Findings from the Houston Health Survey

WHAT ACCOUNTS FOR The 2012 housTon HEALTH DISPARITIES?

educaTion survey: FINDINGS FROM THE Public PercePtions HOUSTON SURVEYS in a(2001-2013) critical time

#

November April 20142013

Kinder Institute for Urban Research Rice University, MS 208 6100 Main Street Houston, TX 77005 Telephone: (713) 348-4132 http://www.kinder.rice.edu Written by Written by Stephen L. Investigator Stephen L.Klineberg, Klineberg,Principal Principal Investigator Jie Wu, Research Project Manager Jie Wu, Research Project Manager Kiara Douds, Post-Baccalaureate Fellow Cristina Barrera, Research Intern Contact us for more information at kinder@rice.edu.

Contact us for more information at kinder@rice.edu

Copyright Š 2013 by Kinder Institute for Urban Research. All rights reserved.

Copyright Š 2014 by Kinder Institute for Urban Research. All rights reserved.

TABLE OF CONTENTS INTRODUCTION..................................................................................................................................... 3 SOCIAL STRUCTURES AND HEALTH DISPARITIES.......................................................................... 4

Measuring Self-Rated Health..................................................................................................................... 4 Figure 1 Self-Rated Health Status in Harris County, from Two Data Sources

Aging and Health ........................................................................................................................................ 5 Figure 2

Figure 3

Self-Rated Health Status by Age Group (2001-2013 KIHAS, Combined) The Prevalence of Serious Health Problems by Age Group among the Respondents Who Reported Fair or Poor Health (HAHS)

The Relationship between Inequality and Health.................................................................................... 6

Self-Rated Health Status across Major Metropolitan Counties, Ranked by Their GINI Coefficients (2011 ACS 5-Year Estimates)

Table 1

Figure 4

The Percent of Residents in Fair or Poor Health, by the GINI Coefficients of their Counties (2013 BRFSS)

The Individual Contributions of Education, Income, and Ethnicity..................................................... 7 .

The Role of Education................................................................................................................... 8 Figure 5

Self-Rated Health Status by Educational Attainment (2001-2013 KIHAS, Combined)

The Importance of Income............................................................................................................ 9 Figure 6

Self-Rated Health Status by Annual Household Income (2001-2013 KIHAS, Combined)

The Impact of Ethnicity................................................................................................................. 10 Figure 7

Levels of Household Income in Five Ethnic Communities (2001-2013 KIHAS, Combined)

Figure 8 Differences in Self-Rated Health Status in Five Ethnic Communities (2001-2013 KIHAS, Combined)

Figure 9 Differences in Self-Rated Health Status among U.S.-Born Asians and in

Houston’s Four Largest Asian Immigrant Communities (2001-2013 KIHAS, Combined)

The Structural Correlates of Self-Rated Health..................................................................................... 12

Table 2 Logistic Regression of the Structural Correlates on Self-Rated Health

Status (2001-2013 KIHAS, Combined)

Figure 10 The Perceived Determinants of a Person’s Health, Average Ratings (HAHS)

ACCESS TO HEALTH CARE...............................................................................................................................15

Health Insurance Coverage...................................................................................................................... 15

Figure 11

Health Insurance Coverage by Ethnicity among Respondents Aged 18 to 64 (2001-2013 KIHAS, Combined)

Table 3 Logistic Regression of Health Insurance Coverage on Self-Rated Health

Status Controlling for the Structural Variables (2001-2013 KIHAS, Combined)

Immigration and Health Insurance among Hispanics. .......................................................... 17

Table 4 Logistic Regression of Immigration Status and of Insurance Coverage on

Self-Rated Health Status among Latinos, Controlling for the Structural Variables

Health Disparities

1

Access to Quality Health Care in Houston............................................................................................. 18

Figure 12

Figure 13

Respondents’ Ratings of the Quality of Health Care Available to Them in the Houston Area, by Household Income (HAHS) Respondents’ Ratings of the Quality of Health Care Available to Them in the Houston Area, by Insurance Coverage (HAHS)

Medical Visits................................................................................................................................ 19

ENVIRONMENTS, NEIGHBORHOODS, AND HEALTH........................................................................... 20

The Importance of Air and Water Quality.............................................................................................. 20

Air Quality.................................................................................................................................... 20 Figure 14

Ratings of Air Quality and Concerns about Air Pollution (HAHS)

Figure 15 Self-Reported Health Status by Proximity to the APWL ZIP Codes (2001-2013 KIHAS, Combined)

Table 5 Logistic Regression of Proximity to the APWL ZIP Codes on Self-Reported

Health, Controlling for the Structural Variables and for Insurance Coverage

Water Quality............................................................................................................................... 23

Figure 16 Ratings of Water Quality and Concerns about Water Pollution (HAHS)

Neighborhood Characteristics and Health-Related Behaviors .......................................................... 24

Food Insecurity............................................................................................................................. 24

Figure 17 The USDA Map of Houston Area Food Deserts, Reflecting Both Income and Access

Physical Activity........................................................................................................................... 25

Figure 18 Respondents’ Satisfaction with Their Level of Physical Activity by

Their Self-Rated Health Status (HAHS)

Figure 19 The Reported Frequency of Vigorous Physical Activity by Income Level (HAHS)

Walkability................................................................................................................................... 27

Figure 20

Neighborhood Amenities in the Houston Area by the Location of Home Residence (HAHS)

The Use of Public Resources........................................................................................................ 29

Figure 21

The Perceived Physical Activity in the Neighborhood by Proximity to Local Parks and Trails (HAHS)

Figure 22 The Availability of Health-Promoting Resources and Their

Frequency of Use during the Past Year among Harris County Residents (HAHS)

Participation and Belonging....................................................................................................... 30 Table 6

Logistic Regression of the Composite Measures of Neighborhood Belonging and Participation on Self-Reported Health Status (2012 KIHAS)

SUMMARY AND CONCLUSIONS.................................................................................................................... 32 APPENDIX A. THE 2012 HAHS METHODOLOGY...................................................................................... 35 APPENDIX B. LOCAL HEALTH ADVISORS................................................................................................. 36 REFERENCES......................................................................................................................................................... 37

2 Kinder Institute for Urban Research

INTRODUCTION This report seeks to identify the forces that account for health disparities in the Houston region. It makes use of questions asked in the past 13 years of the annual “Kinder Institute Houston Area Survey” (KIHAS), buttressed by findings from the third of the institute’s three focused surveys, known as the SHEA studies (“Surveys of Health, Education, and the Arts”). Supported by a grant from Houston Endowment Inc., the studies were designed to assess the experiences, beliefs, and attitudes of Harris County residents with regard specifically to these important areas of urban life. The three separate surveys and the reports1 presenting their central findings complement the Kinder Institute’s continuing annual studies (the KIHAS), which have been tracking, through 33 years of systematic surveys (1982-2014), the demographic patterns, experiences, attitudes, and beliefs of Houston area residents during a period of remarkable change.2 In the process of developing the survey instrument to be used for the SHEA “Houston Area Health Survey” (HAHS), the research team at the Kinder Institute met periodically during 2011 and 2012 with local and national public health experts to fashion the questions we would ask of a scientificallyselected representative sample of 1,200 Harris County residents. The goal was to measure systematically area residents’ self-reported health status, their experience with Houston’s health care delivery systems, and the health-related characteristics of their neighborhoods. The actual telephone interviews were conducted by the Survey Research Institute at the University of Houston between June 6 and July 17, 2012, with 70 percent of the respondents reached by landline and 30 percent by cell phone.

After the interviews were completed, the data were “weighted” to correct for variations in the likelihood of selection and to align the sample more closely with known population characteristics, such as gender, age, education, race and ethnicity, home ownership, and county population and density.

The weighting procedures provide a more accurate and reliable reflection of the actual attitudes and experiences to be found within the Harris County population as a whole. As in previous surveys, the final (weighted) sample distributions still slightly overrepresent respondents aged 60 and older and underrepresent those aged 30 to 44, but they mirror almost perfectly the census figures for Harris County with regard to the distributions by gender, ethnicity, and educational attainment. Unless otherwise indicated, the findings presented in the following charts and discussed below are based on the weighted data. Appendix A provides additional information about these and other aspects of the survey methodology.

Listed in Appendix B are the local advisors who helped the research team develop the survey instrument and offered comments on the preliminary findings and on earlier drafts of this report. The survey instrument that was used in the HAHS and the responses given by Harris County residents on all the questions can be found on the institute’s website at kinder.rice.edu/shea.

We begin this report with an analysis of the findings from both the HAHS and the past 13 years of the KIHAS in an effort to identify empirically the most important structural forces (such as age, education, income, and ethnicity) that demonstrably underlie and help to account for the wide disparities in the health and well-being of Harris County residents. We go on to explore the bases for individual differences in area residents’ access to quality health care and to assess the importance of the relationship between health insurance coverage and self-reported health status. Finally, we ask about the health effects of environmental pollution and of the social characteristics of neighborhoods, and we briefly mention some of the initiatives under way in the Houston area that may make a significant difference in shaping the overall health of this community. ■

Health Disparities

3

SOCIAL STRUCTURES AND HEALTH DISPARITIES Measuring Self-Rated Health

of neighborhoods, and we briefly mention some of the initiatives under way in the Houston area 1 presents thethis responses on this question that make a significant difference shaping theFigure overall health of community.

Earlymay in the Houston Area Health Survey in (HAHS),

from both sources: the HAHS and the combined 13 years of the KIHAS (2001-2013). A plurality of all Harris County residents (32 to 38 percent) said that they were in “good” health; another 44 to 46 percent reported that their health was either “excellent” or “very good.” Fully one-fifth (19 to 21 percent) of the survey participants rated their current state of health as “fair” or “poor.” The “2010 Health of Houston Survey,” conducted by the School of Public Health at the “In University of would Texas, Early in the Houston Area Health Survey (HAHS), the respondents were asked: general, also found that 20 percent of Harris County you that your overall state ofinphysical health these days is excellent, very good, good, fair,resior Thatsay same question was included each of the dents reported status as fair or poor.5 past 13 years (2001-2013) of the Kinder health has poor?”

This simple measure ofannual self-reported been foundtheir to behealth a remarkably reliable The close correspondence among these various data Institute Houston Area Survey which Responses has indicator of a person’s actual (KIHAS), state of health. on this question closely correlate with the 3 sourcesand strengthens confidence in the reliability of reachedofmore than 20,000 Harris County residents they are strong predictors of results physical examinations by health professionals, the findings.

We will be particularly interested to 4 during those years. When asking about the role of mortality. learn what the surveys can tell us about those 20 demographic or structural factors in accounting for percent of area residents who indicate that their individual in self-rated That samedifferences question was includedhealth, in eachweofwill the pastcurrent 13 years (2001-2013) of the annual Kinder state of health is only fair or poor. make use of that more extensive set of data.

We will Institute Houston Area Survey (KIHAS), which has reached more than 20,000 Harris County draw on the HAHS for its rich additional measures residents during those years. When asking about the role of demographic or structural factors in of various environmental and neighborhood characcounting for individual differences in self-rated health, we will make use of that more extensive acteristics and their impact on health outcomes. the respondents were asked: “In general, would you say that your overall state of physical health these days is excellent, very good, good, fair, or poor?”

This simple measure of self-reported health has been found to be a remarkably reliable indicator of a person’s actual state of health. Responses on this question closely correlate with the results of physiMeasuring Self-Rated Health 3 and they cal examinations by health professionals, are strong predictors of mortality.4

SOCIAL STRUCTURES AND HEALTH DISPARITIES

set of data. We will draw on the HAHS for its rich additional measures of various environmental and neighborhood characteristics and their impact on health outcomes.

Figure 11 –- Self-Rated Health Status Status in in Harris HarrisCounty, County,from fromTwo TwoData DataSources Sources Figure Self-Rated Health 38

40 35

32

PERCENT OF RESPONDENTS

30 25 20

“In general, would you say that your overall state of physical health these days is excellent, very good, good, fair, or poor?”

25 26 19 20 16 13

15 10

6

5

5

The 2012 HAHS 2001-2013 KIHAS, combined

0 Excellent

Very good

Good

Fair

Poor

4 Kinder Institute for Research Figure 1 presents theUrban responses on this question from both sources: the HAHS and the combined 13 years of the KIHAS (2001-2013). A plurality of all Harris County residents (32 to 38 percent) said that they were in “good” health; another 44 to 46 percent reported that their health was either “excellent” or “very good.” Fully one-fifth (19 to 21 percent) of the survey participants rated their

Not surprisingly, age is significantly associated with self-reported health statu participants were consistently more likely than the younger respondents to giv when asked about their current state of health. In the HAHS, 24 percent of tho the age of 60 reported that their overall physical health was fair or poor, comp percent of respondents aged 18 to 29. Figure 2 shows that this relationship is c more extensive KIHAS data: Over the 13 years of surveys, 32 percent of all th 60 or older at the time of the interviews said they were in fair or poor health, b just 16 percent of the survey participants who were under the age of 30.

Aging and Health Not surprisingly, age is significantly associated with self-reported health status: The older survey participants were consistently more likely than the younger respondents to give negative ratings when asked about their current state of health. In the HAHS, 24 percent of those who were over the age of 60 reported that their overall physical health was fair or poor, compared to just 15 percent of respondents aged 18 to 29. Figure 2 shows that this relationship is confirmed by the more extensive KIHAS data: Over the 13 years of surveys, 32 percent of all the respondents aged 60 or older at the time of the interviews said they were in fair or poor health, but this was true of just 16 percent of the survey participants who were under the age of 30.

Figure 2 – Self-Rated Health Status by Age Group (2001-2013 KIHAS, Combined)

The HAHS asked the respondents if they had “any physical health problems right now that prevent you from doing any of the things people your age normally can do?” More than a third (34 percent) of the survey participants who were over the age of 60 said they did indeed have a physical problem of that sort, but this was true for just 5 percent of the respondents under age 30. Not surprisingly, those of any age who said their health was fair or poor were far more likely to report having a physical problem that affected their daily lives.

The HAHS asked the respondents if they had “any physical health problems r you from doing any of the things people your age normally can do?” More tha percent) of the survey participants who were over the age of 60 said they did i physical problem of that sort, but this was true for just 5 percent of the respon Not surprisingly, those of any age who said their health was fair or poor were report having a physical problem that affected their daily lives.

Figure 2 - Self-Rated Health Status by Age Group (2001-2013 KIHAS, Combined)

PERCENT OF RESPONDENTS

“In general, would you say that your overall state of health these days is excellent, very good, good, fair, or poor?”

60

53

49

50 40 30 20

44 33

31 16

36

32

32 32

24

18

10 0 18-29 (N=4,101) 30-44 (N=5,262) 45-59 (N=5,639) "Excellent" or "Very Good"

"Good"

60+ (N=4,365)

"Fair" or "Poor"

It is interesting to note that people’s assessments of their overall state of physi be based on different considerations at different ages. Looking just at the onerespondents who rated their current state of health as fair or poor, Figure 3 sho 6

Kinder Institute for Urban Research

from engaging in normal activities. Fewer than a tenth (8 percent) of the younger respondents who

63

30

80

62

69

PERCENT OF RESPONDENTS

70

PERCENT OF RESPONDENTS

they assessments were in generally poor health also indicated that they were currently dealing with a It is interesting to note that said people’s of their from engaging in normal Fewerconsistently than a tenthwith (8 percent) of the you debilitating problem, number for whom this isactivities. true increased age overall state of physical health seem tohealth be based on but thesaid they in generally poor health also indicated that they were currentl to reach 69 percent among those aged 60 andwere older. different considerations at different ages. Looking just debilitating health problem, but the number for whom this is true increased c at the one-fifth of all the respondents who rated their to reach 69 percent among those aged 60 and older.

Figure 3 - TheAge Prevalence of Serious Health ProbFigure Fewer 3 –The than Prevalence Health Problems Group among from engaging in normal activities. a tenthof(8Serious percent) of the youngerbyrespondents who the Respondents Who current state of health as fair or poor, Figure 3 shows lems by Age Group among the Respondents Who Reported Poor Health said they were in generally poor health Fair alsoorindicated that(HAHS) they were currently dealing with a from engaging in normal activities.

Fewer than a tenth (8 percent) of the younger respondents who the Re the percent in each of the four age groups who also said Figure 3 –The Prevalence of Serious Health Problems by Age Group among Reported Fair orwith PoorageHealth (HAHS) debilitating health problem, but the number for whom this is true increased consistently said they were in generally poor health also indicated that they were currently dealing with a Reported Fair or Poor Health (HAHS) they had69a physical healththose problem that them to reach percent among aged andprevented older.

10060 92 health problem, but the number for whom this is “Do debilitating trueyou increased consistently with age have any physical health from engaging in normal activities.

Fewer than a tenth to90reach 69 percent among those aged 60 and older. problems right now that prevent 100 92 (8 percent) the younger respondents who said they Figure 3 –Theof Prevalence of Serious Health you from doing any of the things “Do you hav 80 Problems by Age Group among the Respondents Who 90 Reported Fair or Poor Health (HAHS) 69 people your age normally can do?” problems rig were in generally poor health also indicated that they Figure 3 –The Prevalence of Serious Health Problems by Age Group among the Respondents Who 70 63 62 [N=205] you from do were currently dealing with a debilitating health probReported Fair or Poor Health (HAHS) 80 69 60 people your 100 92 70 have any physical health lem, but the number for whom this is true increased 63 “Do you 62 [N=205] 50 90 100 among consistently with age to reach 69 percent 6038 right now that prevent 92 those 37 problems “Do you have any physical health 40 you from doing any of 31 the things aged 6080and older.

90 50 problems right now that prevent people your age normally can do?”

38 doing any of the things 37 you from 40 [N=205] 31 can do?”

69 Nopeople seriousyour health ageproblems normally 30 63 62 [N=205]

0

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

20 As other are 60 research has shown,6 younger 70people 8 10 health status on more likely to evaluate their current 50 At least one serious health problem 60 No seriou 20 38 37 0 the basis of their participation in risky behaviors (e.g., 8 40 50 10 31 At least o 18-29 60-95 smoking, contrast, older30-44 37 45-5938 30 poor diet, lack of exercise). In 40 0 31 adults 20 overwhelmingly refer to functional limitations 18-29 problems 30-44 45-59 60-95 No serious health 30 As other has their shown,6 younger people are more likely to evaluate their current health and specific physical whenresearch assessing 8 problems 10 status on the20basis of their participationAt inleast riskyone behaviors (e.g.,problem smoking, poor diet, health lack of No serious problems serious health 6 state of health.

8 adults As younger people are likely to evaluate the other researchrefer has shown, exercise). In contrast, older overwhelmingly to functional limitations andmore specific 10

serious health on the of their participation At in least riskyone behaviors (e.g.,problem smoking, po problems their state of basis health.

45-59 when assessing 60-95 status 0 overwhelmingly Health Disparities 5 exercise). In contrast, older adults refer to functional limitat 18-29 30-44 45-59 when assessing 60-95 their state of health. physical problems 6 younger people are more likelyInequality to evaluate their As other research has shown, The Relationship between andcurrent Healthhealth status on the basis of their participation in risky behaviors (e.g., smoking, poor diet, lack of 6 younger people and are more likely to evaluate their current health As other researchrefer has to shown, exercise). In contrast, older adults functional limitations specific The Relationship between Inequality and(BRFSS) Health Tableoverwhelmingly 1 shows the data from the 2013 Behavioral Risk Factor Surveillance System 18-29

physical 30-44

The Relationship between Inequality and Health

The GINI coefficients listed in the right-hand column of Table 1 provide a standard measure of income inequality within each of these urban counties: If all households in the region had exactly the same level of income, the GINI coefficient would be zero; a value of one would mean that a single household had all the income, and all others had no income at all. The localities with a high GINI coefficient are those with a disproportionate number of rich and poor residents; lower GINI coefficients reflect populations with more even income distributions. The table shows clearly that the differences in self-rated health across these metropolitan areas tend to mirror their differences in income inequality.

