Dataset: The Health of Houston Survey 2010

Abstract

The Health of Houston Survey 2010 (HHS2010) is the first household survey to provide communities with information about the health of people living in the area of Harris County and City of Houston (Houston area). The survey supports the efforts of health agencies, service providers, and community organizations to have more accurate and up-to-date health information about Houston communities. The goal of the survey is to create a reliable, efficient and flexible infrastructure for a biennial survey to track emerging health issues, assess the impact of health programs, and document health improvements in valid and fully-comparable ways in the Houston area. HHS 2010 collects extensive information for multiple segments of the population on health status, conditions, behaviors, insurance coverage, and access. The survey provides health and sociodemographic information about one adult and one child in the case of households with children. Within each sampled household, we randomly chose an adult to interview. If the selected adult was the parent, guardian or caregiver of one or more children in the household, the adult provided health information regarding a randomly selected child.

The HHS 2010 sample is representative of Harris County and the City of Houston's non-institutionalized population living in households. The HHS2010 was designed to capture reliable data for a number of subpopulations:

· White, African American, Hispanic, Vietnamese, and Other Asian populations
· A standard range of age and income cohorts
· The total population of Harris County and the City of Houston

Using an addressed-based sample, over 30,000 residents received invitations to participate in the HHS2010. The survey was administered using three modalities. These were computer-assisted telephone interviews (CATI), web-based and mailed questionnaires. The survey was available in English, Spanish and Vietnamese. Certain populations were oversampled using a disproportionate stratified design. These populations included Vietnamese, Asian and African-American populations. The primary strata were the 1-percent PUMAs, and the secondary strata clusters were Census block groups.

To download study documentation, once you have entered Nesstar, click the disc icon download, on the top right side of the screen. To download the data please fill out a standard public use file data agreement form (www.healthofhouston.org) and send it to HealthofHouston@uth.tmc.edu. Study documentation/codebook can be downloaded prior to signing the PUDF agreement. If you would like to receive updates and notices about dataset updates, please register at the survey website.

Variable Groups

Document Description

Full Title

The Health of Houston Survey 2010

Identification Number

HHS2010

Authoring Entity

Institute for Health Policy, University of Texas School of Public Health , University of Texas Health Science Center Houston

Producer

Name Affiliation Abbreviation Role
Institute for Health Policy, University of Texas School of Public Health University of Texas Health Science Center Houston IHP, UTSPH

Copyright

Copyright © Institute for Health Policy, University of Texas School of Public Health, 2011

Date of Production

2011-09-28

Depositor

Name Affiliation Abbreviation

Full Title

HHS2010_F1

Depositor

Name Affiliation Abbreviation

Study Description

Full Title

The Health of Houston Survey 2010

Identification Number

HHS2010

Authoring Entity

Name Affiliation
Institute for Health Policy, University of Texas School of Public Health University of Texas Health Science Center Houston

Producer

Name Affiliation Abbreviation Role
Institute for Health Policy, University of Texas School of Public Health University of Texas Health Science Center Houston IHP, UTSPH
Social Science Research Solutions AUS and ICR SSRS Contracted company served as collaborator in questionnaire development and translation, data collection, data cleaning and processing and weighting.

Date of Production

2011-09-28

Funding Agency/Sponsor

Name Abbreviation Role Grant
Houston Endowment, Inc. Health of Houston Survey 2010
The Center for Clinical and Translational Sciences of the University of Texas Health Science Center Houston CCTS Support for outreach efforts

Data Distributor

Name Affiliation Abbreviation
Institute for Health Policy, University of Texas School of Public Health University of Texas Health Science Center Houston IHP, UTSPH

Contact Person

Name Affiliation E-mail address Universal Resource Identifier
HHS 2010 Support Support healthsurvey2010@uth.tmc.edu http://www.sph.uth.tmc.edu/hhs2010/

