Rue, H. (2013). Bayesian computing with INLA: New features. Computational Statistics & Data Analysis, 67:68-83.
6. Rue, H. and Held, L. (2005). Gaussian Markov Random Fields. Theory and Applications.Chapman & Hall.
7. Rue, H., Riebler, A., Sigrunn, H.S., Illian, J. B., Simpson, D. P., and Lindgren, F. K. (2017). Bayesian computing with INLA: A review. Annual Review of Statistics and Its Application, 4(1):395-421. Available from: https://www.annualreviews.org/doi/abs/10.1146/annurev-statistics-060116-054045external icon
8. National Center Health Statistics. NCHS data release and access policy for micro-data and compressed vital statistics files. Available from: https://www.cdc.gov/nchs/nvss/nvss-restricted-data.htm.
9. Carlin BP, Louis TA. Bayesian methods for data analysis. Boca Raton, FL: CRC Press, 2009.
10. National Center for Health Statistics. County geography changes: 1990–2012. Available from: https://www.cdc.gov/nchs/data/nvss/bridged_race/County_Geography_Changes.pdfpdf icon.

Suggested citation:

Khan D, Hamilton B, Rossen LM, He Y, Wei R, Dienes E. Teen birth rates for age group 15–19 in the United States by county, 2003–2020. National Center for Health Statistics. 2022.

Designed by J Keralis, D Khan, B Hamilton, A Lipphardt and Y Chong. CDC/National Center for Health Statistics.