Table 1 shows the data from the 2013 Behavioral Risk Factor Surveillance System (BRFSS) presenting the proportion of residents in the nation’s ten most populous urban counties (not including the New York boroughs) who indicated that they were in only fair or poor health.7 When compared to the other large urban regions, Harris County ranks in the middle, with 18 percent of Houston area residents in this study reporting fair or poor health, compared to 23 percent in Los Angeles and 14 percent in San José.

Table 1 – Self-Rated Health Status across Major Metropolitan Counties (2013 BRFSS, Excluding NYC Table 1 - Self-Rated Health Status across Major Metropolitan Counties (2013 PRFSS, Excluding NYC Boroughs), Ranked by Their GINI Coefficients (2011 American Community Survey 5-Year Estimates) Boroughs, byHealth TheirStatus GINIacross Coefficients (2011 American Community 5-Year Table 1 –Ranked Self-Rated Major Metropolitan Counties (2013 BRFSS, Survey Excluding NYC Estimates) Boroughs), Ranked by Their GINI Coefficients (2011 American Community Survey 5-Year Estimates)

Major U.S. City

Philadelphia, PA

County Percent in Fair or GINI Coefficient (2007-2011 ACS Population Poor Health (2013 five-year County County Percent in Fair or GINI Coefficient (2007-2011 ACSestimates) BRFSS) Population Poor Health (2013 five-year estimates) BRFSS) Philadelphia 1,536,471 20 0.497

Los Angeles, CA

Los Angeles

Dallas, TXTX Dallas,

Dallas Dallas

Houston, TXTX Houston,

Harris Harris

Chicago, Chicago, IL IL

Cook Cook

San Antonio, TX

Bexar

SanPhoenix, Diego, AZ CA

San Diego Maricopa

Major U.S. City

Philadelphia, PA Los Angeles, CA

San Antonio, TX San Diego, CA

Phoenix, AZCA San José, San José, CA

County

Philadelphia

1,536,471

20

Los Angeles

9,889,056

23

9,889,056

2,416,014 2,416,014

Bexar

17

17

4,180,894 4,180,894

18

5,217,080 5,217,080

17

1,756,153

17

3,140,069

16

3,880,244

15

1,756,153

San Diego

3,140,069

Maricopa Santa Clara

3,880,244 1,809,378

Santa Clara

14

1,809,378

0.497

23

0.494

0.494

0.493

0.493

18

0.491

0.491

17

0.491

0.491

0.466

17

0.455

16

0.454

15

0.452

14

0.466 0.455 0.454 0.452

Figure 4 – The Percent of Residents in Fair or Poor Health, by the GINI Coefficients of Their Counties Figure 4 -BRFSS) The Percent of Residents in Fair or Poor Health, by the GINI Coefficients of Their Counties (2013

FigureBRFSS) 4 – The Percent of Residents in Fair or Poor Health, by the GINI Coefficients of Their Counties (2013 (2013 BRFSS) 24

LA

24

PA

20 18

SA

2016 1814

SD SJ

12

PX

SA SD

16

10

LA

LEGEND SJ—San José (Santa Clara Co.) HO PX—Phoenix (Maricopa Co.) PA SD—San Diego (San Diego Co.) CH DA LEGEND SA—San Antonio (Bexar Co.) SJ—San CH—Chicago (Cook Co.) José (Santa Clara Co.) HO DA—Dallas (Dallas Co.) PX—Phoenix (Maricopa Co.) HO—Houston (Harris Co.) SD—San Diego (San Diego Co.) PA—Philadelphia (Philadelphia Co.) CH DA Antonio LA—Los AngelesSA—San (Los Angeles Co.) (Bexar Co.)

22

OR FAIR HEALTH

PERCENT OF RESIDENTS IN POOR OR FAIR HEALTH PERCENT OF RESIDENTS IN POOR

22

CH—Chicago (Cook Co.) DA—Dallas (Dallas Co.) SJ County GINI Coefficient HO—Houston (Harris Co.) 12 PA—Philadelphia (Philadelphia Co.) 6 Kinder Institute for Urban Research LA—Los Figure 4 illustrates the positive linear relationship between the levels of income inequality in eachAngeles (Los Angeles Co.) 10 metro area and the proportion of residents reporting that they were in fair or poor health. The 0.44 0.46 are lower 0.47 than they 0.48 0.49 (e.g., San0.5 localities where the 0.45 levels of inequality are in Houston José, 14 0.44

0.45

PX

0.46

0.47

0.48

0.49

0.5

Figure 4 illustrates the positive linear relationship between the levels of income inequality in each metro area and the proportion of residents reporting that they were in fair or poor health. The localities where the levels of inequality are lower than they are in Houston (e.g., San JosĂŠ, Phoenix) tend to have healthier populations, whereas those with higher levels of inequality (Philadelphia, Los Angeles) have more residents reporting that their health was only fair or poor. Lopez (2004) examined metropolitan areas across all 50 states and found that, with each 0.1 point rise in the GINI index, the proportion of individuals in the area who reported fair or poor health increased by 4 percent.8 Socioeconomic factors are strongly linked to health outcomes.

The Individual Contributions of Education, Income, and Ethnicity The data in Table 1 clearly suggest that income differences contribute importantly to the health disparities that exist across the country. Higher levels of socioeconomic status (SES) generally mean greater household wealth, healthier neighborhood conditions, and more support systems that encourage health-promoting activities.

Each of the separate components that define the structural inequalities in America (education, income, and ethnicity) may be significant individual contributors to the relationship between SES and self-reported health, or they may be associated with health outcomes simply because of their correlation with other more powerful structural variables. Thus, African Americans and Latinos may have worse health outcomes than Anglos or Asians because they also have lower levels of education and income and not because of any distinctive factors (such as wealth disparities, stress, or discrimination) that are specific to ethnic differences. Or does ethnicity independently contribute to the health disparities, even after the separate effects of education and income are statistically controlled?

We will make use of logistic regression to disentangle the degree to which each of these structural components is individually associated with self-reported health. First, we examine the straightforward correlations of health status with education, income, and ethnicity, and consider the reasons why each factor might indeed be a separate and cumulative predictor of health status. Then, with the use of logistic regression, we will be able to test the independent correlation of each separate factor with self-rated health, after controlling for the impact of all the others.

Health Disparities

7

The Role of Education. A person’s educational attainment can influence health in a variety of ways. Several national studies suggest that more education in and of itself, even when controlling for its close association with income, enhances life expectancy and fosters better health.9, 10 Educational attainment, at all levels of income, increases access to basic health information and the likelihood of engaging in healthy behaviors, such as regular physical exercise and scheduling check-ups on a regular basis.11 The 2003 National Assessment of Adult Literacy found that education is highly correlated with measures of “health literacy,” the ability to obtain, process, and understand basic health information and services in order to make appropriate health decisions.12

Making use of the rich data from 13 years of the KIHAS (2001-2013), Figure 5 reveals a powerful linear relationship between educational attainment and self-reported health status. Over one-third (35 percent) of area residents who had less than a high school education said their current state of health was fair or poor, but this was the case for just 11 percent of the respondents with college degrees.

Fully 63 percent of the survey participants who had post-graduate educations reported their health as excellent or very good, compared to just 31 percent of those who had dropped out of high school.

Clearly, as education levels increase, health status improves. We will explore below whether this relationship continues to hold even after rigorously controlling for the separate effects of income, ethnicity, gender, and age.

Education is also associated with important psychological factors, such as the sense of efficacy, or the belief in personal control over one’s life circumstances: The extent to which individuals are confident that they can influence what happens in their lives has been shown to be linked with higher levels of self-rated health.13 Educational attainment, even beyond its relationship with income, may improve health by reducing overall stress and promoting healthier lifestyle choices. below whether this relationship continues to hold even after rigorously controlling for the separate effects of income, ethnicity, gender, and age.

below whether this relationship continues to hold even after rigorously controlling for the separate Figure 5of– income, Self-Rated Health Status

by Educational (2001-2013 KIHAS, Combined) effects ethnicity, gender, age. Attainment Figure 5 - Self-Rated Health Statusand

by Educational Attainment (2001-2013 KIHAS, Combined)

PERCENTPERCENT OF RESPONDENTS OF RESPONDENTS

705 – Self-Rated Health Status by Educational Attainment (2001-2013 KIHAS, Combined) Figure 63 60 70 50 60 40 50 30 40 20 30 10 20 0 10

61

49 35 31 33 35 31 33

49

34 42 34

63

61

42 32

28

24 32 18 24 18

28

27

11

27

10

"Excellent" or "Very Good" "Good"

"Excellent" or "Very Good" "Fair" or "Poor" 11 10 "Good" Less than HS HS diploma Some college College Postgraduate "Fair" or "Poor" (N=2,411) (N=4,664) (N=5,954) degree education 0 (N=4,484) (N=2,729) Less than HS HS diploma Some college College Postgraduate "What is the highest grade of school or year of college completed?"

(N=2,411) (N=4,664) (N=5,954) degreethat you've education (N=4,484) (N=2,729)

The Importance Income. directorimpact of income on health seems likely to be at least as "What is theof highest gradeThe of school year of college that you've completed?"

Institute for Urban Research 8 Kinder powerful as the relationship with education. People with higher household incomes typically live longer, healthier lives than less affluent individuals, since their higher earnings are associated with The Importance of Income. The direct impact of income on health seems likely to be at least as access to resources that influence health and mortality14,15 and foster healthier lifestyle choices.16 powerful as the relationship with education. People with higher household incomes typically live

The Importance of Income. The direct impact of income on health seems likely to be at least as powerful as the relationship with education. People with higher household incomes typically live longer, healthier lives than less affluent individuals, since their higher earnings are associated with access to resources that influence health and mortality,14,15 and foster healthier lifestyle choices.16 Higher-paying jobs are more likely to have health-related employee benefits and employer-sponsored health insurance.

Beyond their greater access to quality health care, more affluent individuals are likely to experience less stress and to live in neighborhoods with higher levels of social cohesion, greater access to nutritious foods and ample opportunities for physical activity and recreation. In this era of growing inequalities, income differences in health and well-being are likely to be particularly stark.

Figure 6 shows the relationship between selfreported health status and household incomes in the KIHAS surveys. Of all area residents who reported household incomes of $15,000 or less (below the poverty level for a family of two), considerably more than a third (38 percent) said they were in fair or

poor health. As income levels rise above $25,000, the percent of respondents reporting poor health declines, while the number saying that their current physical health is excellent or very good increases in a stepwise fashion. At the highest levels (the households reporting incomes of more than $75,000), only 10 percent of the survey participants said their current state of health was fair or poor, and 63 percent reported that they were in very good or excellent health.i Again, we need to ask whether this relationship is due specifically to differences in household income: Does the connection between income and health exist because respondents with higher incomes are also likely to have higher levels of educational attainment, or are education and income separate and independent contributors to a person’s state of health? Meanwhile, of course, both education and income are closely tied to differences in ethnic backgrounds, so we also need to consider the role of ethnicity as we seek to identify the specific social structural forces that independently account for health disparities among Houston-area residents.

Figure 6 – Self-Rated Health Status by Annual Household Income (2001-2013 KIHAS, Combined) Figure 6 - Self-Rated Health Status by Annual Household Income (2001-2013 KIHAS, Combined) 70

63

PERCENT OF RESPONDENTS

60

54 48

50 40 30

38 32 30

34

31

35

38 37

32 25

20

31 20

27 15

10 0

Less than $15,000 (N=1,359)

$15,001 to $25,000 (N=1,975)

$25,001 to $35,000 (N=2,066)

$35,001 to $50,000 (N=2,210)

$50,001 to $75,000 (N=2,633)

10

"Excellent" or "Very Good" "Good" "Fair" or "Poor"

More than $75,000 (N=4,282)

"Please stop me when I reach the category that includes your total household income in the past year; that is, the income for all members of the household during the past year." i Again, we30need to of ask thisinrelationship due specifically to differences in household Almost percent thewhether respondents the 2001-2013isKIHAS declined to report their household incomes. There were income: Does differences the connection incomeand andthehealth because respondents with higher no significant betweenbetween these individuals rest of exist the respondents in their distributions on education, ethnicityare or age. incomes also likely to have higher levels of educational attainment, or are education and income separate and independent contributors to a person’s state of health? Meanwhile, of course,

Disparities both education and income are closely tied to differences in ethnic backgrounds, soHealth we also need 9 to consider the role of ethnicity as we seek to identify the specific social structural forces that independently account for health disparities among Houston-area residents.

The Impact of Ethnicity. The recent unprecedented immigration streams into this country since the 1965 reform of the previously restrictive laws have transformed the ethnic composition of the Houston and American populations. Throughout almost all of its history, Houston was essentially a bi-racial, Anglo-dominated Southern city. In the course of the past 30 years, after the oil-boom collapse in 1982, it has been transformed into the single most ethnically diverse large metropolitan region in the country. In 2010 the U.S. Census counted 4.1 million people living in Harris County, of whom just 33 percent were Anglos (or non-Hispanic whites). Harris County’s population was now 41 percent Hispanic, 18 percent African-American, and 8 percent Asian or other.

closely associated with self-reported health. Well over half (55 percent) of the U.S.-born Anglos said their current state of health these days was excellent or very good; this was the case for 48 percent of the Asians, 46 percent of the U.S.-born Hispanics, 41 percent of the U.S.-born blacks, and just 32 percent of the Hispanic immigrants.

Fully 30 percent of the Hispanic immigrants and 28 percent of the U.S.-born blacks said they were currently in fair or poor health. The U.S.-born Hispanics generally reported better health than blacks or Latino immigrants, but worse health than Anglos or Asians.

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

PERCENT OF RESPONDENTS

Figure 7 - Levels of Household Income in Five EthFigure 7 utilizes the combined data from the KIHAS nic 7Communities (2001-2013 Combined) (2001-2013) to show the distributions of household Figure – Levels of Household Income in Five KIHAS, Ethnic Communities (2001-2013 KIHAS, Combined) income across HarrisFigure County’s five major ethnic 60 7 – Levels of Household Income in Five Ethnic Communities (2001-2013 KIHAS, Combined) 52 49 49 groups, with Hispanics divided into U.S.-born and 50 45 60 43 43 KIHAS, 43 Combined) Figure 7 – Levels of Household Income in Five Ethnic Communities (2001-2013 52 immigrant communities. Among the U.S.-born An49 38 49 40 35 60 50 glo respondents, 43 percent reported total household 45 – Levels of Household Income in Five Ethnic Communities (2001-2013 KIHAS, Combined) 43 43 52 43 27 30 49 incomes of more than $75,000; this was true for 38 38 49 20 50 45 4043 35 19 Less than $25,000 43 43 16 20 52 for just 20 percent of the percent of the Asians, but 14 49 49 38 $25,001 to $75,000 27 7 40 Hispanics, 16 percent U.S.-born of U.S.-born blacks, 45 35 30 10 43 43 43 and 7 percent of the Hispanic immigrants.

Fully 35 20 More than $75,000 38 19 0 27 30 Less than $25,000 35 16 20 percent of the African-American 14 respondents and US-Born Anglos All Asians US-Born US-Born Blacks Hispanic 20 19 (N=4,392) (N=893) Hispanics (N=4,201) Immigrants 27 45 percent the Hispanic immigrants reported Less than $25,000 $25,001 to $75,000 7 16 20 of14 (N=2,516) (N=1,833) 10 20 total household incomes of less than $25,000 (which 19 $25,001 associated to $75,000 More than $75,000 Less than $25,000 7 disadvantages The social and economic with being black or Latino will inevitably 0 poverty16 14 would10 put them just above the level for a generate differences in health status. Moreover, the ethnic disparities in life circumstances have US-Born Anglos US-Born US-Born Blacks Hispanic More than $75,000 shown to to accumulate over the life course, and they are typically transferred across $75,000 family of Anglos and All 19Asians 7 been $25,001 0 four); only 14 percent of(N=4,392) (N=893) generations, Hispanics Immigrants impacting (N=4,201) the education and health outcomes of children and grandchildren as well.17 percent ofUS-Born AsiansAnglos had incomes below $25,000.

All Asians US-Born US-Born Blacks Hispanic (N=2,516) (N=1,833) Such ethnic8 differences may wellin beSelf-Rated strong predictors of healthStatus disparities in and of themselves, More than $75,000 Figure - Differences Health (N=4,392) (N=893) Hispanics (N=4,201) even beyondImmigrants their close relationship to education and income. in Five Ethnic Communities (2001-2013 KIHAS, US-Born Anglos All Asians US-Born US-Born Blacks Hispanic Figure (N=2,516) (N=1,833) 8 – Differences in Self-Rated HealthorStatus in Five Ethnic Communities (2001-2013 KIHAS, The social and economic disadvantages associated The social and economic disadvantages associated with being black Latino will inevitably Combined) (N=4,392) (N=893) Hispanics (N=4,201) Immigrants Figure 8 presents the racial and ethnic differences in self-rated health status among the KIHAS Combined) generate health status.

Moreover, the ethnic disparities in life circumstances havewith self-reported with being black or(N=2,516) Latino willdifferences inevitablyingenerate respondents. make will it clear that ethnicity is indeed closely associated (N=1,833) The social and economic associated with blackThe ordata Latino inevitably 60 55and beendisadvantages shown to accumulate over thebeing life course, are typically transferred health.

Well over halfthey (55 percent) of the U.S.-born Anglos across said their current state of health these differences in health status.

Moreover, the ethnic generate differences in health status.

Moreover, ethnicdays disparities in life circumstances washealth excellent or very good; was thehave case for 48 percent of the 46 percent of the 48Five generations, impacting the the education and outcomes of this children and grandchildren asAsians, well.17 al and economic disadvantages associated with being or Latino will disparities intolife circumstances have been 50inevitably Figure 8black –shown Differences in Self-Rated Health Status in Ethnic46 Communities (2001-2013 KIHAS, been shown accumulate over the life course, and they are typically transferred U.S.-born Hispanics, 41 percentacross of the U.S.-born blacks, and just 32 percent of the Hispanic Such ethnic differences mayinwell be strong predictors ofCommunities health disparities in 41 and of themselves, Combined) differencesto inaccumulate health status.

Moreover, the8 –ethnic disparities life circumstances over the life course, and they are immigrants.

Fully have 30 percent of and 28 percent37of the U.S.-born blacks Figure Differences in Self-Rated Health Ethnic (2001-2013 generations, impacting the education and health outcomes ofStatus children grandchildren as immigrants well.17KIHAS, 37the Hispanic 40 in Fiveand even beyond their close relationship to education and income.

33 said they were currently in fair or poor health.

The U.S.-born 32 generally reported better 60 own to accumulate the lifeincourse, and they are typically transferred across Combined) 55 31 Hispanics typically transferred across generations, impacting 30 Figure ethnic 8 over – Differences Self-Rated Health in Five Ethnic Communities (2001-2013 KIHAS, Such differences may well beStatus strong predictors of health health disparities in17and of themselves, 29or Latino 28 Anglos or Asians. than blacks immigrants, but worse health than 30 ons, impacting the education health outcomes of childrenand and income. grandchildren as well.

60 48 Combined) the education andand health outcomes children 55of even beyond their close relationship to education and 46 50 21 Figure 8 presents the racial and ethnic in self-rated health status among the KIHAS nic differences may be strong in anddifferences of themselves, 17predictors of health disparities 17 9 shows the importance country of origin in accounting for the variability in self-rated 41of15 60 well 48 may Figure 20 grandchildren 55 as well. Such50ethnic differences 46 37 respondents.