Depositor

Name Affiliation Abbreviation

Version

Public Use Data File (PUDF) 1.0

List of Keywords

Behavioral Risk Factors I (Smoking, Second Hand Smoke, Alcohol Abuse), Behavioral Risk Factors II (Diet, Physical Activity), Colorectal Cancer, Demographics I (Age, Gender, Race/Ethnicity), Demographics II (Country of Origin, Languages Spoken at Home, Citizenship), Economic Hardship, Employment, Environmental Risks, General Health Status, Health and Dental Care Access, Health and Dental Insurance Status, Health Conditions (Obesity, Diabetes, Asthma, Cancer, Cardiovascular Disease, Hypertension), Household Phone Status, Income, Interpersonal Violence, Mammography, Mental Health Access and Utilization, Mental Health Assessment, Neighborhood & Environment & Housing, Pap Test, Prenatal Care/Breastfeeding, Prior Awareness of Health of Houston Survey, Public Programs (Food Stamps, Supplemental Security Income, Social Security/Pensions, WIC, Child Support), Sedentary Behavior, Sexual Identity/Orientation, Social Cohesion, Social Support, Transportation

Topic Classification

Health

Country

United States of America  (USA)

Geographic Coverage

The geographic coverage was Harris County, Texas and those portions of the City of Houston that exist in other neighboring counties.

Unit of Analysis

The units of analyses include individuals, children ages 0-17 and adults 18 and older, and households.

Universe

The universe includes individuals of all ages.

Time Method

The data were collected from the fall of 2010 to the spring of 2011.

Sampling Procedure

The HHS2010 was designed to have a representative number of residents of the Harris County and the City of Houston area based on the U.S. Census 1-percent Public Use Microdata Areas or SuperPUMAs. Certain populations were oversampled using a disproportionate stratified design. These populations included Vietnamese, Asian and African-American residents. The primary strata were the 1-percent PUMAs, and the secondary strata clusters were Census block groups. In order to capture cell phone-only households in the sample frame for the HHS 2010, the decision was made to utilize an address-based sample (AB sample) for the survey. The AB sample captures households with landline phones, cell phone-only households and non-telephone households. The AB sampling started with generating a file of all Harris County and City of Houston residential addresses currently in use based on the United States Postal Service Delivery Sequence File (DSF). That address file was run against databases from InfoUSA, Experian, TargusInfo, and Acxiom that include all listed landline telephone numbers in the state to identify addresses with a listed telephone number. In order to facilitate the fielding of the survey, the final AB sample was divided into two segments: addresses with a listed landline telephone number and addresses without a listed landline telephone number. The overall sampling design contained a number of features across several dimensions that can be described in terms of sample stratification, household selection criteria, and within household selection criteria. 1) Sample stratification · Set interview targets per Super Public Use Microdata Areas (SuperPUMAs). · Within SuperPUMA creation of strata of addresses by listed Vietnamese and Asian-non-Vietnamese surnames. · Stratification of residual (households without an Asian surname) households by Census block group aggregate incidence of Hispanic, percent African American, and percent Asian. 2) Household-level selection · Screening households with respondents under 18 years of age. · If the person on the phone is younger than 18, interviewer asks for another household member who is 18 or older. · If there is no household member 18 or older, the household is not eligible, and the interview is terminated. · Screening households where every adult was age 65 and older. · If the household contained only adults ages 65 and older, the interview was terminated in 33 percent of such instances. That was designed to balance for the fact that such households more readily respond to surveys compared to other households. 3) Individual-level selection · Respondent is randomly selected from all household members using the "Rizzo" method of selection. · First, the number of people in household is determined. · If it is a single-person household, that person is the respondent. · If it is a two-person household, one of those two people is randomly selected to be the respondent. · If it is a three or more-person household, a random selection of household members is performed by the Web/CATI program. If the current respondent is selected, he or she is the respondent. If another household member is selected, ask for the household member, other than the current respondent, with the most recent birthday. Sample Stratification The number of interviews by SuperPUMA was set to ensure adequate statistical power within each. Stratification by racial/ethnic surname and aggregate Census block group incidence of minority households was done to maximize the number of interviews of African Americans, Asians, and specifically Vietnamese, while maintaining an acceptable number of interviews of both Caucasians and Hispanics. Census block groups were defined as being high Hispanic if 50 percent or more households were Hispanic; high African American if 50 percent or more African American; high Vietnamese if 10 percent or more Vietnamese, and high Asian if Asian-non-Vietnamese incidence was 15 percent or higher.

