Slide 1

Life Expectancy of Oregon AI/ANs
Using Linkage-Corrected Mortality Rates

Jenine Dankovchik, BSc
jdankovchik@npaihb.org

Megan Hoopes, MPH
VictoriaWarren-Mears, PhD, RD, LD

Improving Data & Enhancing Access (IDEA-NW) Project, NW Tribal EpiCenter

Logo Northwest Portland Area Indian Health Board, Indian Leadership for Indian Health

Presented at:
Oregon Public Health Association
Corvallis, OR October 2013

Slide 2

Background

Slide 3

AI/AN disparities & lack of data

• NW AI/AN mortality rates exceeded NHW for 9 out of 10 leading causes
• Life tables are an important population health indicator
• Due to high rates of racial misclassification on death certificates, life tables have not be available for AI/ANs
  • Only published estimate by IHS (1999-2001) found NW AI/AN life expectancy 4 years below US all-races

Slide 4

AI/AN often misclassified on death certificates

• Race not often based on family's own report
• AI/ANs misclassified more frequently than other races/ethnicities
• Net result: mortality measures are underestimated for AI/AN

Slide 5

The IDEA-NW Project

• Improving Data & Enhancing Access (IDEA-NW)
  • Goal: Reduce misclassification of AI/AN race in surveillance systems; disseminate local-level health data to NW tribes
  • Grant funding: AHRQ (2010 to 2013), OMH (2012-2017)
• Northwest Tribal Registry (NTR)
  • All AI/AN registered at IHS or tribal clinic in the NW
  • Augmented with data from urban clinics
• Linkages conducted with public health datasets in OR, ID, and WA

Slide 6

Methods

Slide 7

Data sources

• Death certificates: Oregon, 2006-2010
• Linked with NTR (known AI/AN)
  • Using LinkPlus software, compared data sets to find individuals who appear in both
  • Matched based on identifiers (name, birthdate, SSN, date of death)
  • Probabilistic linkage — allow for errors, misspellings, missing data, nick names, etc.
  • Each pair given a score indicating likelihood of a match
  • "Grey area" matches reviewed by hand

Slide 8

Analysis

• Age — and sex — specific mortality rates computed using 2008-2010
  • race-corrected death certificate data in numerator
  • bridged-race population estimates from NCHS in the denominator
• Three years combined to lessen impact of annual fluctuations
• Complete and abridged life tables constructed using these rates

Slide 9

Results

Slide 10

Racial misclassification

• Linkage with NTR found 699 matches among Oregon deaths
• Of these, 149 were misclassified or missing race
• Overall misclassification rate of 21%
• Linkage with NTR increased case ascertainment of AI/AN deaths by 14%
• Linkage decreased life expectancy at birth by 2.15 years

Slide 11

Life expectancy at birth: Oregon

Bar chart. Life Expectancy at Birth by Race, Oregon 2008 to 2010. Male: AI/AN 73.01, N H W 77.17. Female: AI/AN 76.51, N H W 81.65. All: AI/AN 74.77, N H W 79.43.

Slide 12

Life expectancy at birth: NW Region

Bar chart. Life Expectancy at Birth by Race, Northwest Region 2008 to 2010. Idaho: AI/AN 74.46, N H W 79.45. Oregon: AI/AN 74.77, N H W 79.43. Washington: AI/AN 71.40, N H W 79.95.

Slide 13

Survival patterns by race

Survival Curves for AI/AN and N H W, Oregon 2008 to 2010. Shows lines for N H W, AI/AN, and U S N H W tracking closely until about age 45, then diverging. By age 65, there are 86% N H W and U S AI/AN surviving; and 80% AI/AN surviving. Line diverge further to 43% N H W and U S AI/AN surviving at age 85+ to 28% AI/AN surviving at age 85+.

Slide 14

Survival patterns by sex, AI/AN

Survival Curves by Sex for AI/AN and N H W, Oregon. Shows lines for male and female AI/AN and male and female N H W tracking closely until about age 20, then diverging with AI/AN lower than N H W; and males lower than females. At 75 years, female AI/AN are 13 percentage points lower than female N H W; and male AI/AN are 8 percentage points lower than male N H W.

Discussion

Slide 16

Life tables highlight AI/AN disparities

• Oregon AI/AN life expectancy 4.7 years shorter than the NHW population
  • Gap is larger for females
• Disparity is small among younger AI/ANs, but grows larger as they reach working age and beyond
• This disparity is similar to what has been shown in other indigenous populations
  • 5 years lower for First Nations & Metis in Canada
  • 5 years lower for Native Hawaiians
• And is actually better than in some countries
  • 7 years lower for Maori in NZ
  • 17 years lower for Australian Aboriginals

Slide 17

Impact of racial misclassification

• Found racial misclassification rates of 21%
• If left uncorrected, would have given life expectancy estimates 2.15 years too high
• Federal/state efforts underway improve race data collection upstream, at the point of collection
• Until these efforts have gained traction, data linkage remains an effective and inexpensive option

Slide 18

Tribal uses of data

• Tribes use life tables for:
  • Health assessment
  • Grant writing and reporting
  • Program planning and evaluation
  • Policy and advocacy
• Abridged life tables will be disseminated to Tribes
• Cancer registrars will be conducting AI/AN survival analysis using these life tables

Slide 19

Limitations

• Tribal Registry under-represents urban AI/AN and those with private insurance
  • Captures 75-80% of AI/AN population
• Even with combined data years, small numbers make AI/AN rates unstable
• Age misreporting on death certificates
  • could not correct due to AI/AN race not being identified in Medicare records in the past
• Due to methodological differences, should not be compared directly to national estimates

Slide 20

Thank You!

• The Tribes of Oregon
• Victoria Warren-Mears, PhD (P.I.)
• Idaho, Oregon, and Washington vital statistics staff
• Idaho, Oregon and Washington cancer registrars
• Elizabeth Arias at National Center for Health Statistics
• The International Group for Indigenous Health Measurement
• AHRQ Grant #R01HS19972
• OMH Grant #AIAMP120012
  Contact me:
  jdankovchik@npaihb.org
  503-416-3265