The data make it clear that ethnicity is indeed closely associated with self-reported 37 40 yond their close relationship education and income. health status among Houston’s varied Asian-American communities.

A clear majority of the or "Ver "Excellent" Figure 8 presentstothe racial ethnic differences in self-rated health among theBangladesh KIHAS 33 41status well be health disparities inpercent) and 48 ofand 32 31 46over 50 strong predictors immigrants from India, Pakistan, and (61 percent) from the Philippines (52 10 health.

Well half (55 of the U.S.-born Anglos said their current state ofandhealth these 30 37 28 37 29 40 "Good" respondents.

The data make their it clear that relationship ethnicity is indeed closely associated with 30 41 33 the percent), along with the U.S.-born Asians (53 percent), said that their current state of health was of themselves, even beyond close 32 self-reported days was in excellent or very good; this case 48 percent 31for 30 of the Asians, 46 percent of the "Fair" or "Poor" presents thehealth. racial and ethnic differences self-rated health status among KIHAS 37 was 37 percent) 21good.

29 0 the 28 40 Well either excellent or very This was true forthese just 39 percent of the immigrants from China, over half (55 of the U.S.-born Anglos said their current state of health 30 Hispanics, to education andthat income.

33 closely U.S.-born U.S.-born blacks, just 32 of theimmigrants.

Hispanic 32 the 17 of 2041 percent ents. The data make itexcellent clear ethnicity is indeed with self-reported US-Born Alland Asians US-Born US-Born Hispanic 31associated 15 Taiwan, andofHong Kong, and 34 percent of percent the Vietnamese These contrasts in self30 21 days was or very good; this was the case the Asians, 46 percent of the 29 28 for 48 percent Anglos (N=1,315) Hispanics Blacks Immigrants 30 immigrants.

Fully 30 percent of the Hispanic immigrants and 28 percent of the U.S.-born blacks "Excellent" or "VeryinGood" reported health closely parallel the differences among the Asian communities their overall 17 their current Well over half (55 percent) of the U.S.-born Anglos said state of health these 20 15 U.S.-born Hispanics, said 41 percent of the blacks, andhealth. just 32 percent of theand Hispanic (N=5,916) (N=3,572) (N=5,818) (N=2,659) 10 21U.S.-born levels of income andU.S.-born education, hence in"Excellent" thegenerally life circumstances that they experience.ii they were currently in fair or poor The Hispanics reported better Figure 8 presents the racial and ethnic differences in "Good" or "Very Good" s excellent or very good; this was the case for 48 percent of the Asians, 46 percent of the 17 20 15 10 immigrants.

Fully 30 health percent of the Hispanic immigrants and 28 percent of the U.S.-born blacks than blacks orjust immigrants, but worse health than Anglos or Asians."Fair" or "Poor" self-rated health status among the KIHAS respon0Latino n Hispanics, 41 they percent of currently the U.S.-born blacks, and 32 percent of theHispanics Hispanic "Good" "Excellent" or "Very Good" said were in fair or poor health.

The U.S.-born generally reported better 10 US-Born All Asians US-Born US-Born Hispanic Figure 9 – Differences in Self-Rated Health "Fair" Statusoramong U.S.-born Asians and in Houston’s Four L dents.

The it clear that ethnicity indeed nts. Fully 30 percent ofdata the make Hispanic immigrants andbut 28ispercent of the U.S.-born blacks 0 "Poor" "Good" health than blacks or Figure Latino worse health than Anglos orinAsians.

Anglos (N=1,315) Hispanics Blacks Asian Immigrant Communities (2001-2013 KIHAS, Combined)in self-rated 9immigrants, shows the importance of country of origin accounting forImmigrants the variability US-Born All Asians US-Born US-Born Hispanic y were currently in0 fair or poor health.

The U.S.-born Hispanics generally reported better "Fair" or "Poor" (N=5,818) (N=5,916) (N=3,572) (N=2,659) Anglos (N=1,315) Hispanics Blacks Immigrants A comprehensive report on Harris County’s Asian communities released inof 2012. health status among varied Asian-American communities.

A clear was majority theIt presents the findings from an blacks or10 Latino but worse health thanHouston’s Anglos or Asians.

US-Born AllUrban Asians US-Born US-Born Hispanic Kinder Institute for Research Figure 9 immigrants, shows the importance of country of origin in accounting for the variability in self-rated three focused surveys conducted by the Kinder Institute in 1995, 2002, and 2011, reaching more than 500 Asians in 70 (N=5,916) (N=3,572) (N=5,818) (N=2,659) immigrants from India, Blacks Pakistan,Immigrants and Bangladesh (61with percent) and from theconducted Philippines (52 Anglos (N=1,315) Hispanics each of these three years, 24 percent of the in Vietnamese, Cantonese, Mandarin, and 61 A health status among Houston’s varied Asian-American communities. clearsaid majority ofinterviews thecurrent (N=5,916) (N=3,572) (N=2,659) Korean.60The report is available http://kinder.rice.edu/reports/. percent), along with U.S.-born (53 percent), that their state of healthFour was Figure 9the –(N=5,818) Differences in Asians Self-Rated Health Status at among U.S.-born Asians and in Houston’s Largest shows the importance of country of origin in accounting for the variability in self-rated 53 immigrants from India, Pakistan, andAsian Bangladesh (61 percent) andforfrom the Philippines (52immigrants from China, 52 Immigrant Communities (2001-2013 KIHAS, Combined) either excellent or very good.A This was true just 39 percent ofand the Figure 9 – Differences in Self-Rated Health Status among U.S.-born Asians in Houston’s Four Largest atus among percent), Houston’s varied Asian-American communities. clear majority of the 47 along with the U.S.-born Asians (53 percent), said their current state of health was 50 12 that Kinder Institute for Urban Research ii

Figure 8 – Differences in Self-Rated Health Status in Five Ethnic Communities (2001-2013 KIHAS, Combined) 60

55 48

PERCENT OF RESPONDENTS

Figure509 shows the importance of country46of origin Many independent studies confirm that Anglos in accounting for the variability in self-rated health 41 generally experience the best health and the low37 status40 among Houston’s varied37 Asian-American est age-adjusted mortality over the life course,18 33 32 31 30 Hispanics have much higher whereas communities. A29 clear majority of the immigrants 28 blacks and 30 levels of infant mortality and obesity.19 Are there from India, Pakistan, and Bangladesh (61 percent) 21 and from (52 percent), along with forces beyond differences in socioeconomic 17 20 the Philippines 15 the U.S.-born Asians (53 percent), said that their conditions that account for"Excellent" the racial or and ethnic "Very Good" current disparities in self-reported health status? The 10state of health was either excellent or very "Good"and Latinos may daily lived experiences of blacks good. This was true for just 39 percent of the immiwell differ from those of Anglos for"Poor" reasons havgrants from China, Taiwan, and Hong Kong, and 0 "Fair" or 34 percent of US-Born the Vietnamese immigrants.US-Born These ing to do not only with the well-known contrasts All Asians US-Born Hispanic Anglos (N=1,315) Hispanics Blacks Immigrants contrasts in self-reported health closely parallel the in education and income, but also with the vast (N=5,916) (N=3,572) (N=5,818) (N=2,659) differences among the Asian communities in their discrepancies in family assets, the experience of overall levels of income and education, and hence in persistent discrimination, and the continual high the life circumstances that they experience.ii levels of stress that result from these realities.

Figure 9 – Differences in Self-Rated Health Status among U.S.-born Asians and in Houston’s Four Largest FigureImmigrant 9 - Differences in Self-Rated Health Status U.S.-born Asians and in Houston’s Four Asian Communities (2001-2013 KIHAS,among Combined) Largest Asian Immigrant Communities (2001-2013 KIHAS, Combined) 70 60

61 53

52 47

PERCENT OF RESPONDENTS

50 40 30

34

29

25

24 24

20 10

41

39

35

10

15

12

"Excellent" or "Very Good" "Good" "Fair" or "Poor"

0 Indians / Pakistanis (N=292)

US-born Asians (N=303)

Filipinos (N=62)

Chinese / Taiwanese (N=254)

Vietnamese (N=226)

Many independent studies confirm that Anglos generally experience the best health and the lowest age-adjusted mortality over the life course,18 whereas blacks and Hispanics have much higher levels of infant mortality and obesity.19 Are there forces beyond differences in socioeconomic conditions that account for the racial and ethnic disparities in self-reported health status? The daily lived experiences of blacks and Latinos may well differ from those of Anglos for reasons having to do not only with the well-known contrasts in education and income, but also with the vast discrepancies in family assets, the experience of persistent discrimination, and the continual high levels of stress that result from these realities.

A comprehensive report on Harris County’s Asian communities was released in 2012. It presents the findings from three focused surveys conducted by the Kinder Institute in 1995, 2002, and 2011, reaching more than 500 Asians in each of these three years, with 24 percent of the interviews conducted in Vietnamese, Cantonese, Mandarin, andSurvey Korean. The The SHEA Health 13 report is available at http://kinder.rice.edu/reports/. ii

Health Disparities

11

The Structural Correlates of Self-Rated Health

Table 2 presents the results of the logistic regression, measuring the separate associations of these distinct structural variables with the odds that Harris County The Structural Correlates of Self-Rated Health residents will indicate that their current state of health is “fair” or “poor.”

The rich data from the KIHAS make possible a The rich data from the KIHAS make possible a rigorous statistical examination of the individual rigorous statistical examination of the individual impact on self-reported health of each of these major structural factors, after taking account of the impact on self-reported health of each of these major The regression coefficients in the first column of the effects of all the others. To test these relationships, we use the unweighted from the 13 years table indicate the estimated increasedata (or decrease) structural factors, after taking account of the effects of pooled random samples of Harris County residents.

Table 2 presents the results of the in the log odds of reporting fair or poor health.

Forlogistic of all the others. To test these relationships, we use regression, measuring the separate of odds theserepresent distinct the structural with the odds age, the increasevariables in the proporthe unweighted data from the 13 years of pooled associations that Harris County residents will indicate that their current state of health is “fair” or tion reporting poor health with each additional“poor.” year; random samples of Harris County residents.

Table 2 – Logistic Regression of the Structural Correlates on Self-Rated Health Status (2001-2013 KIHAS, Table 2 - Logistic Regression of the Structural Correlates on Self-Rated Health Status (2001-2013 KIHAS, iii Combined) Combined)iii Self-Reported “Fair” or “Poor” Health

Independent Variable Coefficient (A) Socioeconomic Status a (1) Education High School Diploma -.189 Some College -.366** College Degree -.863*** b (2) Income $35,501 to $50,000 -.462*** $50,001 to $75,000 -.631*** More than $75,000 -.958*** (B) Demographic Characteristics c (3) Ethnicity Blacks +.423*** Latinos +.308*** Asians +.269 (4) Age +.021*** (5) Female -.047 ----------------------Constant -.685*** N 6,792 a Reference group: Less than high school. b Reference group: Less than $35,000. c Reference group: Anglos. *p < .05, **p < .01. ***p < .001 (two-tailed tests).

(KIHAS 2001-2013) Standard Odds Ratio Error (OR)

1-OR

.105 .105 .116

.828 .694 .422

-.172 -.306 -.578

.096 .095 .093

.630 .532 .384

-.370 -.468 -.616

.083 .086 .219 .002 .064

1.527 1.361 1.309 1.022 .954

+.527 +.361 +.309 +.022 -.046

.111

.504

-.496

To determine if there were ethnic differences in the relationships with health of the structural variables, we tested the interaction terms and found that none were statistically significant. We also ran the regressions separately for the years 2001-2007 and 2008-2013; we found no differences between the two models in the nature and strength of the coefficients iii for age, gender, andif ethnicity, and only differences very slight differences for education. haveoftherefore combined all 13 we tested To income, determine there were ethnic in the relationships with We health the structural variables, years in thethe regression models. interaction terms and found that none were statistically significant. We also ran the regressions separately for the years 2001-2007 and 2008-2013; we found no differences between the two models in the nature and strength of Institute for Urban Research 12 Kinder the coefficients for age, gender, income, and ethnicity, and only very slight differences for education. We have therefore combined all 13 years in the regression models. iii

for all the other independent variables in the model the coefficient represents the increase or decrease in the odds of reporting poor health compared to the respondents in the designated reference category.

Thus, after controlling for the effects of income, age, ethnicity, and gender, the regression model shows that the odds of reporting poor health decrease at each higher level of education (high school diploma, some college, college degree), when respondents with that amount of education are compared to those who have less than a high school education (the referent category). The results indicate that it is only when the levels of schooling reach beyond high school that educational attainment is significantly associated with better health.

Just having a high school diploma offers a statistically indistinguishable health advantage over having less than a high school education. In this increasingly global, high-tech, knowledge-based economy, education beyond high school is now a critical determinant of a person’s ability to obtain a job that pays more than poverty wages. This new reality is increasingly clear to the general public as well. In the 2013 KIHAS, 73 percent of all Harris County residents agreed with the suggestion that, to succeed in today’s world, “it is necessary to get an education beyond high school.” Fewer than one in four believed instead that “there are many ways to succeed with no more than a high school diploma.”20 The table indicates that respondents with a college degree have an Odds Ratio (OR) of .422 of reporting their health as only fair or poor. This means that they are 42 percent as likely as the respondents without a high school education, or 58 percent less likely (1OR) than those in the reference category, to report being in fair or poor health. And, again, this relationship holds even after rigorously controlling for the effects of differences among the respondents in their household incomes, ethnic backgrounds, age, and gender. In sum, educational attainment, at least when it entails education beyond high school, does indeed appear to be an independent factor that in and of itself is significantly associated with a person’s self-reported health status.

Similarly, each increase in the respondents’ levels of household income is significantly associated with decreasing likelihoods that they will report being in fair or poor health. Respondents with incomes between $35,000 and $50,000 had 37 percent lower odds (1-OR) of reporting poor health compared to those in households making less than $35,000.

Respondents with incomes above $75,000 (after controlling for differences in education, ethnicity, and age) were almost 62 percent less likely (1-OR) to report being in poor or fair health. Income, like education, is indeed a separate and significant independent correlate of self-reported health status.

Table 2 indicates further this is also the case for ethnicity. Controlling for differences on all the other variables in the model, the African-American respondents were 53 percent more likely than Anglos (the Latinos were 36 percent more likely) to report being in fair or poor health. With all the other variables controlled, Asians are only slightly, and not significantly, more inclined than Anglos to report being in fair or poor health. For reasons other than, or in addition to, the ethnic differences in socioeconomic status, the black and Latino residents of Harris County are significantly more likely than Anglos to give low evaluations of their current state of health.

Age, too, remains a significant predictor, with the other structural variables controlled: each one-year increase in the respondents’ ages, from 18 to 95, raises the odds of their reporting that they are in fair or poor health by an average of 2.2 percent.

After considering all these other important factors, men and women have statistically identical odds of reporting poor health.

In sum, the statistical analysis confirms that the separate structural forces captured by measures of educational attainment, household income, ethnicity, and age are indeed significantly associated with the health disparities, and each of these critical factors contributes independently to self-reported health status. We will continue to make use of the regression models to control for the effects of age, ethnicity, education, and income, as we consider the individual importance for self-reported health

Health Disparities

13

of other more subtle factors, such as having health insurance, or engaging in health-related physical activities, or living in areas that are particularly prone to air pollution, or feeling comfortable and engaged in one’s neighborhood.

how much physical activity that person gets” (with an average score of 8.6) or the “amount of stress a person experiences” (at 8.3). They viewed structural inequalities, such as “a person’s level of income” (7.2) or “a person’s experience with racial or ethnic discrimination” (6.2) as having less impact on that Although (as we shall see) individually-based risk person’s health. Not surprisingly, African Amerifactors such as diet and physical activity may well cans (at 63 percent) were more likely than Latinos contribute to health outcomes, the structural forces (53 percent), who were more likely than Anglos (45 Age, too, remains a significant predictor, with the variables controlled: each onemeasured in Table 2 are generally considered byother structural percent), to assert that living with “racial discriminayear increase in the respondents’ ages, from 18 toand 95, raisestion” the odds their effects reporting they health, giving it health scholars to be by far the more pervasive has of strong on that a person’s are in fair or poor health by anfor average of 2.2differencpercent. Afteraconsidering these other critical forces in accounting individual rating of 7 all or more on theimportant ten-point scale.

21 factors, men and women have statistically es in self-reported health status.

As theidentical findingsodds of reporting poor health. from the Houston surveys and numerous other Do the individual and behavior factors that the In sum, the statistical analysis confirms that the separate structural forces captured by measures of studies confirm, the sharp disparities in health that public identifies really contribute on their own to educational attainment, household income, ethnicity, and age are indeed significantly associated exist in this country primarily derive from the vast health outcomes after all the more powerful strucwith the health disparities, and each of these critical factors contributes independently to selfsocial and economic inequalities that are so deeply tural forcesmodels are taken into account? reported health status. We will continue to make use of the regression to control for the Aside from age, 22 ingrained in American society. The socioeconomic status, and institutionalized racial effects of age, ethnicity, education, and general income, public, as we consider the individual importance for selfhowever, is more likely to view a person’s health as inequalities, is self-rated health status also affected reported health of other more subtle factors, such as having health insurance, or engaging in primarily a matter of individual volition and specific by more immediate andpollution, more individualized expehealth-related physical activities, or living in areas that are particularly prone to air or health-related behaviors. riences, such as having health insurance, particifeeling comfortable and engaged in one’s neighborhood. pating in healthy behaviors, or benefiting from the Although (as we shall see) individually-based riskmuch factors such as diet and physical activity may The respondents in the HAHS were asked how health-related aspects of the surrounding social and well health outcomes, the structural Table 2 are generally theycontribute thought atoperson’s health is determined byforces measured naturalinenvironment?

We turn next to an assessconsidered by health scholars to beusing by fara scale the more pervasive and of critical forces in accounting each of several different factors, from ment the importance of factorsfor such as these in 21 individual differences in self-reported health status.

As the findings from the Houston surveys 1 (“very little effect”) to 10 (“very strong effect”). As determining health outcomes, as we seek to assess and numerous other studies the sharp disparities inhow health that exist country indicated in Figure 10, theconfirm, survey participants atwell this cityinisthis meeting its obligation to provide primarily derive from the vast social and economic inequalities that are so deeply ingrained in tributed the most22powerful effects to individual and quality health care and health-promoting resources American society. The general public, however, is more likely to view a person’s health as behavioral factors, such as “what a person eats and to all its residents. ■ primarily a matter of individual volition and specific health-related behaviors.

Figure - The Perceived Determinants of a Person’s Average Ratings (HAHS) Figure 1010 – The Perceived Determinants of a Person’s Health, Health, Average Ratings (HAHS) “How much do you think a person's health is determined by each of these factors? Using a scale from 1 to 10, where 1 means it has 'very little effect' and 10 means it has a 'very strong effect,' how important is each of the following?”

