Are Local Health Department Expenditures Related to Racial Disparities in Mortality? David Grembowski Douglas Conrad Betty Bekemeier William Kreuter University.

Are Local Health Department Expenditures Related to Racial

Disparities in Mortality?

David GrembowskiDouglas ConradBetty BekemeierWilliam Kreuter

University of Washington

Funded by the Robert Wood Johnson Foundation &Changes in Health Care Financing and Organization

Healthy People 2010

Goal 1: Increase quality and years of healthy life

Goal 2: Eliminate health disparities Gender, race/ethnicity,

education/income, disability, rural/urban, sexual orientation

Population Health Paradox

Population health has increased in many developed countries

Disparities in population health have increased in many developed countries

Absolute Change vs. Relative Disparity

in Infant Mortality by Racial/Ethnic Group

2002- Absolute 1950 2004 Change

Black 44 14 30 White 27 6 21 B/W Rate Ratio* 1.63 2.33 .70 bigger

gap* 1950 Black/White Rate Ratio = 44/27 = 1.63

Infant mortality rate: rate of deaths in children less than 1 year old per 1,000 live births

Health, United States, 2007

Absolute Change vs. Relative Disparityin All-Cause, Age-Adjusted Mortality by

Racial/Ethnic Group

Absolute1950 2004

ChangeBlack 1,722 1,027 695White 1,411 786 625B/WRate Ratio* 1.22 1.31 .09 bigger

gap* Black-White rate ratio in 1950 is 1,722/1,411 = 1.22

Rates are per 100,000 population

Health, United States, 2007

Race & Gender

Health, United States

Determinants of Population Health& Health Disparities

Physical & SocialEnvironment

Behavior

Human Biology

PopulationHealth &Disparities

Medical Care

Determinants of Population Health& Health Disparities

Physical & SocialEnvironment

Behavior

Human Biology

PopulationHealth &Disparities

Medical Care

Public Health

Health Policy

To reduce racial/ethnic disparities in mortality,

How much of society’s resources should be invested in local public health systems vs. other determinants of population health and health disparities?

What is the evidence that local public health funding reduces racial/ethnic disparities in mortality?

Can Local Public Health Reduce Racial/Ethnic

Disparities?

Yes! Population-based interventions that influence everyone

have the potential to reduce health disparities Water fluoridation reduces disparities in oral health Raising the health of the worst off fastest

No! Policies to improve overall population health may

increase disparities because people with more resources are more likely to take advantage of them

Goal of improving population health may conflict with goal of reducing health disparities

Link & Phelan 2005; Mechanic 2002

Few Studies Medical Care

10% increase in medical expenditures per capita associated with 1-2% reduction in mortality rates (Hadley, 1982)

Public Health 10% increase in local health department

spending per capita associated with 1-7% decline in mortality rates (Mays & Smith 2007)

In 1907-1910, 1 standard deviation increase in city health expenditures associated with decline in infant mortality from 14.9 to 11.5 deaths per 100 children under age 1 (Costa & Kahn 2006)

What Matters for Policy?

For local health departments (LHDs):

Is it the amount of LHD spending per capita- or -

Is it the share of total public revenue in a local area that goes to a LHD

that matters in reducing racial/ethnic disparities in mortality rates?

Methods

Time-Trend Ecologic Study Design

Test whether 1990-1997 changes in LHD funding

are associated with

1990-1997 changes in Black/White mortality rates

Data Sources LHD expenditures: 1990 & 1997 National

Profiles of Local Health Departments from the National Association of County and City Health Officials (NACCHO)

1990 & 1997 Black and White mortality rates from the CDC Compressed Mortality File

1989-91 & 1996-98 Black and White infant mortality rates

Data Sources U.S. Census of Government Organizations

U.S. Population Census, 1990 county

Area Resource File (Bureau of the Health Professions), county

Medicare county expenditures per capita (Centers for Medicare and Medicaid Services)

Rural/urban county commuting codes (Department of Agriculture)

Number of Local Areas Population: 3,256 LHDs in 1990 (NACCHO)

857 local areas with Black & White all-cause mortality rates 562 local areas (66%) have 1990 & 1997 Profile

Surveys

187 local areas with Black & White infant mortality rates in 1989-1991 & 1996-98 133 local areas (71%) have 1990 & 1997 Profile

Surveys

All-Cause, Age-Adjusted Mortality Variables

Absolute change for each race Change = (1997MR - 1990MR)

Males & Females Males Females

Change in Black/White Rate Ratio Change = (1997 B/W Ratio - 1990 B/W

Ratio)

Policy Variables Change in LHD Expenditures per capita

Change = (1997 LHD$ - 1990 LHD$)

