Data & Measurement Glossary

County-Wide Indicators.  At the top of each county’s Community Health Improvement webpage, you will find two county-wide indicators: Years of Potential Life Lost (YPLL) and Disparities in Life Expectancy at Birth.  These broad county indicators allow us to track progress on community health improvement through a wider, long-term lens.  The measure of YPLL gives us a glimpse into the overall health of the county, while the disparities indicator ensures that gaps in health are noted.  More detail about these indicators and how they are calculated is provided below.  However, it’s important to note that these are not measures we expect to change quickly.   These are long-term indicators of the health status of our community. 

Issue Specific Indicators.  Each NJHC county-based workgroup has selected population-level indicators specific to the topic they are working on.  These indicators appear in a white box, under the workgroup’s results statement.  You will see that some workgroups have many indicators for one issue, while others are still working to find the most relevant data sources.  Important to note is that we see these indicators as leverage points to improving the health of the community overall – if we improve these indicators in the short term, we are likely to see the county-wide indicators (YPLL and Disparities in Life Expectancy at Birth) improve in the long-term. 

YPLL is a measure of premature death.  It captures early deaths in a community by subtracting the age of each person who died during the designated time period from their life expectancy, and converting the total amount of “years lost” to a population rate.   

For example, the average life expectancy for New Jersey Residents is 80 years.³   If an individual were to die at 70 years old, he would contribute 10 years of lost life to this calculation.  If he were to die at 50 years old, he would contribute 30 years of lost life.  In this manner, earlier deaths are weighted heavier in this statistic.

Unlike traditional mortality measures, the calculation approach for YPLL emphasizes the impact of premature death among younger age groups and its cost to the community. Those familiar with the YPLL measure will note that it is usually portrayed using data in 3 year periods and benchmarked against a life expectancy of 75 years.  Since we, in the collaborative, are interested in the trend in these data overtime – not any individual period – we present this as year-by-year numbers so that we have a better picture of the curve or trend of the data.  However, caution needs to be taken to not interpret any significant dip/climbs in specific year data.  Additionally, we chose to use 80 years as our benchmark rather than 75 because it is more reflective of actual life expectancy in New Jersey, and thus gives us a more accurate picture.

Life expectancy at birth is an indicator of the overall health of a population, and refers to the average length of time a newborn is expected to live if mortality patterns (at the time of birth) were to remain consistent in the future⁴.  Life expectancy is influenced by a number of things – genetics and DNA, health behaviors, environment and social factors. 

Because our collaborative is organized at the county level, we wanted to look at within-county disparities—differences in life expectancy between individuals born into different towns within the same county.  For example, a disparity score of 0 would represent the same life expectancy across all towns in the county.

In order to do this, we first calculated the life expectancy at birth for each municipality. Because the population size varies dramatically, we used the South East England Public Health Observatory (SEPHO)⁵ Tool to complete this calculation. This tool has been recommended by the Sub-County Assessment of Life Expectancy (SCALE) Workgroup, which is organized by the Centers for Disease Control and Prevention (CDC) and the Council of State and Territorial Epidemiologists (CSTE). The SEPHO tool applies an adjusted version of the Chiang methodology⁶ for small populations⁸ which incorporates Silcocks’ methodology for adjustments to the variance of the final age interval⁷.

Our calculations used weighted moving sums to get at both the population size and the number of deaths by age group (the two variables that go into calculating life expectancy at birth) in every town, annually.

For those who are interested, here are the specific calculations we used:

• Population size. First, we assigned population counts to every year included in our calculation. These population counts were based on the 2000 or 2010 Census (or a combination of the two). 2004 and 2005 were assigned the values from the 2000 census; for the years 2006-2009, an average of the two censuses were used; for 2010-2015, the 2010 Census value was used.  Second, using these assigned population counts, we calculated a 3-year population count by summing the count from a particular year, with its previous and following year.  For example, the population count for Town A in 2014 is calculated by adding the counts for 2013, 2014 and 2015. For data reliability purposes, we excluded seven municipalities within the five counties, that had 3-year population sums of less than 5,000 residents.

• Death counts. We calculated the 3-year sum of deaths during a particular year, the deaths from the previous and the following years. For example, the 2011 death count for Town A is calculated by adding all deaths from 2010, 2011 and 2012. Death count data is provided annually by NJSHAD.

The disparity in life expectancy at birth score for each county represents the difference between the town with the highest and the town with the lowest life expectancy for each year. It’s important to note that the towns with the highest and lowest life expectancy estimates vary from year to year.

The county-wide indicators were calculated by our Data Committee.  Data comes from:

• Annual data on death counts in five-year age groups used to calculate YPLL and life expectancy at birth was obtained from the New Jersey State Health Assessment Database (NJ SHAD).
• Population estimates for each municipality were taken from the 2010 U.S. Census.
• The error bars on each graph represent the margin of error, calculated at a 95% confidence interval. These statistics will be updated as more recent data becomes available.

For the issue-specific indicators, detailed source information is available by clicking on each indicator.

Years of Potential Life Lost:  Since public health data is variable year-to-year, and reporting periods often lag, it’s important to not interpret any one year of this graph – but rather to look at it as a trend over time.   

Disparities in Life Expectancy:  When calculating life expectancy at birth for each municipality in order to derive a disparity score, a clear limitation is the small population size of some municipalities. While the Chiang (II) methodology is intended to be used with small populations, there is still greater variability in life expectancy scores across years among municipalities with smaller populations as compared to those with larger populations as a result of the methodology utilized.

Issue Specific Indicators: Individual limits can be identified for each indicator.  However, overall there is a lag time with these metrics, as well.

(1) Fontaine, K. R., Redden, D. T., Wang, C., Westfall, A. O., & Allison, D. B. (2003). Years of life lost due to obesity. Jama, 289(2), 187-193.

(2) Centers for Disease Control and Prevention (CDC). (2012, May 18). Section 3: Mortality Frequency Measures. Principles of Epidemiology in Public Health Practice, Third Edition. An Introduction to Applied Epidemiology and Biostatistics. Retrieved from http://www.cdc.gov/ophss/csels/dsepd/ss1978/lesson3/section3.html

 (3) New Jersey Department of Health (NJDOH). (2016, October 27). Important Facts for Life Expectancy at Birth. Retrieved from https://www26.state.nj.us/doh-shad/indicator/important_facts/LifeExpectancy.htm

 (4) United Nations. (2000). Charting the Progress of Populations. Retrieved from http://www.un.org/esa/population/publications/charting/9.pdf

 (5) Council of State and Territorial Epidemiologists (CSTE). (2014). Sub-County Assessment of Life Expectancy (Scale) Project. Retrieved from http://www.cste.org/page/SCALE/Sub-County-Assessment-of-Life-Expectancy-SCALE-Project.htm

 (6) Chiang, C.L. The life table and its construction. In: Introduction to stochastic processes in biostatistics. New York: Wiley, 1968:189–214.

 (7) Silcocks, P.B.S., D.A. Jenner, and R. Reza, Life expectancy as a summary of mortality in a population: statistical considerations and suitability for use by health authorities. Journal of Epidemiology and Community Health, 2001. 55(1): p. 38-43.