11 10 9

EFFECT ON HEALTH (SCALE FROM 1 TO 10)

8 7 6 5 4

8.6

3

8.3

7.35

7.26

7.23

6.93

Level of Income

Genetic Makeup

6.16

2 1 0 Diet and Physical Activity

Amount of Amount of Air and Water Stress Social Support Pollution

Racial Discrimination

Kinder Institute forHAHS Urbanwere Research 14 respondents The in the asked how much they thought a person’s health is determined by each of several different factors, using a scale from 1 (“very little effect”) to 10 (“very strong effect”). As indicated in Figure 10, the survey participants attributed the most powerful effects to

ACCESS TO HEALTH CARE Health Insurance Coverage The respondents in the HAHS were asked if they were “currently covered by any type of health insurance or health care plan.” One-fifth (19 percent) said that they had no health insurance coverage at all. This figure somewhat underreports the actual number of the uninsured. The 2009-2011 ACS 3-year estimates from the U.S. Census indicated that 27 percent of all Harris County residents and 34 percent of those aged 18 to 64 were uninsured. Nationally, 15.4 percent of the total U.S. population and 21 percent of the population aged 18 to 64 were uninsured, meaning that there were some 48 million uninsured Americans in 2012, according to the Census report.23

between $37,501 and $62,500 were uninsured. In addition, close to 34 percent of the respondents who did not have a high school diploma said they had no health insurance, compared to 23 percent of high school graduates, 14 percent of those who had some college education, and just 9 percent of college graduates.

When asked about their source of health insurance or health care coverage, 51 percent of the respondents who were insured said they were covered through an employer, and 21 percent through a government assistance program such as Medicare or Medicaid.

Another 6 percent purchased their plan directly from a private insurance company, and 1 percent through a military or veteran program. Fewer than one percent of the respondents reported that they used the Harris The likelihood of having some kind of health insurCounty Hospital District’s Gold Card (now known as ance plan differs significantly by ethnic background the “HarrisHealth Program”) as their health care plan; and increases with age, income, and education. In

the plan enables these individuals to receive limited the HAHS, well over one-quarter (29 percent) of the subsidized medical care at specified locations, but respondents under age 30 said they were uninsured, through a government assistance program such as Medicare or Medicaid. Another 6 percent technically they are not Another insured. as did 21 percent of those aged 30directly to 44. The proporthrough a government assistance program such as Medicare or Medicaid.

6 percent purchased their plan from a private insurance company, and 1 percent through a military tion droppedpurchased to 18 percent at ages 45 to 59, and to 10 plan directly fromone a private company, and 1 percent through military or veterantheir program.

Fewer than percentinsurance of the respondents reported that they usedathe Harris Figure 11 makes use ofthat the KIHAS tohealth show the orCounty veteran program.

Fewer than one percent the respondents reported they used the Harris percent among the respondents who were 60 years oldofknown Hospital District’s Gold Card (now as the “HarrisHealth Program”) asdata their County Hospital District’s Gold Card (now known as the “HarrisHealth Program”) as their health distribution of health insurance coverage in Houston’s care plan; the plan enables these individuals to receive limited subsidized medical care at specified and older. care plan; the enables they theseare individuals to receive limited subsidized medical care at specified locations, butplan technically not insured. five major ethnic communities, among all responlocations, but technically they are not insured. dents aged 18 to 64. Remarkably, almost half (48 Health insurance coverage also declines at the lower Figure 11 makes use of the KIHAS data to show the distribution of healthimmigrants insurance coverage in County percent) of the Hispanic in Harris ends of the socioeconomic Nearly 28 percent Figure 11 makes use of the KIHAS data to show thealldistribution of aged health coverage in Houston’s fivescale. major ethnic communities, among respondents 18insurance to 64. Remarkably, said wereCounty uninsured.

This was uninsured. the case for 28 of the survey Houston’s respondents whose household incomesimmigrants five major ethnic communities, among all they respondents aged 18 64. Remarkably, almost half (48 percent) of the Hispanic in Harris saidtothey were percent of theCounty Hispanics, 21 percent of the were less thanalmost $37,500 they did not any halfsaid (48case percent) thehave Hispanic in Harris saidofthey uninsured.

This was the for 28of percent of the immigrants U.S.-born Hispanics, 21U.S.-born percent the were U.S.-born blacks, 17 17 percent of Asians,blacks, and 1317percent was of theAsians, case forand 28 percent thenative-born U.S.-bornU.S.-born Hispanics, 21 percent of the U.S.-born type of healthThis insurance or health care plan.ofAlmost percent 13 percent of Anglos. blacks, of Asians, andhousehold 13 percent of native-bornof Anglos. native-born Anglos.

20 percent ofpercent those who reported incomes Figure 11 – Health Insurance Coverage by Ethnicity among Respondents Aged 18 to 64 (2001‐2013 Figure 11 – Health Insurance Coverage by Ethnicity among Respondents Aged 18 to 64 (2001‐2013 KIHAS, Combined) KIHAS, Combined) Figure 11 - Health Insurance Coverage by Ethnicity among Respondents Aged 18 to 64 (2001-2013 KIHAS, 100 10090 9080 8070 7060 6050 5040 4030 3020 2010 10 0 0

PERCENT OF RESPONDENTS PERCENT OF RESPONDENTS

Combined)

87 87

82 82

“Do you and your family currently “Do you and your family currently have any health insurance?”

71 have any health insurance?”

71

79 79

51 48 51 48

13 13

17 17

21 21

US‐Born Anglos All Asians US‐Born Blacks (N=3,761) (N=636) US‐Born Blacks (N=3,959) US‐Born Anglos All Asians (N=3,761) (N=636) (N=3,959)

28 28

US‐Born Hispanics US‐Born (N=2,791) Hispanics (N=2,791)

Hispanic Immigrants Hispanic (N=1,931) Immigrants (N=1,931)

Insured Insured Uninsured Uninsured

Table 3 uses the KIHAS data and the logistic regression model to test the separate impact of Table 3 uses theinsurance KIHAS data and the logistic regression modelthe to effects test theofseparate impact of having health on self-reported health status, when all the key structural the keyHealth Disparities having on self-reported health status, whenentered the effects of all structural factorshealth (age, insurance ethnicity, education, and income) have been into the model. When all four of factors (age, ethnicity, education, and income) have been entered into the model.

When all four of the structural factors are controlled, is the sheer fact that the respondent is or is not insured the structural factors are controlled, is the sheer fact that the respondent is or is not insured significantly associated with self-reported health status? The answer is “yes.” The analyses make it significantly associated self-reported health status? Thecorrelated answer iswith “yes.”

The analyses clear that having healthwith insurance is indeed independently reporting better make health.it

15

Table 3 uses the KIHAS data and the logistic regression model to test the separate impact of having health insurance on self-reported health status, when the effects of all the key structural factors (age, ethnicity, education, and income) have been entered into the model. When all four of the structural factors are controlled, is the sheer fact that the respondent is or is not insured significantly associated with self-reported health status? The answer is “yes.” The analyses make it clear that having health insurance is indeed independently correlated with reporting better health.

Across all categories of education, income, ethnicity, and age, the respondents who said they were currently covered by some type of health insurance or health care plan had on average 38 percent (1-OR=.381) lower odds of reporting that their current state of health was fair or poor, compared to those who were uninsured. This suggests that simply expanding

area residents’ access to health insurance may well have a significant positive impact on their overall health status, although the effects of differences in education, income, and ethnicity, not surprisingly, are generally even more important than having insurance in accounting for the health disparities.

The Congressional Budget Office projects that the Affordable Care Act should result in its first year in 14 million previously uninsured Americans being able to obtain health insurance. The data reported here suggest that this expansion in access to health insurance coverage may well have positive effects on the overall health of residents in Harris County and across America. Unfortunately, the benefits of the new act will be less evident in Texas than elsewhere in the nation, since the state government so far has decided not to accept the federal funds that would expand Medicaid to the working poor, leaving a large gap in insurance coverage.

Table 3 - Logistic Regression of Health Insurance Coverage on Self-Rated Health Status, Controlling for Table 3 – Logistic Regression of Health Insurance Coverage on Self-Rated Health Status, Controlling for the Structural Variables (2001-2013 KIHAS, Combined) the Structural Variables (2001-2013 KIHAS, Combined) Self-Reported “Fair” or “Poor” Health (KIHAS 2001-2013) Independent Variable (A) Socioeconomic Status a (1) Education High School Diploma Some College College Degree b (2) Income $35,501 to $50,000 $50,001 to $75,000 More than $75,000 (B) Demographic Characteristics c (3) Ethnicity Black Latino Asian (4) Age (5) Female

B

SE

OR

1-OR

-.090 -.256* -.781***

.124 .123 .135

.914 .774 .458

-.086 -.226 -.542

-.383*** -.524*** -.777***

.108 .107 .104

.682 .592 .460

-.318 -.408

+.411*** +.216** +.096 +.023*** -.068

.092 .097 .267 .002 .072

1.508 1.242 1.100 1.024 .935

+.508 +.242 +.100 +.024 -.065

-.540

d

(C) Having Health Insurance -.480*** .093 .619 -.381 ----------------------Constant -.425** .139 .654 -.346 N 5,435 a Reference group: Less than high school. b Reference group: Less than $35,000. c Reference group: Anglos. d The question about health insurance coverage was included in the 2001, 2003-2012 surveys. *p < .05, **p < .01. ***p < .001 (two-tailed tests).

16 Kinder Institute for Urban Research

The Congressional Budget Office projects that the Affordable Care Act should result in its first year in 14 million previously uninsured Americans being able to obtain health insurance. The data reported here suggest that this expansion in access to health insurance coverage may well have

Immigration and health insurance among The first regression model indicates that Latino Hispanics. We have seen (in Figure 8) that, among immigrants, despite their differences from U.S.-born the participants in the 13 years of the KIHAS, the Latinos in age, education, and income, do indeed (by Hispanic immigrants (at 30 percent) were more 25 percent) have higher odds than their U.S.-born likely than the U.S.-born Latinos (21 percent) to counterparts of reporting their current health status report their health as fair or poor. The immigrants as fair or poor. The second regression reveals that were also much more likely than the U.S.-born this relationship is no longer statistically significant among thetoLatino respondents, controlling for age, gender, education, Latinosreported (by 47 health to 27 percent) say they had no first after when health insurance coverage is entered into the and income, and then with health insurance coverage also entered into the equations. health insurance. Table 4 tests the relationship model. The association between being an immibetween immigrant status and self-reported health grant and reporting low levelsfrom of personal The first regression model indicates that Latino immigrants, despite their differences U.S.- health is amongborn the Latino respondents, first after conlargely explained, the data indicate, the fact that Latinos in age, education, and income, do indeed (by 25 percent) have higher odds thanbytheir trollingU.S.-born for age, gender, education, and income, and the Latino immigrants are so much less likely to be counterparts of reporting their current health status as fair or poor. The second then with health insurance coverage also entered regression reveals that this relationship is no longer statistically when health insurance covered bysignificant health insurance.

Once again, these analinto thecoverage equations. is entered into the model. The association between being anthe immigrant and reporting low access to yses reinforce importance of increasing levels of personal health is largely explained, the dataaffordable indicate, byhealth the fact that the Latino insurance across the entire Housimmigrants are so much less likely to be covered by health insurance. Once these analyses the overall ton community, if theagain, goal is to improve reinforce the importance of increasing access to affordable health insurance across the entire health of area residents.

Houston community, if the goal is to improve the overall health of area residents.

Table 4Table - Logistic Regression of Immigration Status and Insurance Coverage on Self-Reported Health 4 – Logistic Regression of Immigration Status and Insurance Coverage on Self-Reported Health iv iv Status Status among Latinos, Controlling theStructural Structural Variables (2001-2013 KIHAS, Combined) among Latinos, Controlling for for the Variables (2001-2013 KIHAS, Combined) Self-Reported “Fair” or “Poor” Health (Latinos only) (KIHAS 2001-2013) Independent Variable (A) Socioeconomic Status a (1) Education High School Diploma Some College College Degree b (2) Income $35,501 to $50,000 $50,001 to $75,000 More than $75,000 (B) Demographic Characteristics (3) Age (4) Female c

(5) Latino Immigrants d (C) Having Health Insurance ----------------------Constant N

B

SE

-.307* -.419*** -.815*** -.350*** -.788*** -.929***

(KIHAS 2001-2013)

OR

1-OR

.100 .109 .137

.736 .658 .442

-.264 -.342 -.558

.108 .123 .135

.705 .455 .395

-.295 -.545 -.605

B

SE

OR

1-OR

-.319** -.419** -.815***

.114 .122 .153

.727 .712 .459

-.273 -.288 -.541

-.280* -.727*** -.817***

.120 .135 .147

.756 .483 .442

-.244 -.517 -.558

+ .024*** +.144

.002 .077

1.025 1.155

+.025 +.155

+.024*** +.128

.003 .085

1.024 1.137

+.024 +.137

+.224**

.082

1.251

+.251

+.180 -.284**

.092 .097

1.197 .753

+.197 -.247

-1.706*** 4,185

.148

.182

-.818

-1.533*** 3,472

.167

.216

+.784

a

Reference group: Less than high school.

Reference group: Less than $35,000. c Reference group: U.S.-born Latinos. d The question about health insurance coverage was included in the 2001, 2003-2012 surveys. *p < .05, **p < .01. ***p < .001 (two-tailed tests). b

iv

To determine the relationships of the structural variables variables with health differdiffer amongamong the Latinos by To determine if theifrelationships of the structural withself-reported self-reported health the Latinos by their their immigrant status, we tested the interaction terms and found that none were statistically significant, so they are immigrant status, we tested the interaction terms and found that none were statistically significant, so they are not innot included in the final models. cluded in the final models. iv

20

Kinder Institute for Urban Research

Health Disparities

17

PERCENT OF RESPONDENTS PERCENT OF RESPONDENTS

standards patient and medical research.

Despite such quality of for health carecare available to you in the Houston area” asresources, excellent, too go have difficulty accessing quality health care.

Respondents in the HAHS we one-third (34 percent) rated the health care they could access as fair or poo quality of health care available to you in the Houston area” as excellent, go one-third percent) theinhealth care they could access as fair or poo The same(34 question wasrated asked the KIHAS on two different occasions.

In negative ratings to the quality of health care available to them in the Houst The same was asked in the KIHAS on two different occasions.

In percent in question 2009. The data are consistent in suggesting that more than one-t negative ratings to the quality of health care available to them in the Houst rate the health care they can obtain as only fair or poor, and that there has b Access to Quality Health Care in Houston percent 2009.of The in their suggesting that more than one-t over theincourse thedata pastare twoconsistent decades in assessments of the quality rate the health care they can obtain as only fair or poor, and that there has b dents said that thethe quality of care available to them Houston is home to the largest medical complex in over the12 course past two decades in household their assessments the responde quality Figure showsofthe relationship between incomesofand in the Houston area was excellent or good, compared the world and boasts some of the highest standards quality of health care available to them in the Houston area. Over half (51 to just1238shows percent ofrelationship theHAHS uninsured.

Fully 41 percent for patient care and medical research. Despite such Figure between household responde respondents in thethe 2012 who reported annual incomes incomes and of less than $ of the uninsured rated the quality of care they could resources, too many area residents have difficulty ac- quality of health care available to them in the Houston area.

Over half quality of health care available to them was fair or poor. This was true (51 for access in Houston as fair poor. cessing quality health care. Respondents in the HAHS respondents in the 2012 HAHS who annual incomes of less than $ with household incomes oformore thanreported $100,000. were asked to rate “the quality of health care available quality of health care available to them was fair or poor. This was true for to you in the Houston area” as excellent, good, fair, with household incomes of more than $100,000.

12 –12 Respondents’ Ratings of theofQuality of Health Care Available to The Figure - Respondents’ Ratings the Quality or poor. Over one-third (34 percent) rated the health Figure byof Household Income (HAHS) Health Care Available to Them in the Houston care they could access as fair or poor.

Figure 12 – Respondents’ Ratings of the Quality of Health Care Available to The Area, by Household Income (HAHS)

by Household Income (HAHS) “What about the quality of health care available to The same question was asked in the KIHAS on two you in the Houston area? Would you say: excellent, different occasions. In 1992, 37 percent gave negative “What about the quality of health care available to good, fair or poor? ” ratings to the quality of health care available to them 70 you in the Houston area? Would you say: excellent, in the Houston area, as did 38 percent in 2009. The good, fair 60 51 or poor? ” data are consistent in suggesting that more than one70 44 50 39 42 third of all area residents rate the health care they can 60 37 36 40 33 32 51 31 obtain as only fair or poor, and that there has been no 44 50 30 39 42 37 36 20 improvement over the course of the past two decades 19 17 32 40 33 31 20 in their assessments of the quality of care they can Having30 health insurance, not surprisingly, is also strongly correlated with 20 the respondents’ r 10 19 17 access.

20 of the quality of health care available to them in the Houston area. As shown in Figure 13, th

PERCENT OF RESPONDENTS

0 insured10 and uninsured respondents differ dramatically on this question. More than 70 percen Less thansaid that $37,501 More than the quality of$62,501 care available to them in the Houston area wa Figure 12 shows the relationship between household the insured respondents 0or good, $37,500 62,500 100,000 $100,000 excellent compared to just 38 percent of the uninsured.

Fully 41 percent of the unin incomes and respondents’ ratingsinsurance, of the quality of Having health not surprisingly, is also strongly correlated with the respondents’ ratings Less thanthey could $37,501 - in Houston $62,501as- fair or More than rated the quality of care access poor. quality health care available to them in the Houston area. As shown in Figure 13, the"Fair" or "Poor" "Excellent" "Good" health care availableof tothethem in of the Houston Having health insurance, notarea. surprisingly, is also strongly correlated with the respondents’ $37,500 62,500 100,000ratings $100,000 insured and uninsured respondents differ dramatically on this question.

More than 70 percent of Having health insurance, not surprisingly, is also strongly correlated with the respondents’ ratings Over half (51 percent) ofquality the respondents in available the 2012toThis of the of health care themdifference in the Houston area.

As shown Figureone 13, of thethe reasons why having health insur in access to quality care in is surely the insured said thatarea. the quality of care available to them in the"Good" Houston area was"Fair" or "Poor" of the quality health care to respondents them in the Houston Aswith Figure "Excellent" insured and uninsured respondents differ dramatically on 13, this question.

Moreofthan 70 percent isshown so strongly related tothe higher ratings personal health. of The logistic regression model (not Having healthof insurance, notavailable surprisingly, is also strongly correlated the in respondents’ ratings HAHS who reported annual incomes of less than excellent or good, compared to just 38 percent ofthan the 70 uninsured.

Fully brackets 41 percentcan of the uninsured insured and uninsured respondents differ dramatically on this question.

More percent People higher income afford insurance tran that health insurance, controlling for the coverage, effects of incom thequality insured said thatarea. the quality of confirms care available toof them in the Houstoneven areaafter wasbetter of the $37,500 quality ofsaid health care available toofrespondents them in the Houston Asshown) shown in in Figure 13, having the that the health care available thequality qualityofofcare careavailable they could access asarea fairand or age, poor.is indeed significantly associated with the respondents’ ratings the insured respondents saidrated that the to them in in theHouston Houston was education, ethnicity, insured and uninsured respondents differ dramatically on this question.

Moreofthan 70 percent ofFully excellent or good, compared to just 38 percent the uninsured.

41 percent of the uninsured clinics and hospitals, and access to a wider array of medical specialists.

Th to them was fair or poor. was true foruninsured. just one-Fullyquality excellent or good, compared to justThis 38 percent of the 41 percent of the uninsured of health care available to them in the Houston area. the insured respondents saidrated that the to them inPeople the Houston area was in higher income brackets can afford better insurance coverage, tran thequality qualityofofcare careavailable they could access in Houston as fair or poor. income is also reflected in ethnic differences: 81 percent of the Anglo resp rated the quality of care they could access inin Houston as or poor.

This difference access to fair quality care is41 surely one of reasons why having health insurance fifth those with household incomes of more than excellent orof good, compared to just 38 percent of the uninsured.

Fully percent of the uninsured Figure 13 -the Respondents’ Ratings the Quality clinics and hospitals, and access tomodel aorof wider array of medical Thc said they had access to “excellent” “good” health care, butspecialists. this was the is so strongly related to higher ratings of personal health.

