Frequently Asked Questions
Community Health Survey
WHAT IS THE CHS?
The New York City Community Health Survey (CHS) is a telephone survey conducted annually by the DOHMH, Division of Epidemiology, Bureau of Epidemiology Services. CHS provides robust data on the health of New Yorkers, including both neighborhood and citywide estimates on a broad range of chronic and behavioral risk factors.
CHS is based upon the CDC's Behavioral Risk Factor Surveillance System (BRFSS). The cross-sectional survey samples approximately 10,000 adults aged 18 and older from all five boroughs of New York City - Manhattan, Brooklyn, Queens, Bronx, and Staten Island. A computer-assisted telephone interviewing (CATI) system is used to collect survey data, and interviews are conducted in a variety of different languages. All data collected are self-reported.
The survey results are analyzed and disseminated in
order to influence health program decisions, to increase the understanding of
the relationship between health behavior and health status, and to support
health policies. For more information and public-use datasets, visit Community Health Survey.
back to top
WHEN WILL MORE RECENT CHS DATA BE AVAILABLE?
Survey data from 2002, 2003, 2004, 2005, 2006, and 2007 are currently available on the DOH website. CHS 2008 data will be available in the late summer of 2009.
back to top
HOW ARE CHS DATA DISSEMINATED AND USED?
The Community Health Survey is conducted in order to inform public health policy and practice in New York City, making data use and dissemination cornerstones of the project. The CHS data have three main roles in public health work:
- Instrumental:
Influencing health program decisions
- Conceptual:
Increasing the understanding of the relationship between health behavior and
health status
- Persuasive: Supporting health policies
Many data derived from the CHS are linked to objectives in the Healthy People 2010 initiative and Take Care New York. Assessment of progress toward meeting these objectives is carried out at the local, state, and the national levels. Using CHS findings for assessment provides policymakers with informed choices in public health decisions. The CHS findings also can assist planners in designing public health intervention strategies and evaluating their impact, such as the campaign to reduce the rate of cigarette smoking in NYC.
Disseminating the CHS findings is also an integral function. CHS data are an important source of information on current health status, behaviors, and lifestyle patterns, as well as outcome trends. The NYC DOHMH uses CHS data to prepare the Vital Signs Reports on various health topics and the Community Health Profiles on neighborhood-specific
health indicators. The CHS data also are used to create Fact Sheets, Issue
Briefs and recommendations for health providers and the public. For more
data resources, go to My Community's
Health.
back to top
HAVE OTHER CITIES AND STATES CONDUCTED HEALTH SURVEYS SIMILAR TO THE CHS?
The CHS, conducted by the NYC DOHMH, is very similar to a survey conducted by state health departments in conjunction with the U.S. Centers for Disease Control and Prevention (CDC) called the Behavioral Risk Factor Surveillance System (BRFSS).
All 50 states, including New York State, conduct surveys as part of BRFSS. However, because the BRFSS is statewide, the sample sizes for individual neighborhoods are too small to find neighborhood-level estimates. Some local and city health departments, such as the NYC DOHMH, have conducted their own independent versions of BRFSS, such as the Los Angeles County Health Department.
back to top
WHAT ARE SOME OF THE LIMITATIONS AND SOURCES OF ERROR IN CHS DATA?
Since the CHS is modeled after the BRFSS, a good source of information on error within the survey is the BRFSS User's Guide. Highlighted below are a few key limitations that should be considered when using CHS data.
- Sampling error occurs because estimates are based on
only a sample of the population rather than on the entire population. This
type of error can occur in even the most sophisticated sampling design with an
adequate sample size. Strict adherence to correct surveillance procedures,
however, can eliminate some causes of sampling error. In the CHS, one
avoidable cause of sampling error is associated with the random selection of
residential telephone numbers.
- The CHS does not sample from the entire population -
some groups cannot be reached using a basic telephone survey. Excluded groups
include people living in nonresidential settings, such as hospitals, nursing
homes, prisons, and college dormitories; and households without telephone
service. Persons without telephones tend to have lower household incomes, and
low income is associated with certain health risk behaviors. This means that
estimates based on the CHS and other telephone surveys may result in
underestimates of risk behavior for subpopulations in certain neighborhoods
and subgroups. In addition, some households in New York City are
cell-phone-only households, which were not included in the survey. There is no
easy solution to this type of error in the CHS. Although data can be weighted
to account for households without telephones, neighborhood-specific estimates
of households without a telephone are only available once every 10 years.
- The CHS relies on self-reported data on health risks,
which means that survey respondents answer the questions without any outside
check on the truth of their answers. This means that the accuracy of
respondents' answers depends on how well they understand the questions and
remember the information needed to answer the questions. Also, people may
answer a question the way they think that the interviewer wants them to
answer, instead of giving a truthful answer, or they may not want to tell the
interviewer the correct answer to a personal question so they provide false
information, instead of just not answering the question.
- Nonresponse error - the inability to obtain data for all questionnaire items from persons in the sample population - is a common problem in surveys. There are two levels of nonresponse: unit nonresponse and item nonresponse. Unit nonresponse occurs when an eligible sampling unit (i.e., a household or a person) does not respond or a respondent refuses to participate in the survey. Some persons may never be located after multiple call attempts have been made; others may be willing but unable to respond because of hearing problems, or other causes. Item nonresponse can occur when a respondent believes a question is too sensitive or the respondent does not know or cannot recall the answer to a question.