Mode of Data Collection

Data collection relied on three interview modes: telephone (CATI), web, and mail. The survey options were explained to those sample members in advance letters and reminder letters. Advance letters and reminder letters in three languages were mailed to all in the sample, offering the options of telephone and web survey models. In addition, sample for which listed telephone numbers could be obtained, traditional telephone interviewing methods are used as well. The specific steps for the data collection process were as follows: 1) Advance letters in three languages were sent to all households. The advance letter invited the household to participate in the study and offered the option of calling in to the survey center using a toll-free telephone number or completing a web-based survey. Unmatched sample also had the option of sending their phone number by filling out a postcard that was sent with the advance letter. Letters for AB sample with a listed telephone number also notified people that they would be receiving a call in the next few weeks to complete the survey. Advance letters included a $2 pre-incentive. 2) Telephone interviews were attempted with all households for which we had a telephone number. The initial calls commenced one week after the mailing of the advance letters. 3) Reminder notices were sent to all non-responding households. 4) A final reminder notice was sent to all non-responding households. A copy of the mail questionnaire was included in this final reminder notice. All of the letters and reminder postcards included a 1-800 toll-free number that the respondent could call for additional information on the survey or to complete the survey by telephone.

Weighting

Important note: two survey-related weights are provided. It is important that these weights are used correctly when using the "tabulation" and "analysis" capabilities of Nesstar. A survey weight is a value assigned to each case in the data file and is used to make statistics computed from the data more representative of the population. Survey data are weighted to adjust for differential sampling probabilities, to reduce any biases that may arise because of differences between respondents and non-respondents (i.e., nonresponse bias), and to address gaps in coverage in the survey frame (i.e., coverage bias). Survey weights, when properly applied in surveys can reduce the effect of nonresponse and coverage gaps on the reliability of the survey results. The weight named "Population weight" (pop_wgt) is to be used when frequencies or percentages that reflect the overall population are desired. This weight is best used in tabulation analyses. It is not to be used when performing correlation and regression analyses. The weight named "Person weight, normalized" (norm_wgt) is to be used when performing correlation and regression analyses. This is a normalized weight, which, when applied, totals the same number of respondents as the sample. The methodology used for calculating population and person weights is described in this PDF: http://hhs2010.sph.uth.tmc.edu/methods/Methodology_Report_HHS2010_Weights.pdf . For more information about survey weighting and the correct application of weights, see http://help.pop.psu.edu/help-by-statistical-method/weighting/Introduction%20to%20survey%20weights%20pri%20version.ppt/view .

Related Materials

Computer-assisted telephone interview (CATI) and Web survey questionnaires.

This is the questionnaire used by telephone interviewers and the programmers for the Web-based survey.

Mail questionnaire.

This is the English language mail questionnaire.

Health of Houston Survey 28 ZIP Code aggregation areas used for neighborhood analyses

This file contains two worksheets that detail which ZIP Codes were aggregated into the 28 areas used by the Health of Houston Survey for neighborhood analyses. One worksheet can be used in a GIS or mapping system to label the areas. The second worksheet is designed to be printed into hardcopy.

Health of Houston Survey 28 ZIP Code aggregation areas GIS shapefiles

This compressed (ZIP) file contains two shapefiles that show the ZIP28 areas used by the Health of Houston Survey for neighborhood analyses. One shapefile is a boundary (or polygon) file that shows the outlines of all 28 areas. The other shapefile is a point file (centroids). These files are optimizied for use with ESRI ArcMap products. If you are using other GIS software, please see their instructions for importing ESRI shapefiles.

Data Files Description

File Name

HHS2010_F1.NSDstat

Overall Case Count

6494

Overall Variable Count

201

Type of File

Nesstar 200801

Download

Metadata Index

This is the Metadata Index for a Nesstar Server.
Nesstar is a tool used for analysing, visualising and downloading datasets.

Click the "Explore Dataset" button to open the dataset.