The logistic regression (not $100,000. rated the quality of care theyThis could access inin Houston or poor. difference access as to fair quality careFigure isincome surely one ofCare thereflected reasons why having health insurance of Health Available to Them in the Houston 13 – Respondents’ Ratings of the Quality of Health Care Available to Them in the Houston is also in ethnic differences: 81 percent of the Anglo resp This difference in access to quality is surely of thehealth reasons why shown)care confirms thatone having insurance, evenhealth after controlling for effects of income, ofhaving blacks and insurance 52 percent ofthe Hispanics. is so strongly related to higher ratings ofbyregression personal health.

The logistic regression model (not health care, but this was the c Insurance Coverage (HAHS) Area, by Insurance Coverage (HAHS) is so strongly related to higher ratings ofethnicity, personal health.

The logistic model (not said they had access to “excellent” or “good” education, and age, is indeed significantly associated with the respondents’ ratings of the This difference inhigher access to quality care is surely one of the reasons why having health insurance shown) confirms that having health insurance, even after controlling for the effects of income, People in income brackets can afford better inshown) confirms that havingquality health of insurance, evenavailable after controlling the effectsand ofarea. income, health care tologistic them for inregression the Houston of blacks 52 percent of Hispanics. is so strongly related to higher ratings of personal health.

The model (not Having health insurance, not surprisingly, is also strongly correlated with t education, andassociated age, is indeed significantly associated with 45 surance coverage, transportation to even better clinics and education, ethnicity, and age, is indeed ethnicity, significantly with the respondents’ ratings of thethe respondents’ ratings of the 41 area. As show shown) confirms that havingquality health of insurance, after controlling for the effects ofarea. income, 40 of the quality of health care available to them in the Houston health care available to them in the Houston qualityhospitals, of health care available in the Houston area. speandand access totoindeed athem wider array of medical 40 education, ethnicity, age,Figure is significantly associated theHaving respondents’ ratings of the health insurance, not surprisingly, is also strongly correlated with 13 – Respondents’ Ratings of with the Quality of Health Care Available to Them in the Houston Area, insured and uninsured respondents differ dramatically on this question.

Mot The withinincome is(HAHS) also reflected of the qualitycialists. of health carerelationship available to them the Houston area.

35 quality

by Insurance Coverage of health care available to them in the Houston area.

As show 31 Figure 13 –Quality Respondents’ Ratings of the Quality of in Health Care Available Figure in 13 ethnic – Respondents’ Ratings of percent the Health CareresponAvailable to Them the Houston Area, to Them in the Houston Area, differences: 81 ofofthe Anglo 28 30 insured and uninsured respondents differ dramatically on this question. Mo by Insurance Coverage (HAHS)

by Insurance Coverage (HAHS)

in the HAHS saidofthey had access to Care “excellent” Figure dents 13 – Respondents’ Ratings the Quality of Health Available to Them in the Houston Area, 45 22 25 Kinder Institute for Urban Research 41 40 for just 48

by Insurance Coverage (HAHS) or “good” health care, but this was the case 45

40 45

40 45

35 40

40

41 22

31

35not surprisingly, 35 40 Having health insurance,30 is also 31 31 28 25 30 30 35 strongly correlated with the respondents’ ratings of 31 18 the quality of health care20 available to them in 28 the 25 25 30 PERCENT PERCENT OF RESPONDENTS OF RESPONDENTS

PERCENT PERCENT OF RESPONDENTS OF RESPONDENTS

40 percent of blacks 40 and 52 percent of Hispanics.

41

21

Houston area. As shown in Figure 13, the insured 18 15 20 18 and 20 25 12 21

uninsured respondents differ dramatically on this 18 10 1512 12 10 question. More than 70 percent of the insured respon-

15 20 10 15

5 10 0 5

18 0

512 10

20 15

18

21

41

Kinder Institute for Urban Research

10

28

5

28

12

"Excellen 10

"Fair"

21

0

21

Yes, Insured

"Good" "Poor"

No, Not insured "Excellent"

"Are you currently covered by any type of health insurance or health care plan?"

10 "Excellent" "Good" "Excellent"

10 "Good" "Excellent"

"Fair" "Good"

A second regression (also not shown) indicates "Poor" that the association between the respondents "Fair" 10 "Good" "Fair" assessments of the quality of health care available to them and their self-reported health statu 0 5 "Poor" "Fair" "Poor" fully explained by theNot relationships of each of these two variables with socioeconomic status Yes, Insured No, insured 0 residents with higher levels of age, education and income, and Anglos and Asians more than "Are you currently covered by any type of health insurance or health care plan?"

"Poor" Yes, Insured No, Not insured Kinder Institute for Urban Research Yes, Insured insured blacks or Latinos,No, areNot better able to access the kind of high-quality health care that results in "Are you currently covered by any typeyou of health insurance health care of plan?"

"Are currently covered by anyhealth, type health insurance Yes, Insured No, Notorinsured and they are ableortohealth do so,care theplan?" data suggest, because they are also more likely to be c "Are you currently covered by any type of health insurance or health care plan?" by health insurance.

A second regression (also not shown) indicates that the association between the respondents’

assessments of the that quality health care available to them and their self-reported health status is A second regression (also not shown) indicates the of association between the respondents’

This difference in access to quality care is surely one of the reasons why having health insurance is so strongly related to higher ratings of personal health.

The logistic regression model (not shown) confirms that having health insurance, even after controlling for the effects of income, education, ethnicity, and age, is indeed significantly associated with the respondents’ ratings of the quality of health care available to them in the Houston area.

A second regression (also not shown) indicates that the association between the respondents’ assessments of the quality of health care available to them and their self-reported health status is fully explained by the relationships of each of these two variables with socioeconomic status: Area residents with higher levels of age, education and income, and Anglos and Asians more than blacks or Latinos, are better able to access the kind of high-quality health care that results in good health, and they are able to do so, the data suggest, because they are also more likely to be covered by health insurance.

Medical Visits. More than 60 percent of the respondents in the HAHS said they (or someone in their household) had “visited a healthcare provider for medical treatment” at least once in the past 12 months. The majority of the visits (71 percent) were to a doctor’s office to see a primary-care physician; another 21 percent reported visiting clinics, health centers, or hospitals for medical treatment, and 8 percent of the visits were to a hospital emergency room.

Almost all of the respondents (92 percent) said there was “a particular place where you usually go when you’re sick or in need of health care.” The survey participants with children at home said they generally go to the same health care providers when their children are in need of care. Three-fourths (76 percent) of these parents said they most often take their child to a doctor’s office or a primary care physician, and another 21 percent visit a clinic; 2 percent said that they most often take a sick child to the hospital emergency room, and three-fourths of these respondents said they take their child to the ER even for routine or preventive care, such as a physical exam

or check-up. That seemingly small figure translates into some 14,000 (2 percent) of Harris County’s 700,000 or so residents with children, who say they bring their child to an ER for basic health care.

People without adequate insurance are severely limited in the types of health services they can access.

Many physicians refuse to accept uninsured patients because of their low rates of payment compared to those with health insurance.24 Harris County has developed a variety of safety-net clinics that offer free or low-cost care to the uninsured and to the Medicaid populations. These include the three publicly-funded Harris Health System (HHS) hospitals (Ben Taub, Lyndon B. Johnson, and Quentin Mease), 16 HHS community health centers, six school-based clinics, and 13 federally qualified health centers (FQHCs).25 A recent University of Texas study found that these providers are able to meet only about 30 percent of the demand for care among Houston’s low-income citizens.26 When such safety net providers are unavailable, many of the uninsured turn to the emergency room for the care they need, even in connection with non-emergency health issues. Patients who rely on the ER as their primary source of health care not only contribute to overcrowding in hospital emergency rooms, but also lose out on the quality and continuity of care that is provided in clinical settings.27 Despite the vast distribution of hospitals, health care facilities, and safety net clinics across Harris County, fully 12 percent of the respondents in the HAHS said that there was a time during the past year when they were unable to get the medical care they needed.

More than half (56 percent) of these respondents cited their lack of insurance or inadequate coverage as the main reason they were unable to access medical care.

Another 26 percent mentioned cost, unemployment, or lack of transportation as the barrier. In the 2010 BRFSS, 19 percent of all Harris County residents said they were unable to get the health care they needed in the past 12 months because of cost.28 ■

Health Disparities

19

ENVIRONMENTS, NEIGHBORHOODS, AND HEALTH Aside from the direct effects of age, education, Most of Houston’s air pollution is due to (a) onincome, ethnicity, and insurance coverage on indiand off-road emissions from motor vehicles, ships, viduals’ overall health outcomes, the characteristics trains, and construction equipment; (b) industrial of the neighborhoods in which people live have sources such as the petroleum refineries along the also been linked to their physical and mental health Ship Channel and the Port of Houston; and (c) area status, mortality, birth weight, chronic conditions, sources, such as surface coating processes and dry 29,30 and other important health issues.

The prescleaners.31 In 2000 the city announced an ambience, for example, of sidewalks, parks, and trails tious plan to reduce its industrial emissions by 90 Thethe Relationship Health with Airfood and may Water Quality and availability ofofaffordable healthy percent in order to meet by 2007 the standards esencourage health-promoting behaviors and make it tablished by the Environmental Protection Agency Air quality. Air pollution levels in Houston are considered to be unacceptable both by experts and 32 easier maintain them. to environmen(EPA).

Houston is still unable 7, to comply with the by theto general public.

TheExposure city’s problems in this regard were made fully evident on October tal hazards, of course, has obvious health implicasmog limits set in the 2007 deadline, 1999, when the headline in the U.S.A. Today newspaper was, “Houston, cough, cough … We’ve but air quality tions: Poor air cough, and water quality andLos proximity has improved overall: The ozone readings across got a problem, cough!”

As the Angeles to Times proclaimed on that particularly hot and industrial that Capital emit hazardous substances steadily sincehad the late 1980s balmy day,facilities “New Smog of America Declared!” Forthe thearea firsthave time fallen in history, Houston surpassed Los Angeles in the number of dangerously dayssignificant recorded during a singlegrowth, year. and Houston are likely to have adverse effects on the health of polluted despite population residents. now ranks as only the seventh most ozone-polluted 33 Most of Houston’s air pollution is due to (a) on- and off-road from motor vehicles, cityemissions in the country. ships, trains, and construction equipment; (b) industrial sources such as the petroleum refineries The Importance of Air and Water Quality

along the Ship Channel and the Port of Houston; and (c) area sources, such as surface coating According to EPA measurements, Harris County processes and dry cleaners.31 In 2000 the city announced an ambitious plan to reduce its industrial experienced days during the year 2012 when Air quality.byAir levelstoinmeet Houston arethe standards emissions 90 pollution percent in order by 2007 set by the23Environmental 32 the air was unhealthy groups, three considered to be unacceptable both by experts and setsensitive in the 2007 Protection Agency (EPA). Houston is still unable to comply with the smog limitsfor days of “generally” unhealthy air quality, and one by the general public.

The city’s problems in this deadline, but air quality has improved overall: The ozone readings across the area have fallen 34 day of “very unhealthy” air quality; this equates steadilywere sincemade the late 1980s despite regard fully evident on significant October 7,population 1999, growth, and Houston now ranks as only 33 the seventh most ozone-polluted in the country. to about one out of every 14 days when the air was when the headline in the U.S.A.city Today newspa-

demonstrably unhealthy. Poor air quality aggravates per was, “Houston, cough, cough … We’ve got a According to EPA measurements, Harris County experienced 23 days during thesuch year 2012 when According to respiratory illnesses as asthma. problem, cough, cough!” As the Los Angeles Times the air was unhealthy for sensitive groups, three days of “generally” unhealthy air quality, and one the 2010 BRFSS, five percent of Houston residents proclaimed on that particularly hot and balmy day, day of “very unhealthy” air quality;34 this equates to about one out of every 14 days when the air suffer from asthma. This was confirmed in the 2006 “New Smog Capital of America Declared!”

For the was demonstrably unhealthy. Poor air quality aggravates respiratory illnesses such as asthma.

KIHAS,suffer which found thatThis 7 percent first time intohistory, Los According the 2010Houston BRFSS, had five surpassed percent of Houston residents from asthma. was of all Harris residents said they were suffering Angeles in in thethenumber of dangerously polluted confirmed 2006 KIHAS, which found that 7 percentCounty of all Harris County residents said currently they asthma, and 11 percent said they had at some days a single year. and 11 percent said from wererecorded currently during suffering from asthma, they had at some point in their lives point in their lives been diagnosed with the disease. been diagnosed with the disease.

Figure 14 - Ratings of Air Quality and Concerns about Air Pollution (HAHS) Figure 14 – Ratings of Air Quality and Concerns about Air Pollution (HAHS) 60

52

48

PERCENT OF RESPONDENTS

50 38

40 32 30

26

Excellent / Not at all concerned

20 10

34

31

13

12 4

4

5

0 Ratings of air quality in Concerns about air Concerns about fumes the Houston area pollution from industry and smoke from cars and trucks

20 Kinder Institute for Urban Research 24

Kinder Institute for Urban Research

Good / Not very concerned Fair / Somewhat concerned Poor / Very concerned

PERCENT OF RESPONDENTS

The respondents in the HAHS were asked to evalDrawing on the combined data from the 13 years uate the quality of Houston’s air. As indicated in of the KIHAS, Figure 15 compares respondents Figure 14, 38 percent ranked Houston’s air quality from three different locations on their self-reas “fair” and another 26 percent said “poor.” A ma- ported state of health. The three groups are: (a) jority (52 percent) of area residents also said they the survey participants who are living in one of were “very concerned” about the “effects of air the ten specified APWL ZIP codes, along with pollution from industry on your family’s health,” the sixteen ZIP codes that are immediately adjawith another 31 percent saying that they were cent to them;37 (b) the respondents living in the “somewhat concerned.” In addition, almost half other ZIP codes located inside the 610 Loop; and The respondents in the were asked to evaluate thethey quality of Houston’s As indicated in codes located in Harris (48 percent) of HAHS the survey participants said (c) thoseair. residing in ZIP Figurewere 14, 38“very percent ranked Houston’s “fair” and 26 percent saidthe “poor.”

A concerned” about air thequality healthaseffects of another County outside Loop. majority (52 percent) of area residents also said they were “very concerned” about the “effects of “fumes and smoke from cars and trucks.” air pollution from industry on your family’s health,” with another 31 percent saying that they were As indicated in said the Figure, the KIHAS respon“somewhat concerned.” In addition, almost half (48 percent) of the survey participants they In order to heighten public awareness of air dents who live in or near the APWL-designated were “very concerned” about the health effects of “fumes and smoke from cars and trucks.” quality issues and to encourage efforts to reduce ZIP codes were more likely (at 25 percent) than In order to heightenthe public awareness of air quality issues and to encourage efforts to reduce in Harris County (19 peremissions, Texas Commission on Environthose living elsewhere emissions, the Texas Commission on Environmental Quality (TCEQ) maintains an “Air Pollutant mental Quality (TCEQ) maintains an “Air Pollutcent) to report that their current state of physical Watchant List” (APWL), designating the areas in various cities and counties that have elevated Watch List” (APWL), designating the areas health was fair or poor. The finding is confirmed concentrations of air pollutants and toxics. In 2012 the TCEQ placed ten of Harris County’s ZIP in various cities and counties that have elevated in the HAHS data: 26 percent of those living in the codes on its APWL designation, citing the high levels of benzene in the Galena Park and Pasadena 35 concentrations of air pollutants and toxics.

In APWL ZIP codes surrounding areas reported the Houston Shipand Channel areas and of styrene in the Lynchburg Ferry area , all of which are near 36 the section TCEQofplaced ten of Harris County’s ZIP that their health was fair or poor, compared to 16 in the 2012 southeast the city. codes on its APWL designation, citing the high percent of the respondents from other areas inside Drawing on the combined in data from the 13Park yearsand of the KIHAS, Figure compares respondents levels of benzene the Galena Pasadena the15Loop and 18 percent of those from other parts 35 The three groups are: (a) the from three on the theirLynchburg self-reportedFerry state of health. areasdifferent and of locations styrene in area, of Harris County. survey participants who are living in one of the ten specified APWL ZIP codes, along with the all of which are near the Houston Ship Channel in sixteen ZIP codes that are immediately adjacent to them;37 (b) the respondents living in the other 36 the southeast section of the city.

Further confirming ZIP codes located inside the 610 Loop; and (c) those residing in ZIP codes located in Harris the importance of location, the residents in each of these three areas also County outside the Loop. differed in the extent to which they expressed Figure 15 - Self-Reported Health Status by Proxconcern about the impact of air pollution on their Figure 15 – Self-Reported Health Status by Proximity to the APWL ZIP Codes (2001-2013 KIHAS, imity to the APWL ZIP Codes (2001-2013 KIHAS, family’s health. More than half (54 percent) of the Combined) Combined) survey participants living in or near the APWL ZIP codes said they were “very concerned” about 70 57 the health effects of air pollution, compared to 50 60 50 percent of those living in other ZIP codes inside 50 38 37 the 610 Loop, and 42 percent of those living else40 31 where in Harris County outside the Loop.

25 25 30 19

20

19

10 0 APWL ZIP codes and Inside the 610 Loop Outside the 610 Loop surrounding areas "Excellent" or "Very Good"

"Good"

"Fair" or "Poor"

Area residents who are living in the most polluted areas of the city are generally older, less affluent, less educated, and more likely to be minorities than those who are able to live elsewhere in the county. Table 5 makes use of the regression model

As indicated in the Figure, the KIHAS respondents who live in or near the APWL-designated ZIP codes were more likely (at 25 percent) than those living elsewhere in Harris County (19 percent) to report that their current state of physical health was fair or poor. The finding is confirmed in the HAHS data: 26 percent of those living in the APWL ZIP codes and surrounding areas reported that their health was fair or poor, compared to 16 percent of the respondents from other areas inside the Loop and 18 percent of those from other parts of Harris County.

26

Kinder Institute for Urban Research

Health Disparities

21

to determine if the relationship between the ZIP codes of the respondents’ home residences and their self-reported health status remains statistically significant, even after rigorously controlling for the impact of age, income, education, ethnicity, and insurance coverage.

The model indicates that the survey participants living in or near the APWL ZIP codes did indeed (by 28 percent) have significantly greater odds than the respondents living elsewhere in Harris County of reporting that their current state of health was fair or poor, even among all ethnicities and at the same levels of education, income, and The ambient Further confirming the age. importance of location, the residents in each of these three areas also air quality in the neighborhood is clearly associated differed in the extent to which they expressed concern about the impact of air pollution on their with self-reported health, over and above the family’s health. More than half (54 percent) of the survey participants living in or near the APWL existing relationship between ZIP codes said they residential were “verylocation concerned” about the health effects of air pollution, compared to and socioeconomic status.

