Linking IPUMS versions of NHIS and MEPS with restricted versions in the FSRDC
What is the FSRDC?
- Federal Statistical Research Data Centers (FSRDCs) are partnerships between federal statistical agencies and research institutions. At these secure environments, authorized researchers may access restricted-use microdata for statistical purposes.
- Approved researchers can access data from the Agency for Healthcare Research and Quality (AHRQ), Bureau of Labor Statistics (BLS), Census Bureau, and National Center for Health Statistics (NCHS).
Why use the IPUMS versions of NHIS and MEPS in the FSRDC?
- Take advantage of the data harmonization and cleaning already done by IPUMS Health Surveys by bringing your IPUMS extract into the FSRDC.
- Browse our searchable web-based documentation system for available variables, recommended sampling weights, and analytic considerations.
- Access versions of the data with consistently named and coded variables across all available years of data.
- Select only the variables and years of data that you need for your analysis using an intuitive web "shopping" interface.
- Download the data in your preferred format: Stata, SAS, SPSS, R, fixed width (.dat), or comma-separated values (.csv).
- Record your selections in your IPUMS account to make it easy to quickly revise your extract to add/drop variables and years.
- IPUMS has identified variables common to NHIS and MEPS between the datasets (as well as within each dataset) and used the same variable names and coding schemes in the integrated versions of both datasets. The list of shared variables is available here.
- Familizarize yourself with MEPS data before accessing restricted versions of the data. MEPS data are complex and there is not internet access in the FSRDC. Explore IPUMS MEPS documentation and/or contact our user support with questions outside of the FSRDC.
- Get a jump start on the portions of your analysis that utilize public-use versions of the data. Write and troubleshoot code to construct your unique analytical variables and prepare analyses while you have access to resources not available inside the FSRDC (e.g., internet), allowing you to maximize your time in the FSRDC.
What information is in the Restricted NHIS and MEPS Files?
- Finer geographic detail
- Country of birth and related immigration data
- State and year of birth
- Detailed race and Hispanic origin data
- Exact dates of events
- For more information, visit the NCHS/CDC page on accessing NHIS in the FSRDC.
- Link files to NHIS
- Fully specified ICD-9 codes
- Fully specified industry and occupation codes
- Finer geographic detail
- Federal and state marginal tax rates
- Confidential data files (Medical Provider Component, Household Component-Insurance Component Linked File)
- For more information, visit the AHRQ page on MEPS data available in the FSRDC.
Note that if you are requesting the linking keys for connecting MEPS and NHIS data, your application may need to address requirements for both data providers (AHRQ and NCHS). This page is focused on FSRDC access requests for linking NHIS and MEPS, using IPUMS versions of one or both datasets. If you are not interested in linking the NHIS and MEPS data, but rather linking only the IPUMS and restricted access versions of a single dataset, see the respective instructions for NHIS and MEPS.
How do I talk about IPUMS data in the proposal?
- IPUMS provides harmonized versions of the NHIS data collected by the National Center for Health Statistics (NCHS) and MEPS data collected by the Agency for Healthcare Research (AHRQ). IPUMS does not collect these data. IPUMS versions of data should be listed as an external or user-supplied data source.
- You will be required to submit a data dictionary as part of your proposal; highlight the variables necessary to merge IPUMS and restricted access versions of each dataset in this section of the proposal.
- Remember that IPUMS variable names are not necessarily the same as original NHIS or MEPS variable names. For NHIS data, you may compare variable names between IPUMS and NCHS versions using the concordance tool.
- The NHIS linking keys that allow users to associate MEPS participants with their NHIS records are part of the restricted access MEPS data. In addition to requesting that AHRQ include the NHIS linking keys on your MEPS data, you will also need to provide linking information for connecting IPUMS MEPS to restricted MEPS data and IPUMS NHIS to restricted NHIS data.
- To combine IPUMS versions of the data with restricted access versions, you will need to provide NCHS and AHRQ with the IPUMS data file(s) you plan to bring into the FSRDC; NCHS requires that user-provided data be SAS-formatted data files.
- Here is more information on converting data formats between SAS and other statistical packages.
- For NHIS data requests, an FSRDC analyst for NCHS will merge the restricted variables and any identifiers onto your data file for you and have it uploaded to your FSRDC project space.
- When noting data sources in table shells in your proposal, remember that the restricted access NCHS and AHRQ versions of the data are different than the IPUMS versions.