Change in percentage share of total local public revenue going to LHD Percentage = LHD expenditures/(total

revenue from county govt + cities & towns + school districts + special districts)

Change = (1997% - 1990%)

Local Area Control Variables

Social Environment Area Characteristics Female Rural/Urban Education Home value Income Rent amount

Income inequality White collar occupation Unemployment Medical Care Household size Medicare

expenditures Single female household Percent Black Health Care System

Foreign residents Hospital beds English speaking Physicians Veterans Drive to work People in mental institutions People in correctional institutions

Data Analysis

Descriptive statistics

General Estimating Equations (GEE)

40 binary (0,1) State variables to control for State-level influences on local areas

Standard errors adjusted for clustering of local areas by Federal Region

Test for 2-way causation between LHD funding and mortality

ResultsLHD Funding (Medians)

1997-1990 1990 1997 Change

LHD expenditures $22* $30 $7 per capita

% share of public0.99% 1.17% 0.17 revenue to LHD * Adjusted 1997 dollars

Average Mortality Rates

1997-1990 1990 1997 Change

Black Mortality Total 1,235 1,152 -83

Female 969 932 -37 Male 1,646 1,479 -167

White MortalityTotal 956 914 -41

Female 754 743 -11 Male 1,244 1,150 -94___________________________________________________________________________________

Rates are per 100,000 population

Absolute Change in Black Mortality Rates

(1997 – 1990)

Mortality Rates & 1990 + 1997 Profile Surveys

Areas Areas with without Profile Profile Surveys Surveys Both Yrs Both Yrs p-value

1997-1990 Change Black mortality -83 -34 .004 White mortality -41 -40 .868

Black infant mortality -3.0 -1.8.070

White infant mortality -1.5 -1.2.060

___________________________________________________________________________________

1997 - 1990 Absolute Change in Black Mortality Rates

1997-1990 change in LHD expenditures per capita

are not associated with

1997-1990 change in Black total mortality, Black female mortality and Black male mortality

Change in LHD Share of Public Revenue

& Change in Mortality Rates

A 1-percent increase in share of public revenue to LHD associated with small declines in Black and White mortality rates:

Avg Change1997-1990

Mortality Decline

Black Total Mortality -83 -1.24#White Total Mortality -41 -1.65#

Black Female Mortality -37 -0.98White Female Mortality -11 -1.35#

Black Male Mortality -167 -2.23White Male Mortality -94 -1.41#___________________________________________________________________________________

Rates are per 100,000 population

# Statistical test for 2-way causation significant

1990 & 1997 Black/White Rate Ratios

(Averages)

1997-1990

1990 1997 Change

Total 1.30 1.27 -.03

Female 1.30 1.28 -.02 Male 1.34 1.30 -.04

____________________________________________________________________________

Change in Black/White Rate Ratios

(1997 – 1990)

1997 - 1990 Change inBlack/White Mortality Rate Ratios

1997-1990 changes in LHD expenditures per capita

or changes in share of public revenue for LHD

are not associated with

1997-1990 changes in Black/White rate ratios

Average Infant Mortality Rates

1997-1990 1990 1997 Change

Absolute Change Black 18 15 -3

White 8 6 -2

B/W Rate Ratio 2.35 2.43 +.08_________________________________________________________________________

Rates are deaths in children less than 1 year old per 1,000 live births

1997 - 1990 Change inWhite Infant Mortality Rates

1997-1990 changes in LHD expenditures per capita

or share of public revenue for LHD

are not associated with

1997-1990 changes in White infant mortality rates

1997 – 1990 Change in Black Infant Mortality Rates

A 1-percent increase in share of public revenue to LHD associated with: 0.36 increase in Black infant mortality rate (p =.002) 0.07 increase in B/W rate ratio (p<.001)

A $1 increase in LHD expenditures per capita associated with: 0.01 increase in Black infant mortality rate

(p=.009)

Conclusions

Per capita LHD expenditures and share of public revenue going to LHD generally not related to disparities in Black and White mortality rates

A greater share of public revenue for LHDs related to small decline in Black male mortality

Conclusions

Per capita LHD expenditures and share of public revenue going to LHD related to greater Black infant mortality

We also are analyzing whether LHD services are related to Black/White disparities in mortality

Limitations

Correlation or causation

Omitted variables (population-level smoking, medical care rates)

No data for LHDs without Profile Surveys

No data for how LHDs spent funds in 1990/97 Profile Surveys

No data for public health spending by other community agencies

Black county populations often too small to calculate cause of death mortality rates (by gender)

Only 2 racial/ethnic groups