National Health Interview Survey (NHIS) Geocodes
The National Health Interview Survey (NHIS) is representative at the national and Census Region level. Each public use annual file is representative of the Nation and includes a variable REGION which indicates the four Census Regions (aggregates of states). The public use annual files and documentation are available at Questionnaires, Datasets, and Related Documentation.
Lower levels of geography are often restricted because they pose a disclosure risk. However, these lower levels can be useful in research for various reasons, such as
- Preparing estimates at lower levels (Examples of using NHIS for state estimates: State, Regional, and National Estimates of Health Insurance Coverage)pdf icon
- Controlling for geographic differences
- Merging in additional (contextual, political, etc.) information from other data sources.
When proposing to use these data, it is important to indicate
- How you intend to use these data
- The variable name/column location, description, and file from which the data come. The chart below indicates the variable source, name, and description for the purpose of helping you write the RDC proposal.
There are various sources of geographic data for NHIS depending on the years of interest. It is the researcher’s responsibility to investigate the possible strengths and limitations to the different sources and choose the one that best suits his/her project. It also is important for the researcher to understand the relationships between different NHIS annual files and the reference values of the geographic data that are available.
- In-House Files – These are the NHIS data files that are used by NCHS to create the NHIS public use data files that are released annually. These files contain everything in the public use data files, along with information such as geographic area detail (e.g., state), metropolitan area detail, an urban/rural code, and other items such as detailed industry and occupation codes that are not provided in the NHIS public use data files.
- HUD Geocode Files (1986-Present NHIS ) – In response to demand for geocode information for recent National Health Interview Survey (NHIS) data files, NCHS established an agreement with the Department of Housing and Urban Development (HUD) for HUD to geocode NHIS addresses. The output from the HUD GSC geocoding process is at the NHIS Household File level and includes the Census block, block group, tract, county, state, latitude, longitude, and other geographic data, along with return codes indicating the reliability of the geographic codes assigned by the HUD GSC geocoding process.
- 1986 – 2016 NHIS HUD Geocode Documentation (2000 Census geocodes)pdf icon
- 1986 – present NHIS HUD Geocode Documentation (2010 Census geocodes)pdf icon
- When requesting levels of geography below county, for the years 2006-2015, it is recommended that you request access to both the UCF geocodes and the HUD geocodes. The geocodes in the 2006-2015 Census Unit Control Files are based on 2000 decennial census geography and are high quality. For the area frame portion of the NHIS sample, state, county, tract, block group, and block are available from the Census Unit Control File. NCHS recommends that these codes be used if there is disagreement with the HUD geocodes. The geocodes in the Census Unit Control File consist only of state and county for the permit frame portion of the NHIS sample; the HUD geocodes, when present, are the only available source for tract, block group, and block.
- A separate set of HUD Geocode Files also is available for 1986-Present NHIS with 2000 Census geocodes and 2010 Census geocodes. These geocode files are at the NHIS Person File level. They contain the variables described above from the HUD GSC geocoding process (Census block, block group, tract, county, state, latitude, longitude, and return codes indicating the reliability of the geographic codes assigned by the HUD GSC geocoding process). They do not contain other auxiliary variables described in the documentation that is available above.
- Special Project Geocode File – It includes state, county, tract, block group, and block codes for Census 1990 and Census 2000 for NHIS households from 1995-2005. There are some missing values for tract, block group, and block. There also is a 1985-1994 geocode file that provides Census 1990 codes (tract/block numbering area and block group) for 1985-1994 NHIS households. This file does not contain any 1990 geocodes for some 1985-1994 NHIS households. No evaluation has been done to compare the quality of the HUD geocodes with the previous geocodes.
- Unit-Control File and Master Segment Tape File – The Unit control File (UCF) includes additional metropolitan area detail for the entire NHIS sample. It also includes lower levels of geography for the “area frame” portion of the NHIS sample, specifically census tract and census block information and therefore only includes geography for 85-90% of survey respondents. (For more information about the “area frame”, consult the Procedures and Methodology document.
- Master Segment Tape File The 1973-1994 precursor to the Unit Control File, the Master Segment Tape File, contains similar information. For the 1973-1978 NHIS data files, state and county must be obtained from the Master Segment Tape files, these variables are not available on the In-House Files prior to 1979. No resource of this type exists for the 1963-1972 period.
- 1963-1972 PSU_State_SMSA File – The only resource available for 1963-1972 to merge sub-Census Region geography (state and 1960 metropolitan area codes only) to NHIS data files.
Reference Values of the Geographic Data
- For a given NHIS sample design period, e.g., 1995-2005, the geographic data on the NHIS In-House Files and Unit Control Files such as state, county, census tract, etc. come from the decennial census prior to the beginning of the sample design period.
- For example, the reference value for the 1995-2005 NHIS data is 1990 (from the 1990 Census).
- Please refer to the NHIS Methods for more information about the NHIS sample design periods for which microdata are available, which are: 1963-72, 1973-84, 1985-94, 1995-2005 2006-present..
- Metropolitan area detail also is fixed for a given NHIS sample design period.
- For the 2016-present NHIS data, the metropolitan area definitions are those that became effective in February 2013.
- For the 2006-2015 NHIS data, the metropolitan area definitions are those that became effective in June 2003.
- For the 1995-2005 NHIS data, the metropolitan area definitions are those that became effective in June 1993.
- For the 1985-1994 NHIS data, the metropolitan area definitions are those that became effective in June 1983.
- For the 1973-1984 NHIS data, the metropolitan area definitions are those that became effective in February 1971.
- For the 1963-1972 NHIS data, the metropolitan area definitions are those that became effective in November 1960.
National Health Interview Survey (NHIS) Geocodes chart