Important are codes inside the 610 Loop, and 42 percent of those living 50 percent of those living inefforts other ZIP under way in these communities, spearheaded

by Loop. elsewhere in Harris County outside the groups such as Air Alliance Houston, to educate residentsArea aboutresidents local air who pollution issuesin and are living thetomost polluted areas of the city are generally older, less advocate affluent, for more less effective environmental educated, and morepolicies. likely to be minorities than those who are able to live elsewhere in the county. Table 5 makes use of the regression model to determine if the relationship between the ZIP codes of the respondents’ home residences and their self-reported health status remains Table 5 - Logitics Regression of Proximity the APWL Zip Codesfor onthe Self-Reported Health, Controlling statistically significant, even aftertorigorously controlling impact of age, income, education, for the Structural and for Insurance Coverage (2001-2013 KIHAS, Combined) coverage. ethnicity,Variables and insurance

Coefficient

Independent Variable (A) Socioeconomic Status (1) Educationa High School Diploma Some College College Degree (2) Incomeb $35,501 to $50,000 $50,001 to $75,000 More than $75,000 (B) Demographic Characteristics (3) Ethnicityc Black Latino Asian (4) Age (5) Female (C) Having Health Insuranced

(D) Living in or near APWL ZIP Codes

e

Self-Reported “Fair” or “Poor” Health Standard Error Odds Ratio (OR)

1-OR

-.110 -.261* -.781***

.126 .126 .139

.896 .770 .458

-.104 -.230 -.542

-.449*** -.522*** -.776***

.112 .109 .106

.638 .593 .460

-.362 -.407 -.540

+.423*** +.195* +.038 +.023*** -.101 -.460***

.095 .100 .281 .002 .073 .096

1.526 1.216 1.039 1.023 .904 .631

+.526 +.216 +.039 +.023 -.096 -.369

+.248**

.090

1.281

+.281

----------------------Constant -.457*** .144 .633 -.367 N 5,275 a Reference group: Less than high school. b Reference group: Less than $35,000. c Reference group: Anglos. d The question about health insurance coverage was included in the 2001, 2003-2012 surveys. Reference group: Living in areas other than APWL ZIP codes or its surrounding areas. *p < .05, **p < .01. ***p < .001 (two-tailed tests).

22 Kinder Institute for Urban Research

Water quality. Houston’s water sources – its rivers, As indicated in Figure 16, almost half (47 percent) reservoirs, and groundwater – have also been subof the respondents in the HAHS rated “the quality jected to various forms of pollution. Harris County’s of the water in Houston’s bayous and streams” increasingly crowded urban centers, its large areas as “poor,” and another 29 percent gave it a rating of undeveloped land, and its dense industrial develof “fair.” Strikingly, more of the survey responopments present many challenges for water quality dents were concerned about the health effects of management. A study by the Environmental Workpoor water quality than of poor air quality: As ing Group, a research and advocacy organization indicated earlier, 52 percent said they were “very specializing in environmental protection, found that concerned” about the health effects of “air polthe City of Houston’s drinking water contains 18 lution from industry,” but almost three-fourths hazardous chemicals in concentrations that exceed (73 percent) expressed that level of concern with federal health guidelines.38 Some of the most serious regard to “toxic wastes leaking into our streams have to that do with high levels of bacteria, Theconcerns model indicates the survey participants living in or near APWLAnother ZIP codes indeed andthe bayous.”

72did percent said they were oxygen, nutrients, and PCBS/dioxin (by dissolved 28 percent) have significantly greater odds thancomthe respondents living elsewhere in Harris “very concerned” about “standing water that can pounds, all of which have negative effects on aquatic County of reporting current state of health was fair or poor, even among all ethnicities The model indicatesthat thattheir the the APWL ZIP codes did indeed 39 survey participants living in or near breed mosquitoes in your neighborhood.”

The contamination is dueThe ambient life andsame human health. and at the levels of education, income, andthan age. quality in the neighborhood (by 28 percent) have significantly greater odds the respondentsairliving elsewhere in Harris primarily to urban and agricultural runoff and to is clearly with self-reported health, over and existing between County ofassociated reporting that their current state of health wasabove fair orthe poor, even relationship among all ethnicities illegal dumping into the sources of the area’s water residential location and socioeconomic status.

Important efforts are under way in these and at the same levels of education, income, and age. The ambient air quality in the neighborhood supply.associated communities, spearheaded by groups such as Air Alliance Houston, to educate residentsbetween about is clearly with self-reported health, over and above the existing relationship local air pollution and to advocatestatus. for more effective environmental policies. residential locationissues and socioeconomic Important efforts are under way in these communities, spearheaded by groups such as Air Alliance Houston, to educate residents about local air pollution issues and to advocate for more effective environmental policies.

Figure 16 - Ratings of Water Quality and Concerns Figure 16 –Water RatingsPollution of Water (HAHS) Quality and Concerns about Water Pollution (HAHS) about Figure 16 – Ratings of Water Quality and Concerns about Water Pollution (HAHS)

PERCENT OF RESPONDENTS PERCENT OF RESPONDENTS

80

73

72 70 80 73 72 60 70 47 50 60 40 47 50 29 30 40 17 17 20 29 30 10 8 10 3 3 5 17 17 20 1 10 0 8 10 Ratings of water Concerns about 3 3 5 about Concerns 1 quality in toxic wastes standing water 0 Houston's leaking into the that can breed Ratings of water Concerns about Concerns about bayous and streams and mosquitoes quality in toxic wastes standing water streams bayous Houston's leaking into the that can breed bayous and streams and mosquitoes Excellent streams / Not at all concerned bayous Good / Not very concerned Fair / Somewhat concerned Excellent / Not at all concerned

Poor / Very concerned Good / Not very concerned

Fair / Somewhat concerned

Poor / Very concerned

Neighborhood Characteristics and Health-Related Behaviors

Health Disparities

Food insecurity. The respondents in the were asked, “How serious a problem has it been Neighborhood Characteristics andHAHS Health-Related Behaviors for you in the past year to pay for the groceries to feed your family?” More than one-fourth (27 percent) said that The this had been a “somewhat” orwere “veryasked, serious” problem. surprisingly, Food insecurity. respondents in the HAHS “How seriousNot a problem has it these been

23

Neighborhood Characteristics and Health-Related Behaviors

Supermarkets are the main way of providing Americans with fresh and healthy food options, but they are not always readily available. A study conducted in 2011 found that the greater Houston area contains one supermarket for every 12,000 people, whereas the national average is one for every 8,600 people.40 Some 440,000 Houston area residents live in areas designated as “food deserts” – neighborhoods that are judged to provide inadequate access to grocery stores and fresh foods.v Not surprisingly, these are primarily found in historically underserved communities, such as inner-city and rural areas, and particularly in low-income communities of color.41

Food insecurity. The respondents in the HAHS were asked, “How serious a problem has it been for you in the past year to pay for the groceries to feed your family?” More than one-fourth (27 percent) said that this had been a “somewhat” or “very serious” problem. Not surprisingly, these responses are strongly associated with education, income, and ethnicity, as well as with the respondents’ own self-rated health status: 31 percent of the survey participants who said they had difficulty paying for the groceries to feed their families also reported that their current health The USDA created a helpful tool to capture the locastatus was fair or poor, compared to 14 percent of tion of food deserts and to measure access to healthy arms Rolling Green Market,” the initiative was“not recognized those who said that paying for food was much by the U.S. Conference of and affordable foods nationwide. Figure 17 shows as an innovative way to of a problem. ” combat childhood obesity in a large urban community. the census tracts in the Houston region that were identified as both low-income and as having inad– The USDA Map of Houston Area Food Deserts, Reflecting Both Income and Access equate access to healthy foods.v These underserved Figure 17 - The USDA Map of Houston Area areas are heavily concentrated on the eastern side of Food Deserts, Reflecting Both Income and the city, in the triangle of neighborhoods between Access U.S. 59 North and the Interstate 10 East; and east of Highway 288 on the south side. Residents in areas that have no large local grocery stores are more likely to experience diet-related health problems such as obesity and diabetes.42

Almost 15 percent of the respondents in the HAHS disagreed with the suggestion that they had “easy access to fresh fruit and vegetables in your neighborhood.” Almost half (48 percent) of those who said they had easy access to fresh foods rated their current health as excellent or good, compared to only 21 percent of those who strongly disagreed with the suggestion that healthy foods were readily accessible in their neighborhoods. In the regression model (not shown), when the cumulative impacts of income, education, and ethnicity are entered into the equation, the relationship between access to healthy foods and self-reported health status disappears. The interconnection, important as it is, exists only because access 43 Source: USDA Food Access Research Atlas. Data are to healthy foods and self-reported health are both from the 2012 report,43“Access to Affordable and NutriSDA Food Access Research Atlas. Data are from the 2012 report, “Access Affordable and Nutritious sotoclosely associated with income, education, and tious Food: Updated Estimates of Distances to Supermarated Estimates of Distances to Supermarkets Using 2010 Data.” ethnicity. kets Using 2010 Data.”

The U.S. Department of Agriculture defines a “low-income census tract” an Health vProgram at the Kinder Institute(USDA) for Urban Research has developed a is as one in which the poverty rate is 20 percent or higher, or the median family income is less than or equal to 80 percent of the metropolitan area’s overall ensive guide to all farmers’ markets operating within the city of Houston. The report, median family income. A “low-access community” is one in which at least 500 people and/or at least 33 percent of the “A Guide on Local Agriculture for Houstonians,” lists all farmers’ markets and other local census tract’s population live more than one mile away from a supermarket or large grocery store. of fresh produce in the Houston metropolitan area (inside Beltway 8), and provides a map Kinder Institute for Urban Research 24 of the location markets where organic or local produce are available and where vendors ayment via credit/debit cards, SNAP, or WIC.44 Many other Houston initiatives (such as O Houston, Healthy Living Matters45, and the Go Healthy Houston Task Force) are also

The Urban Health Program at the Kinder Institute for Urban Research has developed a comprehensive guide to all farmers’ markets operating within the city of Houston. The report, entitled “A Guide on Local Agriculture for Houstonians,” lists all farmers’ markets and other local sources of fresh produce in the Houston metropolitan area (inside Beltway 8), and provides a map showing the location of markets where organic or local produce are available and where vendors accept payment via credit/debit cards, SNAP, or WIC.44 Many other Houston initiatives (such as CAN DO Houston, Healthy Living Matters,45 and the Go Healthy Houston Task Force) are also making important contributions to the ongoing efforts to create more healthy communities.

percent). Only 23 percent of the African-American respondents said they were “very satisfied” with their level of physical activity.

Figure 18 also shows the relationship between such satisfaction and self-rated health. One-fifth (20 percent) of the respondents who reported their physical health as fair or poor said they were not at all satisfied with the amount of physical activity they were getting, compared to just 4 percent who said their health was excellent or very good.

When the respondents who were not satisfied with their activity levels were asked in an open-ended question to indicate “the most important reason that kept you from getting the amount of exercise that you would like,” more than 40 percent said they were “too busy,” had “no time for exercise,” or were “too tired after work.” Another 16 percent admitted that the main reason was “sheer laziness.” Another one-fourth (27 percent) indicated they had existing health problems that prevented them from getting more exercise. Only 5 percent said that there was no convenient place where they could get the physical activity they wanted.

Physical Activity. Regular physical exercise is another important factor in maintaining vitality, helping to control weight, and reducing the risk of a variety of preventable conditions such as heart disease and diabetes.46 Moderate-intensity aerobic activity, for example brisk walking for a total of 150 minutes per week, is highly recommended by the CDC and is safe for most people. As indicated in Figure 18, a third of the HAHS respondents Figure 18 were also shows the relationship said they “very satisfied” with “thebetween amount such satisfaction and self-rated health. One-fifth (20 percent)activity of the you’re respondents who reported health asalmost fair orthree-fourths poor said they were of physical generally getting thesetheir physical As we have seen, of the renot at”all the amount of physicalsatactivity they were getting, just 4with days, andsatisfied anotherwith 40 percent were “somewhat spondents were at leastcompared somewhat to satisfied percent who said health were was excellent or very good. isfied.” Anglos (at their 37 percent) the most likely their level of physical activity; 60 percent of those to be “very satisfied,” followed by Hispanics (33 who expressed such satisfaction reported that they

Figure 18 – Respondents’ Satisfaction with Their Level of Physical Activity, by Their Self-Rated Health Figure(HAHS) 18 - Respondents’ Satisfaction with Their Level of Physical Activity, by Their Self-Rated Health Status Status (HAHS) 50 45

40

40 30

26

25

19

20

38

20 21

28

18

Not at all satisfied

13

15 10

“How satisfied are you with the amount of physical activity you're generally getting these days?”

44

41

33

35 PERCENT OF RESPONDENTS

42

8

6

5

Not very satisfied Somewhat satisfied

4

Very satisfied

0 Total

"Fair" or "Poor"

"Good"

"Excellent" or "Very Good"

Health Disparities

When the respondents who were not satisfied with their activity levels were asked in an openended question to indicate “the most important reason that kept you from getting the amount of exercise that you would like,” more than 40 percent said they were “too busy,” had “no time for

25

had engaged “in enough vigorous activity (through sports or exercise) to work up a sweat” several times a week during the past month. One-fifth (18 percent) of the survey participants said they had not engaged in any physical activity at all during the past month.

Similarly, the 2010 BRFSS found that 22 percent of the Houston respondents reported engaging in no leisure-time physical activity at all during in the previous month.47 The figure for all Texas adults was 27 percent.

People who engage in vigorous sports or other activities are less likely to report being in fair or poor health than those who do not get any meaningful exercise. Among all the Anglo respondents in the HAHS who were aged 45 or older and reported household incomes of more than $62,500, only 7 percent of those who said they had engaged in vigorous activities during the past 30 days said they were in poor health, compared to 14 percent of those who did not get any exercise.

Not surprisingly, as age increases, the likelihood of frequent exercise decreases. Almost two-thirds (63 percent) of the respondents in the HAHS who were under the age of 30 said they exercised several times a week, and just 10 percent said they got no exercise at all. In contrast, 29 percent of area residents aged 60 and older reported no vigorous physical activity at all in the past month.

Regression analyses indicate further that engaging in vigorous physical activity is indeed associated with self-reported health even after all the structural components are controlled. Does better health lead a person to engage in more physical activity, or does the exercise itself lead to better health? Both directions of causality are surely at play, and the high correlation with income found for both physical activity and self-reported health may well help to explain why differences in socioeconomic status are so powerfully related to health status.

Household income is also correlated with the reported frequency of physical activity. As indicated in Figure 19, the percent of the survey participants who said they had not engaged in any vigorous physical activity at all during the past week dropped from 22 percent among those with household incomes below $37,500 to just 7 percent in the highest income bracket. Almost two-thirds (64 percent) of the wealthiest respondents reported exercising multiple times per week, compared to just 40 percent of the low-income respondents.

Figure 19 Vigorous Physical Activity

by Income Level Level (HAHS) Figure 19 –- The The Reported ReportedFrequency Frequencyofof Vigorous Physical Activity

by Income (HAHS) 70

64

PERCENT OF RESPONDENTS

60

40

38

32

30 20

51

49

50

18

44

40

22

36

20

32

30

17

Not at all in the past month 7

10

“During the past 30 days, how many times, if at all, did you engage in enough vigorous activity (through sports or exercise) to work up a sweat?”

At least once Several times a week

0 Total

Less than $37,500

$37,50162,500

26 Kinder Institute for Urban Research

$62,501100,000

More than $100,000

Walkability. Houston is ranked by “WalkScore.com” as the 23rd most walkable large city in America (i.e., among cities with populations greater than 380,000).

The city’s neighborhoods that were identified as the most walkable were all located within the 610 loop; they included the Neartown-Montrose area, the Greenway-Upper Kirby area, and Midtown. When asked about the statement, “my neighborhood has plenty of safe and well-connected sidewalks,” fully 47 percent of the survey participants strongly agreed, and only 24 percent strongly disagreed. As indicated in Figure 20, the respondents who lived outside Houston’s city limits were more likely (at 70 percent) than those residing inside the Loop (at 58 percent) to agree that their neighborhoods have plenty of safe and well-connected sidewalks.

Good sidewalks do not necessarily mean that people can walk to where they need to go, since that depends on proximity to local amenities. Well over half (57 percent) of all the HAHS respondents answered in the affirmative when asked if there were “shops or restaurants within easy walking distance of your home.” Of those living inside the Loop, Figure 20 indicates that 66 percent said they lived within walking distance of shops or restaurants. This was the case for 62 percent of those who lived in the city but outside the Loop and for 51 percent of the respondents

who lived in the surrounding areas of Harris County beyond the city limits. Homes outside Houston may be more likely to have safe and well-connected sidewalks than homes in the city, but they are less likely to have shops or restaurants within easy walking distance.

Walkable neighborhoods with easy access to local amenities offer many health benefits compared to the car-dependent lifestyle that is so much more typical of this low-density, spread-out metropolitan area.

The KIHAS has found over the years that increasing proportions of Harris County residents have been expressing a preference for more walkable options.

In alternating years since 2008, the survey respondents were asked what type of housing they would prefer if they could live anywhere they wanted.

The numbers opting for “a single-family home with a big yard, where you would need to drive almost everywhere you want to go” dropped from 59 percent in 2008 to 47 percent in 2012. The proportion saying they would choose instead “a smaller home in a more urbanized area, within walking distance of shops and workplaces” increased from 36 percent in 2008, to 39 percent in 2010, and to 51 percent in 2012.48 Similarly, one-half of all area residents in the 2013 KIHAS said they would prefer to live in “an

PERCENT OF RESPONDENTS

Figure 20 - Neighborhood Amenities in the Houston Area by Location of Home Residence (HAHS) Figure 20 – Neighborhood Amenities in the Houston Area by the Location of Home Residence (HAHS) 100 90 80 70 60 50 40 30 20 10 0

88

58

65

70

66

71 62 51 34

Inside the Loop Outside the Loop, in the city Outside the city

My neighborhood has Shops or restaurants plenty of safe and well- are in easy walking connected sidewalks distance of my home

I have easy access to public transit, such as bus or light rail

Good sidewalks do not necessarily mean that people can walk to where they need to go, since that HAHS Health Disparities 27 depends on proximity to local amenities. Well over half (57 percent) of all the respondents answered in the affirmative when asked if there were “shops or restaurants within easy walking distance of your home.” Of those living inside the Loop, Figure 20 indicates that 66 percent said

area with a mix of developments, including homes, shops, and restaurants,” rather than in a “single-family residential area.” Houston developers would be well advised to provide effective responses to these changing demands.

Because using public transit involves some walking to and from stations, transit users generally engage in more physical activity than those who commute only by car.49 Houston’s mass transit consists of a large bus system and a well-established 7.5-mile light rail system, which is in the process of expanding to 20 miles. More than a third (36 percent) of the survey respondents strongly agreed with the suggestion that “I have easy access to public transit, such as a bus or light rail, in my neighborhood,” while another third (34 percent) strongly disagreed.

As also shown in Figure 20, there is a strong relationship between reported access to mass transit and the location of the respondents’ residence in the metropolitan area. Fully 88 percent of the survey respondents whose homes were inside Loop 610 agreed that they have easy access to transit, compared to 71 percent of those residing inside the city but outside the Loop, and to just 34 percent of those living in the surrounding areas beyond the city limits. Compared to other large metropolitan areas, this spread-out low-density city is still far behind in the provision and usage of public transit.

In New York City, 31 percent of commuters are public transit users, as are 15 percent in San Francisco and 14 percent in Washington, DC.50 The city of Houston has been actively promoting bicycle commuting with the B-cycle program, and it is in the process, through the remarkable “Bayou Greenways 2020” Initiative, of constructing a network consisting of over 250 miles of off-street bike lanes and bayou trails.51 There is room for growth in this connection: Today, about 0.4 percent of area residents say they bike regularly to work.52 According to the 2013 KIHAS, more than 80 percent of all workers in Harris County drive alone when commuting to their jobs.

28 Kinder Institute for Urban Research

The respondents in the HAHS were asked how often during the past year they rode “on Houston’s mass transit, such as bus or light rail, to get where you wanted to go?” Among the survey participants living in areas of Harris County beyond the city limits, 86 percent said they had not used any public transit at all during the past year, compared to 63 percent of those living inside the Loop. Fewer than 8 percent of all respondents said they used public transit at least once a week, and one-third of those who said they use public transit on a regular basis actually disagreed when asked if they had easy access to public transit in their neighborhood.

Greater use of alternative modes of transportation might encourage area residents to integrate exercise into their daily routines and help reduce obesity.

