Board of Scientific Counselors - Presentations
September 17-18, 2007
Back to September 2007 Presentations
NCHS Mortality Linkage Program
PowerPoint PresentationCdc-ppt
Slide 1
NCHS Mortality Linkage Program
- Designed to maximize the scientific value of the NCHS population-based surveys
- Enables researchers to examine the association of a variety of health factors with mortality, using the richness of NCHS health surveys
Slide 2
Background: Linked Mortality
Public-use Data Release
- Risk of respondent re-identification prevented public-use release of linked mortality data files
- Based upon BSC recommendation, conducted disclosure risk analyses
- Developed data perturbation strategy to protect respondent identity, while maintaining statistical validity
NOTES: Due to confidentiality protections of the NHIS participants, the NHIS Linked Mortality Files were made available only through the NCHS Research Data Center (RDC). Recognizing that this would place restrictions on researchers’ use of previously highly utilized public-use data files, NCHS has developed a data perturbation plan to allow for the continued public-use release of linked mortality data. The data release plan was developed to eliminate re-identification of survey participants and maximize the amount of data to be included in the release, and limit the amount of synthetic data introduced to the data file.
Slide 3
Summary NCHS Linked Mortality Data Files
Slide 4
Supporting Documentation for Linked Mortality Data
- File layouts
- Matching methodology reports
- Analytic guidelines
- Comparative analysis of the public-use and restricted-use data
- Guide to weighting and variance estimation (NHIS)
- Sample input statements for SAS & STATA for Public-use Linked Mortality Files
Slide 5
Access our webpage from the NCHS homepage
Slide 6
- SELECT MORTALITY BY TOPIC AREA AND SEE LIST OF ALL SURVEYS LINKED TO MORTALITY
- OR SELECT DATA BY SPECIFIC NCHS SURVEY
- FROM THE SURVEY SPECIFIC PAGE, ANALYSTS CAN DOWNLOAD THE PUBLIC DATA OR GET INFORMATION REGARDING THE RESTRICTED FILES
Slide 7 through 8
Background: Linked Mortality Public-use Data Release
- Problem/Issue
- Risk of respondent disclosure prevented a public-use release of previously highly utilized mortality linked data files
- BSC recommendation to study ways to release public-use linked mortality files
- Solution
- Conducted disclosure risk analyses to determine extent of NCHS respondent re-identification risk
- Developed perturbation strategy to ensure protection of respondent identity, while maintaining statistical validity
- Cleared by NCHS Disclosure Review Board