Small Area Life Expectancy As a Tool for Policy, System and Environment Change

BACKGROUND:  Core functions of a local health department include assessment of health and social needs, policy development (strategies to protect health/promote well-being), and assurance (measure and evaluate progress towards social and health equity). King County, WA, home to 2.1 million people, often fares well in comparison of health and social factors to other counties in the U.S. However, assessment at a state or county level often masks place-based inequities. Public Health, Seattle & King County (PHSKC) frequently reports health data using 48 sub-county city or neighborhood aggregations called Health Reporting Areas (HRA), recognizing that effective policy, system and environment (PSE) change often begins at the community level. However, even the HRAs mask inequities. Noting a geographic pattern in LE variation across HRA estimates led to LE census tract (CT) mapping.

METHODS:  LE was calculated using Chiang II method with Silcock’s adjustment was used to calculate aggregated 5 year LE estimates for 2008-2012, using geocoded death data and population from the statewide Small Area Demographic Estimates for HRAs and CTs.

RESULTS:  HRA estimates ranged from 86.2 to 76.6, where tract LE varied from 95.5 to 71.5 (after controlling for unreliable tracts), with a county average 81.6 years. The strong spatial pattern led to questions about major contributors, and additional small area analysis of other social, economic, housing, and health conditions at a CT level. Comparison of the lowest 10% and highest 10% of tracts show large disparities in outcomes. Adult residents in CT in the lowest 10% died 13 years earlier, were 3x more likely to report frequent mental distress, 4x more likely to be current smokers, 2x more likely to be obese, 4x as likely to be unemployed, and have almost 3x the rate of diabetes compared to the highest decile tracts. More than half (54%) of the residents in in the lowest decile CT lived below 200% of the Federal Poverty Level.

CONCLUSIONS: Drawing attention to these social determinants of health that mirror the LE pattern brought funders and community partners to the table. The variation between the absolute and relative inequity between HRAs and CTs illustrates the value of being able to disaggregate geographies. Using LE as an initial scan of place-based inequity led to interest in other measures that could contribute to potential explanations of the LE variation, and provided opportunities for funder and community engagement for PSE at the local level.