Only 30 percent of the respondents in the HAHS said they were “very satisfied” with their current weight; 39 percent were “somewhat satisfied,” and 30 percent were either “not very satisfied” or “not at all satisfied.” The 2010 BRFSS estimated that fully 66 percent of Houston area residents were either overweight or obese.53 Close to three of every four Hispanics and blacks were found to be overweight. This condition can have serious consequences for health, including hypertension, diabetes, heart disease, cancer, and many other problems.54 In the HAHS, 19 percent of blacks and18 percent of Latinos said “yes,” when asked if they had “any health issues right now that are affected by your current body weight,” compared to 11 percent of Anglos.

The use of public resources. More than half (58 percent) of the survey respondents said there was “a hike or bike trail within a mile or so of your home.” Yet 57 percent indicated that they had never visited “any of Houston’s bayous or hike and bike trails, whether for exercise, or recreation, or for commuting to work during the past 12 months,” and only 14 percent said they had done so as often as once a week. Living near a hike and bike trail was highly correlated with visiting frequency, but fully 46 percent of those who reported having a nearby hike and bike trail said they had not visited “any of Houston’s bayous or hike and bike trails” during the past year.

Workplaces differ significantly in the degree to which they provide employees not only with health care insurance or paid sick and personal leaves, but also with worksite wellness programs that are designed to lower medical costs and increase productivity by encouraging their employees to engage in more physical activity and healthier behaviors. In a 2012 national survey of 512 large companies (those with 1,000 or more employees), 87 percent were found to have a wellness program in place.55 The programs often include financial incentives for healthy behaviors, such as subsidized gym memberships, discounts on insurance premiums, and even cash payouts to encourage regular health screenings or physical activity.56

The survey found similar trends with regard to the availability and usage of Houston’s parks. Fully 83 perA study by the RAND Corporation found that cent of all respondents said there was “a park or play programs aimed at helping employees with chronic area within a mile or so of your home,” but 38 percent illnesses (e.g., high blood pressure, diabetes) resulted said they had not once visited a park or playground in significant cost savings to the companies. Lifestyle during the past year. Nevertheless, proximity to parks management programs, however, which aim to reduce and trails does have a generalized effect on percephealth risks through weight loss or stress managetions of the neighborhood as a whole. As indicated in ment, did not demonstrably achieve any net savings.57 Figure 21, the survey participants who said they lived within a mile of a “park or play area” or a “hike and Almost half (44 percent) of the survey respondents bike trail” were more likely than those living farther in the HAHS who were employed said that their home,” 38 percent saidwith theythehad not once visited a park playground the past year. away from thesebut amenities to agree suggesworkplace offersor“activities thatduring are designed to keep Nevertheless, proximity parks and trails does have a generalized effectwhose on perceptions tion that “I often see other people to being physically employees healthy.” Of those workplaces of the neighborhood as a whole.

As indicated in Figure 21, the survey participants who said particithey lived active (such as walking, jogging, bicycling, playing offered such programs, 40 percent said they a mile of a “park or play area” or a “hike andinbike were likely than thoseonce living sports)within in my neighborhood. ” pated suchtrail” activities at more their workplace at least farther away from these amenities to agree with the suggestion that12“Imonths, often see a week during the past 27 other percentpeople said being physically active (such as walking, jogging, bicycling, playingatsports) in my neighborhood.” they participated least occasionally, and 32 percent said they made no use at all of those resources.

PERCENT OF RESPONDENTS

Figure 21 – The Perceived Physical Activity in the Neighborhood by Proximity to Local Parks and Trails Figure(HAHS) 21 - The Perceived Physical Activity in the Neighborhood by Proximity to Local Parks and Trails (HAHS) 100 90 80 70 60 50 40 30 20 10 0

87

85 76 67

31

“I often see other people being physically active (such as walking, jogging, bicycling, or playing sports) in my neighborhood.”

23 14

13

Disagree Agree

No park or play area nearby

Park or play area nearby

No hike or bike Hike or bike trail trail nearby nearby

Health Workplaces differ significantly in the degree to which they provide employees notDisparities only with29 health care insurance or paid sick and personal leaves, but also with worksite wellness programs that are designed to lower medical costs and increase productivity by encouraging their employees

It is interesting to note that in all these varied areas – having to do with the use of public transit, of nearby parks or hike and bike trails, and of wellness programs at one’s place of work – the surveys suggest that many area residents are failing to take advantage of readily available opportunities for healthenhancing activities.

Resources and Their Frequency of Use during the Figure 22 – The Availability of Health-Promoting Resources and Their Frequency of Use during the Past Past YearHarris among Residents (HAHS) Year among CountyHarris Residents County (HAHS)

90

participation. As was indicated in Figure 10, when asked what determines a person’s overall health, the HAHS respondents83attributed the strongest effects to “what that person eats and how much physical activity that person gets,” yet many are simply not motivated enough or are not 80 given sufficient encouragement to take advantage of the available opportunities to engage in the healthy behaviors they recognize to be so important.

70

Figure 22 – The Availability of Health-Promoting Resources and Their Frequency of Use during the Past Year among Harris County Residents (HAHS)

PERCENT OF RESPONDENTS

Participation and belonging. Because levels of stress are strongly implicated in health outcomes, a person’s overall health and well-being may also be affected by the nature of the social relationships that characterize their neighborhoods. People who live in places with high levels of social capital, interpersonal trust, and a strong sense of belonging have been found to report better physical and mental health.58 Conversely, living in neighborhoods with low levels of social cohesion and high degrees of alienation and social disorder has been shown to be related to high levels of anxiety and depression,59, 60 and to lower levels of self-rated health.

The data from the Houston surveys can help to clarify the degree to which neighborhood characteristics such as these are associated with the health of residents, even after individual differences in age, education, income, and ethnicity are taken into account.

participation. As was indicated in Figure 10, when asked what determines a person’s overall health, the HAHS respondents attributed the strongest effects to “what that person eats and how much physical activity that person gets,” yet many are simply not motivated enough or are not given sufficient encouragement to take advantage of the available opportunities to engage in the healthy behaviors theyAvailability recognize to be so

Figure 22 - The ofimportant.

Health-Promoting

60 90

58

56 83

50

44

80

A Pew Research Center study in 2012 reported that, of the ten largest metropolitan areas in the country, Houston is the most segregated by income.61 The Census estimates that 18.7 percent of all Harris County residents were living in poverty in 2010, with 11 percent residing in extreme-poverty neighborhoods – areas in which at least 40 percent of the residents are living below the poverty line. Almost a fifth (19 percent) of those whose homes were inside the Houston city limits were living in areas of concentrated poverty.62 Concentrated poverty is often

40 70

30

60

24

58

56

20 50

14 44

10

44

40 0

30

30

24 20 14 10

0

30 Kinder Institute for Urban Research

44

30 PERCENT OF RESPONDENTS

Figure 22 illustrates the discrepancy between resource availability and participation. As was indicated in Figure 10, when asked what determines a person’s overall health, the HAHS respondents attributed the strongest effects to “what that person eats and how much physical activity that person gets,” yet many are simply not motivated enough or are not given sufficient encouragement to take advantage of the available opportunities to engage in the healthy behaviors they recognize to be so important.

36

Kinder Institute for Urban Research

accompanied by very low levels of social capital as a result of social exclusion and fragile social networks.63, 64 Such communities offer far less access to employment opportunities and to public services,65 and they are the most likely to suffer when public spending and social services are reduced.66 The 2012 KIHAS included a special module of questions measuring the respondents’ sense of belonging in their neighborhoods and the extent of their participation in neighborhood activities. The respondents’ feeling of connectedness to their neighborhoods was measured by a scale composed of three questions. The survey participants were asked whether they agreed or disagreed with the following statements: (a) “I feel like I belong in my neighborhood”; (b) “Living in my neighborhood gives me a sense of community”; and (c) “I (do not) think of myself as different from most of the other people in my neighborhood.” Similarly, four items were combined to measure neighborhood participation: the respondents were asked how often, if at all, they participated in neighborhood activities having to do with (a) arts or cultural activities, (b) neighborhood clubs and associations, (c) non-profit activities, and (d) block parties.

As indicated in Table 6, the logistic regression revealed that both of these composite measures of neighborhood involvement were significant independent correlates of self-reported health, after controlling for the effects of differences in age, gender, education, income, ethnic background, and insurance coverage. The analysis revealed that residents who expressed a greater sense of community and stronger feelings of connectedness in their neighborhoods had significantly lower odds of reporting that they were in fair or poor health (OR=0.962, p<0.05).

The composite measure of neighborhood participation was also associated with self-rated health (OR=0.883, p<0.06) when controlling for the other variables, including the scale measuring neighborhood belonging. The more actively involved the respondents were in neighborhood activities, the less likely they were to report being in fair or poor health.

Here, again, the arrow of causality is likely to point in both directions: better health leads to more participation, and more participation may reduce stress and thus contribute to better health.

Other studies have also found that living in neighborhoods with higher levels of social trust is associated with more positive reports of one’s overall state of physical health.67 Residents who participate in close relationships of solidarity and trust are more likely to work together for common objectives, such as sustaining clean and safe public spaces, looking out for neighborhood children, and maintaining informal controls that discourage health-damaging or undesirable behaviors such as smoking or littering.68 The Houston surveys show clearly that the kinds of neighborhoods in which people live play an important role in fostering or discouraging healthy behaviors, with significant implications for the overall health of the community.■ Table 6 - Logistic Regression of the Composit Measures of Neighborhood Belonging and ParticipaTable 6 – Logistic Regression of the Composite Measures of Neighborhood Belonging and Participation tion on Self-Reported Health Status (2012 KIHAS) on Self-Reported Health Status (2012 KIHAS) Self-Reported “Fair” or “Poor” Health

Independent Variable (A) Socioeconomic Status a (1) Education High School Diploma Some College College Degree b (2) Income $35,501 to $50,000 $50,001 to $75,000 More than $75,000 (B) Demographic Characteristics c (3) Ethnicity Black Latino (4) Age (5) Female (C) Having Health Insurance (D) Neighborhood Belonging (three items) (E) Neighborhood Participation (four items) ----------------------Constant N

Coefficient

(KIHAS 2012) Standard Odds Ratio Error (OR)

1-OR

-.307 -.254 -.944**

.265 .262 .302

.736 .775 .389

-.264 -.225 -.611

-.272 -.391 -1.060***

.239 .242 .242

.762 .676 .346

-.238 -.324 -.654

+.583** +.397* +.027*** -.069 -.579**

.209 .219 .005 .164 .200

1.792 1.488 1.028 .934 .561

+.792 +.488 +.028 -.066 -.439

-.039* -.124*

.019 .065

.962 .883

-.038 -.117

-.637 990

.463

.529

-.471

a

Reference group: Less than high school.

Reference group: Less than $35,000.

Those indicating their ethnicity as “Asian” or “Other” were omitted because of their small numbers in this one (2012) survey. Anglos make up the reference group. *p < .06, **p < .01. ***p < .001 (two-tailed tests). b c

SUMMARY AND CONCLUSIONS Introduction. Supported by a grant from Houston Endowment Inc. and aided by an advisory panel of health experts, researchers at Rice University’s Kinder Institute for Urban Research developed the 2012 “Houston Area Health Survey” (HAHS). Through systematic interviews with a representative sample of 1,200 Harris County residents, the surveys measured area residents’ selfreported health status, their experience with Houston’s health care delivery systems, and the health-related characteristics of their neighborhoods. Buttressed by additional rich data from 13 Area Survey” Health Disparities years of the “Kinder Institute Houston (KIHAS), this report seeks 31 to identify the

The SHEA Health Survey 37

SUMMARY AND CONCLUSIONS Introduction. Supported by a grant from Houston Endowment Inc. and aided by an advisory panel of health experts, researchers at Rice University’s Kinder Institute for Urban Research developed the 2012 “Houston Area Health Survey” (HAHS).

Through systematic interviews with a representative sample of 1,200 Harris County residents, the surveys measured area residents’ self-reported health status, their experience with Houston’s health care delivery systems, and the health-related characteristics of their neighborhoods. Buttressed by additional rich data from 13 years of the “Kinder Institute Houston Area Survey” (KIHAS), this report seeks to identify the most important forces that demonstrably underlie and help to account for the wide disparities among Harris County residents in their self-reported health and personal well-being.

Social structures and health disparities. In both the HAHS (2012) and the KIHAS (2001-2013), the survey participants were asked to rate their current state of physical health as “excellent, very good, good, fair, or poor.” This simple measure of self-reported health has been found to be a remarkably reliable indicator of a person’s actual state of health.

In all the survey years and from both sources of data, approximately 45 percent of area residents have said that their health was “excellent” or “very good”; another 35 percent were in “good” health, and fully one-fifth (19 to 21 percent) of all Harris County residents said their current state of health was only “fair” or “poor.” What are the most important factors that account for these disparities in self-reported health?

Metropolitan areas that have higher levels of income inequality also have higher proportions of residents who report that they are in fair or poor health. In the Houston surveys, sophisticated regression analyses make it clear that each increase in the respondents’ level of household income is significantly associated with decreasing odds that they will report being in fair or poor health, even after controlling for the effects of differences in age, education, ethnicity, and gender.

32 Kinder Institute for Urban Research

Similarly, at all levels of income and age, and among all ethnicities, educational attainment beyond high school is directly and powerfully associated with better health outcomes. African Americans and Hispanics, with the other structural variables controlled, have significantly higher odds than Anglos and Asians of reporting that they are in poor health.

Factors beyond and in addition to the ethnic differences in education and income are contributing to the more negative health outcomes reported by the black and Latino residents of Harris County.

Education, income, ethnicity, and age are thus decisive, pervasive, and independent correlates of the health disparities in Harris County. The sharp differences in health and well-being across the region and throughout America exist in large part because the social and economic inequalities are so deeply ingrained and growing larger in this society. When asked to identify the most important determinants of a person’s health, the participants in the HAHS survey were more likely to cite individual and behavioral factors (such as diet and physical activity) than they were to point to the economic and ethnic disparities.

With controls in place for the effects of age, ethnicity, education, and income, we assess the additional importance for self-reported health of these more subtle factors, such as having health insurance, engaging in health-related physical activities, living in areas that are particularly prone to air pollution, or feelings of connectedness and engagement in the neighborhood. Are any of these other factors demonstrably associated with self-reported health status after all the powerful structural forces (age, income, education, and ethnicity) are taken into account?

Access to health care. At all levels of education, income, and age, and among all ethnicities, people with health insurance have readier access to high quality health care. Latino immigrants generally report worse health than the U.S.-born Latinos, but

they are also far less likely to have health insurance. The statistical analyses reveal that the health disparities among Latinos by immigration status disappear when having insurance is entered into the equations. In the general regression model, respondents of all ethnicities and at all levels of age and SES who are covered by health insurance have significantly higher odds of reporting that their state of health is excellent or good, and they are far more likely than those without insurance to give positive ratings to the quality of health care available to them in the Houston area.

Houston is home to the largest medical complex in the world and boasts some of the highest standards for patient care and medical research. At the same time, however, Houston has one of the highest proportions among major American cities of children without health insurance (at 19 percent), and Texas in 2012 had the highest rate of uninsured in the nation (24 percent of the state’s population). People without adequate insurance are severely limited in the types of health services they can access.

Fully 12 percent of the HAHS respondents said there was a time during the past year when they were unable to get the medical care they needed, mainly (more than half of them said) because of a lack of adequate health insurance. Providing greater access to affordable health insurance for all area residents might well contribute importantly, the surveys suggest, to improving the health and wellbeing of the Harris County population as a whole.

Unfortunately, the state of Texas so far has chosen not to expand Medicaid to the working poor, leaving a large gap in insurance coverage.

Environments, neighborhoods, and health. Aside from the direct effects on self-reported health of age, socioeconomic status, ethnicity, and insurance coverage, neighborhood characteristics can support or impede health-promoting behaviors, and exposure to environmental hazards has obvious health implications. The TCEQ has placed 10 local ZIP codes on its Air Pollutant Watch List, signaling particularly the high levels of toxic pollutants along the Houston Ship Channel.

In the KIHAS surveys, the respondents who lived in or adjacent to those locations were indeed significantly more likely than those living elsewhere to report that their own state of health was only fair or poor and to say that they were “very concerned” about the effects of air pollution on their family’s health. The relationship between health and residence remained statistically significant even after all differences by education, income, age, ethnicity, and insurance coverage were entered into the regression models. The data confirm the strong independent effect of environmental hazards on self-reported health, above and beyond the powerful relationship that also exists between area residents’ socioeconomic status and their exposure to pollution.

The HAHS respondents were even more worried about the health effects of poor water quality than of poor air quality. Half of the survey participants said they were “very concerned” about the “effects on your family’s health” of air pollution from industry and of smoke and fumes from cars and trucks, but almost three-fourths of all area residents expressed such concerns over toxic wastes leaking into streams and bayous.

Inadequate access to quality food also has obvious health implications. Despite Houston’s growing economy and low unemployment rates, the inequalities have only deepened in recent years: More than one-fourth of all the survey participants in the HAHS said they had difficulty during the past year paying for the groceries to feed their families.

Houston is undersupplied with supermarkets and oversupplied with “food deserts.” Inadequate access to healthy affordable food is clearly part of the reason why poverty and poor health are so strongly interconnected.

One-third of the HAHS respondents said they were “very satisfied” with their amount of physical activity. Almost half reported that, for at least several times a week during the past month, they had engaged “in enough vigorous activity (through sports or exercise) to work up a sweat.” Regression analyses indicate that engaging in vigorous physical activity is significantly associated with self-reported Health Disparities

33

good health. The strong correlation of income with both activity levels and self-rated health may be another reason why socioeconomic status is so powerfully and consistently related to health outcomes.

Walkable neighborhoods with easy access to local amenities also offer many health benefits compared to the car-dependent lifestyle more typical of this low-density, spread-out metropolitan area. Harris County residents who live inside the city limits of Houston are less likely than those in the suburbs to say that their neighborhood has plenty of safe and well-connected sidewalks, but they are more likely to say that shops or restaurants are within easy walking distance of their home and that they have ready access to public transit.

The annual KIHAS has found that increasing proportions of Harris County residents in recent years have said that they would prefer more walkable alternatives and would choose localities with a mix of developments over a single-family residential area “where you would need to drive almost everywhere you want to go.” The mayor’s recently formulated “Complete Streets and Transportation Plan” is designed to gradually transform Houston’s roadways into streets that will be safe, accessible and convenient not only for motorists, but also for public transit users, pedestrians, people of different abilities, and bicyclists.69 More access to alternative modes of transportation might well result in healthier lifestyles and lower obesity rates. In the 2013 KIHAS, more than 80 percent of all area workers said they drive alone when commuting to their jobs. Fewer than 8 percent said they make use of public transit at least once a week.

Given that an estimated two-thirds of all Harris County residents are either overweight or obese, Houston clearly needs to develop more opportunities for area residents to incorporate meaningful exercise into their daily routines. Some interesting efforts are under way to do so. The surveys indicate that Houston needs not only to provide more opportunities for physical exercise, but also to encourage many more area residents to take advantage of them.

34 Kinder Institute for Urban Research

Because the amount of stress people experience has powerful implications for health, their physical well-being may also be affected by the nature of the social relationships and support systems that their neighborhoods provide. The 2012 KIHAS developed composite measures of the respondents’ feelings of belonging and connectedness in their neighborhoods and the extent of their participation in neighborhood activities. Both of these separate measures were significantly associated with self-reported health, even after controlling for the impact of individual differences in age, gender, education, income, ethnic background, and insurance coverage.

The characteristics of the neighborhoods in which people live are clearly important forces in their own right that foster or discourage the social supports, stress reducers, and healthy behaviors that significantly contribute to physical well-being.