Poststratification is the final adjustment made to the survey weights. This adjustment forces the weighted totals within categories of age and race to be distributed consistent with current estimates of the adult population in New York, based on the 2000 US Census data. Poststratification is performed to reduce potential bias due to differential response and lack of telephone coverage, thereby providing more accurate estimates that are consistent with other statistics.
back to top
WHAT IF I WANT CHS DATA FOR PERSONS UNDER THE AGE OF 18?
While the CHS only sampled New Yorkers aged 18 and older, data on adolescent New Yorkers in grades 9 through 12 in public schools are available from the Youth Risk Behavior Survey (YRBS).
back to top
CAN I COMPARE NEW YORK CITY CHS RESULTS TO OTHER STATES AND THE COUNTRY AS A WHOLE?
For certain variables, yes. The easiest way is to go the CDC BRFSS website and run a query for the health outcome and state of interest, and then compare the results to those from the CHS. However, care must be exercised when making comparisons. It is important to consider whether the questions were asked the same way on the surveys being compared and whether or not the populations being asked the questions were the same (e.g., aged 18-64 in one survey and aged 20-60 in another). Also, it is important to note whether the rates are adjusted to the same standard population. If not, you may want to compare crude or unadjusted estimates.
back to top
HOW DO YOU PROTECT THE CONFIDENTIALITY OF THE INFORMATION OBTAINED FROM THE PEOPLE INTERVIEWED?
Respondents to the CHS are anonymous. A computer generates random telephone numbers to be called, which are then discarded when the survey is complete. Thus, a respondent's telephone number is never linked to his/her survey responses.
back to top
HOW ARE NEIGHBORHOODS DEFINED IN THE CHS?
The CHS used a stratified sample in order to produce citywide as well as neighborhood-specific estimates. Neighborhoods were defined using the United Hospital Fund's (UHF) designation. UHFs were chosen because survey respondents usually know their zip codes, and UHFs are the accepted zip-code-defined neighborhoods, as opposed to Community Districts, which do not follow zip code areas. There are a total of 42 UHF neighborhoods in NYC each defined by several adjoining zip codes. In order to increase statistical power, several neighborhoods were combined in 2002 and in 2003. In more recent years, separate estimates are available for all 42 neighborhoods.
back to top
ARE THE QUESTIONS THE SAME ON THE CHS EVERY YEAR?
The majority of the questions remained consistent from year to year, because the CHS is first and foremost a surveillance tool – we track indicators over time. However, there have been changes between years. All questionnaires are available in PDF format for comparison here.
back to top
HOW DO I EXAMINE TRENDS OVER TIME IN CHS INDICATORS?
Not all questions are asked each
year, although a basic core of items is in every CHS. For specific questions,
users will need to check the questionnaire for each year, or contact the Bureau
of Epidemiology Services at survey@health.nyc.gov to determine how to best to compare survey measures across multiple years.
back to top
CAN I DOWNLOAD THE CHS DATASET FOR ANALYSIS USING MY OWN
STATISTICAL SOFTWARE?
CHS data is available for download on
the NYC DOHMH website, with SAS programs for running analyses. This dataset
is somewhat restricted for privacy reasons, e.g. only five age-groups are
available. Questions about the data or running analyses should be sent to
survey@health.nyc.gov.
back to top
WHAT IS THE DIFFERENCE BETWEEN THE SURVEY POPULATION SIZE AND THE SURVEY SAMPLE SIZE?
The survey population size refers to the size of the population that is eligible to be included in the survey (e.g., for the CHS, all non-institutionalized New Yorkers age 18 and older who live in households with landline telephone service). The sample size refers to the size of the population that is actually contacted by phone and surveyed.
back to top
WHAT IF I WANT TO KNOW MORE ABOUT SURVEY DESIGN?
We suggest the following readings for persons who are interested in learning more about survey design.
- Glossary of Survey Terms (CDC)
- CDC Training Module on Prevalence and Surveys
- Bureau of the Census. Statistical Brief: Phoneless in America. Washington,
DC: Bureau of the Census, 1994.
- Centers for Disease Control and Prevention. Health Risks in America: Gaining Insight from the Behavioral Risk Factor Surveillance System. Atlanta, GA: U.S. Department of Health
and Human Services, Public Health Service, Centers for Disease Control and
Prevention, 1997.
- Fowler FJ Jr. Survey Research Methods. Newbury Park, CA: Sage Publications, 1988.
- Marks JS, Hogelin GC, Gentry EM, Jones JT, Gaines KL, Forman MR, Trowbridge FL. The behavioral risk factor surveys: I. State-specific prevalence estimates of behavioral risk factors. Am J Prev Med 1985;1(6):1-8.
- Remington PL, Smith MY, Williamson DF, Anda RF, Gentry EM, Hogelin GC. Design, characteristics, and usefulness of state-based behavioral risk factor surveillance 1981-1986. Public Health Rep 1988;103:366-375.
- Siegel PZ, Frazier EL, Mariolis P, Brackbill RM, Smith C. Behavioral risk factor surveillance, 1991: monitoring progress toward the nation's year 2000 health objectives. MMWR
1993;42(SS-4):1-22.
- Waksberg J. Sampling methods for random digit dialing. J Am Stat Assoc 1978;73:40-46.
- Weeks MF, Kulka RA, Lessler JT, Whitmore RW. Personal versus telephone surveys for collecting household health data at the local level. Am J Public Health 1983;73:1389-1394.
back to top
Page last updated April
2009.