Sample Proposal Language
Sample language for including IPUMS NHIS as a non-NCHS data:
User-supplied data file: IPUMS Health Surveys NHIS (IPUMS NHIS). This file will include data from the harmonized form of NHIS data that is produced by IPUMS. We will use IPUMS NHIS variables that are based on NHIS household, family, person, sample adult, and sample child files. The IPUMS NHIS contains NHIS public-use-file identification number (NHISPID) that will be used to merge the file to restricted NHIS data. The file delivered to NCHS will be called ipums_nhis_data.sas7bdat.
Sample language for including IPUMS MEPS as user-supplied data:
User-supplied data file: IPUMS Health Surveys MEPS (IPUMS MEPS). This file will include data from the harmonized form of MEPS data that is produced by IPUMS. We will use IPUMS MEPS variables from the following questionnaire sections: Access to Care, Preventive Care, Charge Payment, and Health Status. The IPUMS MEPS contains the MEPS public-use-file identification number (MEPSID) that will be used to merge the file to restricted MEPS data. The file we will submit containing IPUMS data will be called ipums_meps_data.sas7bdat.
Note that this sample language clearly includes:
- Identification of IPUMS as a harmonized version of the original data from NCHS or AHRQ.
- The name of the files that contain IPUMS data to be merged with restricted data.
- Files to be merged with restricted NHIS data are in a SAS format.
- The name of the original NHIS and MEPS questionnaire sections from which the harmonized IPUMS variables are drawn. For IPUMS NHIS, this information is available on the "NHIS Sources" tab of IPUMS variable pages (pictured below).
Sample language for how to merge IPUMS data with restricted data:
IPUMS NHIS merge to NHIS Restricted: Our public use file, which will contain harmonized variables, will be merged to the restricted NHIS file at the person level for each year in 2003-forward. The year-specific linking key can be constructed by concatenating HHX, FMX, and FPX; these variables are included in both the IPUMS and NCHS versions of the data. The NCHS analyst will need to merge the restricted NHIS data with the IPUMS NHIS data one year at a time by this linking key. IPUMS NHIS data will include all years in a single file; the IPUMS variable YEAR reports the survey year.
Sample language for how to merge IPUMS MEPS data with restricted MEPS data:
IPUMS MEPS merge to MEPS Restricted: Our public use file, which will contain harmonized variables, will be merged to the restricted MEPS file at the person level for each year using the IPUMS variable MEPSID. The linking key can be constructed by concatenating DUID, PID, and PANEL in the restricted MEPS data from AHRQ; this key can be used to merge the restricted MEPS data with the public use data one year at a time by the IPUMS variable MEPSID. IPUMS MEPS data will include all years in a single file; the IPUMS variable YEAR reports the Full-Year Consolidated File survey year.
Merging MEPS and NHIS data: We are also requesting the NHIS linking keys from AHRQ to link NHIS to MEPS; the linkage file description can be used to associate the MEPS sample person ID (DUPERSID in original AHRQ data) to the corresponding NHIS person ID.
Note that this sample language clearly includes:
- The name(s) of the appropriate linking keys in the IPUMS and restricted data files for the years of interest. Information on linking keys for all years is available here for NHIS.
- Instructions for appropriate merging of IPUMS and restricted data within each dataset.
- Reference to the request for linking keys to merge MEPS and NHIS data.
Gaining FSRDC access
- Researchers interested in utilizing restricted access linked NHIS-MEPS data in an FSRDC will need to submit a proposal and obtain Census Bureau Special Sworn Status before gaining access. The process of submitting a proposal and acquiring Special Sworn Status takes time (a minimum of 6 months), so get started early.
- Researchers interested in using restricted NHIS data (including NHIS data linked to MEPS data) in an FSRDC should note that there must be an FSRDC administrator on-site to use the data; not all FSRDCs have a full time administrator. Contact your local FSRDC administrator for more information.
- Visit the NHIS and MEPS pages on using restricted access versions of these data in the FSRDC.
- Fees associated with accessing data in the FSRDC vary, but may include:
- NCHS data management fees
- AHRQ data management fees
- FSRDC access fees (depending on the FSRDC and your institutional affiliation, there may be additional fees for daily use of the FSRDC)
- For more information on fees, visit the NCHS/CDC page on Fees and Invoicing for FSRDC access.
- List of existing FSRDCs