The survey findings reviewed in these pages make it clear that the overall health of Houston area residents is impacted by critical forces operating at many different levels. Individual attributes and inclinations, the quality of the physical environment, the social support systems that neighborhoods provide, the extent of access to affordable health insurance, and above all the stark socioeconomic disparities in an era of burgeoning inequality – all of these factors influence a person’s health in demonstrably significant ways, and all of them need to be taken into account in the ongoing efforts to improve the overall health of the Houston population. It will be important to continue to measure systematically the extent and nature of such improvements through subsequent Houston surveys in the years ahead.

APPENDIX A. THE 2012 HAHS METHODOLOGY The Survey Research Institute (SRI) at the University of Houston’s Hobby Center for Public Policy administered the telephone interviews. Using “back translation” and the reconciliation of discrepancies, the questionnaire was translated into Spanish, and bilingual supervisors and interviewers were assigned to the project at all times. A list of all the survey items included in the questionnaire and the distribution of responses given for each question can be found on the institute’s website at kinder.rice.edu/shea.

The SRI conducted the interviews between June 6 and July 17, 2012. Using random digit dialing (RDD), households were sampled in the Harris County area until the goal of 1,200 telephone interviews was achieved, with 840 (70 percent) of the respondents reached by landline and 360 (30 percent) by cell phone. In each household contacted by RDD, a household member aged 18 or over was randomly chosen as the eligible respondent, with initial preference given to adult males.

The response rates (indicating the number of completed interviews in relation to all potentially eligible phone numbers) were 19.1 percent for landlines and 14.5 percent for cell phones. Among all the eligible respondents who were identified and contacted, the survey’s cooperation rate (the ratio of completions to interviews plus refusals and break-offs) was 38.6 percent for respondents reached by landline and 27.6 percent for those on cell phones. Social Science Research Solutions (SSRS), the Philadelphia-based research firm, conducted the weighting process, utilizing the 2008-2010 three-year ACS estimates to correct for nonresponse and coverage biases in the sample, and assigning weights to each case in order to ensure that the final distributions were in close agreement with the actual Harris County distributions with respect to such known population parameters as race and ethnicity, age, gender, education level, home ownership, county population, and density.

The chart below shows the way the weights given to the survey responses resulted in changes that bring the sample figures into closer alignment with the actual distributions, as reported by the 2008-2010 ACS threeyear estimates for Harris County. The chart makes it clear that, despite the relatively low response rate, the demographic distributions in the survey sample are very close to the ACS figures. With the exception of a slight overrepresentation of respondents who were 60 and older and an underrepresentation of those aged 30 to 44, the proportions in the sample of males and females, of Anglos, blacks, Hispanics, and Asians, and of various levels of educational attainment are almost identical to the figures recorded in the most recent U.S. Census data. This alignment offers reassurance that the survey data are providing a representative picture of the Harris County population as a whole.

ACS 20082010 ACS three-year estimates

Survey Sample (N)

Survey Sample (%)

Weighted Sample (%)

Male

549

45.8%

49.0%

49.3%

Female

651

54.3

51

50.7

Non-Hispanic White

464

38.7%

39.0%

37.4%

African American

243

20.3

18.9

18.7

Hispanic, Latino

412

34.3

33.0

36.3

Asian

29

2.4

5.6

6.8

18-29

161

13.4%

21.2%

25.7%

30-44

200

16.7

21.4

31.0

45-59

356

29.7

27.5

26.2

60-98

483

40.3

29.8

17.1

74

6.2%

15.3%

High school diploma

215

17.9%

25.8%

45.7%*

Some college

375

31.3

29.9

26.8*

College degree

310

25.8

18.5

18.1*

Postgraduate

209

17.4

9.5

9.5*

Respondents who have children

867

72.3%

68.7%

--

Respondents who have children at home

295

24.6

29.2

--

Respondents who have children under the age of six

115

9.6

12.9

--

Gender:

(Adults only)

Race and Ethnicity:

Age:

Education: Less than high school

Parents:

*Note: The ACS statistics for educational attainment include only members of the Harris County population who are aged 25 years and over; the margin of error is +/- 0.2 to 0.3.

Health Disparities 42

Kinder Institute for Urban Research

35

APPENDIX B. LOCAL HEALTH ADVISORS Listed here, with deep gratitude, are the advisors and colleagues who helped the research team during 2011-2012 to develop the HAHS survey instrument and to assess the preliminary findings. We are especially grateful to Aileen Beltran, Rosemary Coffey, Justin Denney, Beverly Gor, Necole Irvin, Ken Janda, Rachel Kimbro, Elizabeth Love, Karen Love, John Mendelsohn, and the Health of Houston Survey Research Team at the University of Texas School of Public Health (Stephen Linder, Dritana Marko, Thomas Reynolds, and Jessica Tullar) for their helpful suggestions on earlier drafts of this report.

William B. Baun

Wellness Officer, The University of Texas MD Anderson Cancer Center

Charles E. Begley, Ph.D.

Co-Director, Professor of Management, Policy, and Community Health, Center for Health Services Research, University of Texas School of Public Health

Aileen M. Beltran, M.P.H.

Community Benefits Manager, Texas Childrenâ&#x20AC;&#x2122;s Hospital

Ronald R. Cookston, Ed.D.

Executive Director, Gateway to Care

Justin T. Denney, Ph.D.

Assistant Professor of Sociology, Associate Director of Urban Health Program, Kinder Institute for Urban Research, Rice University

Beverly J. Gor, Ed.D.

Project Director, CAN DO Houston

Necole S. Irvin, J.D., M.P.H.

Program Officer, Houston Endowment Inc.

Ken Janda, J.D.

President and CEO, Community Health Choice, Inc. (CHC)

Rachel T. Kimbro, Ph.D.

Associate Professor of Sociology, Director of Urban Health Program, Kinder Institute for Urban Research, Rice University

June Liu, M.P.H.

Quality Project Manager, Medical Affairs, Community Health Choice, Inc.

Karen H. Love Senior Vice President, Strategic Planning and Partnerships, Network Management, Community Health Choice, Inc. Roksan Okan-Vick

Executive Director, Houston Parks Board

Herminia Palacio, M.D., M.P.H.

Senior Program Officer, Robert Wood Johnson Foundation

Steven B. Schnee, Ph.D.

Executive Director, MHMRA of Harris County

Umair A. Shah, M.D., M.P.H.

Executive Director, Harris County Public Health and Environmental Services

Stephen L. Williams, M.Ed., M.P.A.

Executive Director, Houston Department of Health and Human Services

Stephen H. Linder, Ph.D. Dritana Mark, M.D. Thomas Reynolds, Ph.D. Jessica M. Tullar, Ph.D.

Institute for Health Policy, University of Texas School of Public Health

36 Kinder Institute for Urban Research

REFERENCES 1

The published reports on the Arts Survey and the Education Survey can be found on the Kinder Institute’s website at kinder. rice.edu/shea.

14

Moore, D. E., and Hayward, M.D. 1990. “Occupational careers and the mortality of elderly men.” Demography 27: 31-53.

15

Publications and descriptions of the KIHAS survey findings are available on the Kinder Institute Houston Area Survey website at kinder.rice.edu/has.

House, J. S., Wells, J. A., Landerman, L. R., McMichael, A. J., and Kaplan, B.H. 1980. “Occupational stress and health among factory workers.” Journal of Health and Social Behavior 20: 13960.

3

Idler, Ellen L., and Angel, Ronald J. 1990. “Self-Rated Health and Mortality in the NHANES-I Epidemiologic Follow-Up Study.” American Journal of Public Health 80(4):446-452.

16

4

17

2

Idler, Ellen L., and Benyamini, Yael. 1997. “Self-Rated Health and Mortality: A Review of Twenty-Seven Community Studies.”

Journal of Health and Social Behavior 38(3):21-37.

House, J. S., Stretcher, V., Metzner, H. L., and Robbins, C. 1986.

“Occupational stress and health in the Tecumseh Community Health Study.” Journal of Health and Social Behavior 27: 62-77.

Harper, C., Marcus, R., and Moore, K. 2003. “Enduring Poverty and the Conditions of Childhood: Life course and Intergenerational Poverty Transmissions.” Chronic Poverty and Development Policy 31(3):535-554.

5

“Health of Houston Survey 2010.” The University of Texas School of Public Health: Institute for Health Policy. (www. hhs2010.net) 6

Manderbacka, K., Lundberg, O. and Martikainen, P. 1999. “Do Risk Factors and Health Behaviors Contribute to Self-Ratings of Health?” Social Science & Medicine 48(12):1713-1720.

18

Williams, David R. 2005. “The Health of U.S. Racial and Ethnic Populations.” Journals of Gerontology 60(B):53-62.

19

National Center for Health Statistics, Centers for Disease Control and Prevention. 2007. “Heath, United States, 2007 with Chartbook on Trends in the Health of Americans.” Hyattsville, MD: DHHS Publication No. 2007-1232.

7

“County Health Rankings & Roadmaps.” University of Wisconsin Population Health Institute. (www.countyhealthrankings. org) 8

Lopez, Russ. 2004. “Income Inequality and Self-Rated Health in US Metropolitan Areas: A Multi-Level Analysis.” Social Science & Medicine 59:2409-2419.

9

Cutler, David M., and Lleras-Muney, Adriana. 2008. “Education and Health: Evaluating Theories and Evidence.” Pp. 29-60. Making Americans Healthier: Social and Economic Policy as Health Policy, edited by RF Schoeni, James S. House, George A. Kaplan, and H. Pollack. New York: Russell Sage Foundation.

10

Herd, P., Goesling, B., and House, J. S. 2007. “Socioeconomic Position and Health: The Differential Effects of Education versus Income on the Onset versus Progression of Health Problems.”

Journal of Health and Social Behavior 48:223-238.

11

Pampel, Fred C., Patrick M. Krueger, and Justin T. Denney.

2010. “Socioeconomic Disparities in Health Behaviors.” Annual Review of Sociology 36: 349-370.

12

Kutner, M., et. al. 2006. “The Health Literacy of America’s Adults: Results From the 2003 National Assessment of Adult Literacy.” U. S. Department of Education, Institute of Education Sciences: NCES 2006-483.

13

Cohen, S., Kaplan, G. A., and Salonen, J. T. 1999. “The role of psychological characteristics in the relation between socioeconomic status and perceived health.” Journal of Applied Social Psychology 29: 445-468.

20

Klineberg, Stephen L. 2013. “The 32nd Kinder Institute Houston Area Survey: Tracking Responses to the Economic and Demographic Transformations.” Kinder Institute for Urban Research, Rice University.

21

Link, Bruce G., and Phelan, J. 1995. “Social Conditions as Fundamental Causes of Disease.” Journal of Health and Social Behavior 35:80-94.

22

Link, Bruce G., and Phelan, J. 1995. “Social Conditions as Fundamental Causes of Disease.” Journal of Health and Social Behavior 35:80-94.

23

DeNavas-Walt, C., et al. Sep 2013. “Income, Poverty and Health Insurance Coverage in the United States: 2012.” U.S. Census Bureau, Current Population Reports: 60-245.

24

Feder, Judith, Larry Levitt, Ellen O’Brien, and Diane Rowland.

2001. “Covering the Low-Income Uninsured: The Case for Expanding Public Programs.” Health Affairs 20(1):27-39.

25

To learn more about the organizations that comprise Harris County’s healthcare safety net, please visit https://www.harrishealth.org/en/about-us/who-we-are/pages/safetynet.aspx 26

Begley, C. E., et al. 2012. “Health Reform and Primary Care Capacity: Evidence from Houston/Harris County, Texas.” Journal of Health Care for the Poor and Underserved 23(1):386-397.

Health Disparities

37

27

Howard, Matthew S., Davis, B. A., Anderson, C., Cherry, D., Koller, P., and Shelton, D. 2005. “Patients’ Perspective on Choosing the Emergency Department for Nonurgent Medical Care: A Qualitative Study Exploring One Reason for Overcrowding.”

Journal of Emergency Nursing 31(5):429-435.

28

Center for Health Statistics (CHS). Texas Behavioral Risk Factor Surveillance System Survey Data. Austin, Texas: Texas Department of State Health Services, 2010.

29

Sampson, R., Morenoff, J., and Gannon-Rowley, T. 2002.

“Assessing ‘neighborhood effects’: Social processes and new directions in research.” Annual Review of Sociology 28: 443-478.

30

Robert, S. A. 1999. “Socioeconomic position and health: The independent contribution of community socioeconomic context.” Annual Review of Sociology 25: 489-516.

31

Bethel, Heidi L., et al. 2006. “A Closer Look at Air Pollution in Houston: Identifying Priority Health Risks.” The 16th Annual International Emission Inventory Conference.

32

“Dirty Air: Houston, we have a problem.” The Economist, Jan 6th 2000.

33

“State of the Air.” American Lung Association. (www.stateoftheair.org)

40

Manon, M., and Treering, D. 2010. “Food for Every Child: The Need for More Supermarkets in Houston.” Philadelphia, PA: The Food Trust.

41

Treuhaft, S., and Karpyn, A. 2010. “The Grocery Gap: Who has Access to Healthy Food and Why it Matters.” PolicyLink & the Food Trust.

42

Manon, M., and Koprak, J. 2012. “Roadmap for Encouraging Grocery Development in Houston and Texas.” Philadelphia, PA: The Food Trust.

43

Food Access Research Atlas, USDA. (www.ers.usda.gov/data-products/food-access-research-atlas/go-to-the-atlas.aspx) 44

Pangchang, S. 2012. “A Guide on Local Agriculture for Houstonians.” Kinder Institute for Urban Research, Rice University. The report is downloadable at http://kinder.rice.edu/uploadedFiles/ Kinder_Institute_for_Urban_Research/Programs/UrbanHealthProgram/FM_FullReport.pdf.

45

The Healthy Living Matters Collaborative consists of a multi-sector group of local leaders, engaged in an initiative to curb childhood obesity in Harris County area by using policy actions to enact systemic and environmental change. For more details, please visit http://www.healthylivingmatters.net.

46

34

“AirData.” U.S. Environmental Protection Agency. Retrived June 25, 2013. (www.epa.gov/airdata)

Centers for Disease Control and Prevention. 2011. “Physical Activity and Health.” (http://www.cdc.gov/physicalactivity/everyone/health/)

35

47

Air toxic benzene has been designated as a human carcinogen by several agencies, such as the TCEQ, the U.S. Environmental Protection Agency, etc. Exposure to low levels of styrene has the potential to cause odor-related health effects, such as nausea and headaches.

Center for Health Statistics (CHS). Texas Behavioral Risk Factor Surveillance System Survey Data. Austin, Texas: Texas Department of State Health Services, 2010.

48

“Report on the Air Pollutant Watch List Areas in Texas,” Texas Commission on Environmental Quality, February 2012.

Klineberg, Stephen L. 2013. “The 32nd Kinder Institute Houston Area Survey: Tracking Responses to the Economic and Demographic Transformations.” Kinder Institute for Urban Research, Rice University.

37

49

36

The ten APWL ZIP codes are 77029, 77012, 77017, 77547, 77015, 77503, 77506, 77530, 77571, and 77520. The sixteen ZIP codes immediately adjacent to them are 77011, 77013, 77020, 77023, 77028, 77049, 77061, 77087, 77502, 77505, 77507, 77521, 77523, 77536, 77562 and 77587.

38

“National Drinking Water Database: City of Houston Public Works.” Environmental Working Group. (www.ewg.org/ tap-water/whatsinyourwater/TX/City-of-Houston-PublicWorks-/1010013/) 39

Houston-Galveston Area Council. 2013. “How’s the Water? 2013 Houston-Galveston Area Council Basin Highlights Report.” (http://videos.h-gac.com/CE/water/2013-Basin-Highlights-Report.pdf)

38 Kinder Institute for Urban Research

Lachapelle, U., and Frank, L.D. 2009. “Transit and Health: Mode of Transport, Employer-Sponsored Public Transit Pass Programs, and Physical Activity.” Journal of Public Health Policy 30(1):S73-S94.

50

American Community Survey Reports. 2011. “Commuting in the United States: 2009.” United States Census Bureau.

51

Houston Bikeway Program, City of Houston. (http://www. houstonbikeways.org/) 52

American Community Survey Reports. 2011. “Commuting in the United States: 2009.” United States Census Bureau.

53 Center for Health Statistics (CHS). Texas Behavioral Risk Factor Surveillance System Survey Data. Austin, Texas: Texas Department of State Health Services, 2010.

67 Subramanian, S. V., Kim, D. J., and Kawachi, Ichiro. 2002.

“Social Trust and Self-Rated Health in US Communities: A Multilevel Analysis.” Journal of Urban Health 79(4):521-534.

54

68

“Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults: The Evidence Report.” 1998. National Heart, Lung, and Blood Institute, National Institutes of Health Publication No. 98-4083.

55

“Performance in an Era of Uncertainty: 17th Annual Towers Watson/National Business Group on Health Employer Survey on Purchasing Value in Health Care.” 2012. Towers Watson.

Putnam, Robert D. 1993. “The prosperous community: Social capital and public life.” The American Prospect 4(13):35.

69

Overall the plan has seven major components. To read the executive order signed by the Mayor on November 1, 2013, please go to http://www.houstontx.gov/planning/docs_pdfs/Exec_Order_Complete_Streets.pdf.

56

Finkelstein, Eric A., and Kosa, K. M. 2003. “Use of Incentives to Motivate Healthy Behaviors among Employees.” Gender Issues 21(3):50-59.

57

Caloyeras, J. P., et. al. 2014. “Managing Manifest Diseases, But Not Health Risks, Saved PepsiCo Money over Seven Years,” Health Affairs 33(1): 124-131.

58

Kawachi, Ichiro, Kennedy, B. P. and Glass, R. 1999. “Social Capital and Self-Rated Health: A Contextual Analysis.” American Journal of Public Health 89(8):1187-1193.

59

Ross, C. E. 2000. “Neighborhood disadvantage and adult depression.” Journal of Health and Social Behavior 41(2):177-187.

60

Elliott, M. 2000. “The stress process in neighborhood context.”

Health Place 6(4):287-299.

61

Taylor, P., and Fry, R. August 2012. “The Rise of Residential Segregation by Income,” Pew Research Center.

62

Kneebone, E., Nadeau, C., and Berube, A. 2011. “The Re-Emergence of Concentrated Poverty: Metropolitan Trends in the 2000s.” Metropolitan Policy Program: Brookings Institute.

63

Cattell, Vicky. 2001. “Poor People, Poor Places, and Poor Health: The Mediating Role of Social Networks and Social Capital.” Social Science & Medicine 52:1501-1516.

64

Cleaver, Frances. 2005. “The Inequality of Social Capital and the Reproduction of Chronic Poverty.” World Development 33(6):893-906.

65

Ahmed, A. R., Mohammed, S. A., and Williams, D. R. 2007.

“Racial Discrimination & Health: Pathways & Evidence.” Indian Journal of Medical Research 126:318-327.

66

Williams, David R., and Collins, Chiquita. 2001. “Racial residential segregation: a fundamental cause of racial disparities in health.” Public Health Report 116(5): 404-416.

Health Disparities

39

MISSION: Rice Universityâ&#x20AC;&#x2122;s Kinder Institute for Urban Research conducts scientific research, sponsors educational programs, and engages in public outreach that advances understanding of pressing urban issues and fosters the development of more humane and sustainable cities.

TheThe 2012 wasmade madepossible possible 2012Houston Houston Education Area HealthSurvey Survey was byby

Rice University, MS 208 | 6100 S. Main, Houston, TX 77005 www.kinder.rice.edu | kinder@rice.edu