1 National Center for Health Statistics National Center for Health Statistics Office of Analysis and Epidemiology Office of Analysis and Epidemiology Special Projects Branch Special Projects Branch Record Linkage Program Record Linkage Program Christine S. Cox, SPB Branch Chief, OAE Christine S. Cox, SPB Branch Chief, OAE NCHS Board of Scientific Counselors Meeting NCHS Board of Scientific Counselors Meeting April 24, 2008 April 24, 2008 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Centers for Disease Control and Prevention Centers for Disease Control and Prevention National Center for Health Statistics National Center for Health Statistics
42
Embed
1 National Center for Health Statistics Office of Analysis and Epidemiology Special Projects Branch Record Linkage Program Christine S. Cox, SPB Branch.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
National Center for Health StatisticsNational Center for Health StatisticsOffice of Analysis and EpidemiologyOffice of Analysis and Epidemiology
Special Projects BranchSpecial Projects BranchRecord Linkage ProgramRecord Linkage Program
Christine S. Cox, SPB Branch Chief, OAEChristine S. Cox, SPB Branch Chief, OAENCHS Board of Scientific Counselors MeetingNCHS Board of Scientific Counselors Meeting
April 24, 2008April 24, 2008
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Centers for Disease Control and PreventionCenters for Disease Control and PreventionNational Center for Health StatisticsNational Center for Health Statistics
Data Linkage & Tracking Data Linkage & Tracking TeamsTeams
Data Linkage TeamData Linkage Team Stephanie Bartee (contractor)Stephanie Bartee (contractor) Jim Brittain (contractor)Jim Brittain (contractor) Cordell GoldenCordell Golden Kimberly LochnerKimberly Lochner Donna MillerDonna Miller Gloria WheatcroftGloria Wheatcroft
Data Tracking TeamData Tracking Team Dawn ScottDawn Scott Keith ZevallosKeith Zevallos
4
Why Do LinkageWhy Do Linkage Augments available information for Augments available information for
major diseases, risk factors, and health major diseases, risk factors, and health service utilizationservice utilization Links exposures to outcomesLinks exposures to outcomes Provides longitudinal component to survey Provides longitudinal component to survey
datadata Reduces cost burdenReduces cost burden
Re-contacting survey respondents for follow-Re-contacting survey respondents for follow-up information can be expensiveup information can be expensive
Increases accuracy and detail of data Increases accuracy and detail of data collectedcollected
5
Types of Data LinkageTypes of Data Linkage
Person-level or facility level recordPerson-level or facility level record Person survey data linked with administrative Person survey data linked with administrative
data (e.g. Medicare)data (e.g. Medicare) Hospital survey data linked facility Hospital survey data linked facility
characteristics (e.g. American Hospital characteristics (e.g. American Hospital Associations Annual Survey of Hospitals)Associations Annual Survey of Hospitals)
Contextual dataContextual data Geocoded to standard Census geo-areasGeocoded to standard Census geo-areas
Census data – population & housingCensus data – population & housing EPA data – Environmental air qualityEPA data – Environmental air quality State level data – generosity of Medicaid paymentsState level data – generosity of Medicaid payments
6
How Records are LinkedHow Records are Linked
7
Research Potential of Research Potential of NCHS Linked DataNCHS Linked Data
AgingAging Risk factors for poor health outcomes (hip fractures, Risk factors for poor health outcomes (hip fractures,
stroke, etc.)stroke, etc.) DisabilityDisability
Effects of chronic illness and obesity on disability and Effects of chronic illness and obesity on disability and mortalitymortality
DisparitiesDisparities Mortality patterns by race/ethnicity or socioeconomic Mortality patterns by race/ethnicity or socioeconomic
statusstatus Health ServicesHealth Services
Functional impairment and health care costsFunctional impairment and health care costs GeneticsGenetics
Genetic variants and health outcomesGenetic variants and health outcomes Methodologic StudiesMethodologic Studies
Validation of self-reports vs. administrative recordsValidation of self-reports vs. administrative records
8
NCHS Linkage ProgramNCHS Linkage Program
Early 1980’sEarly 1980’s NMCUES link to Medicare in 1981-1986NMCUES link to Medicare in 1981-1986 NHEFSNHEFS
NDI linkage in 1982-1992NDI linkage in 1982-1992 Medicare linkage in 1980-1986Medicare linkage in 1980-1986
1990’s – NHIS mortality linkage to NDI1990’s – NHIS mortality linkage to NDI 2000 to present – NCHS expanded linkage 2000 to present – NCHS expanded linkage
programprogram Division specific linkagesDivision specific linkages OAE/SPB record linkagesOAE/SPB record linkages
Division of Vital StatisticsDivision of Vital Statistics Linked Birth and Infant Death FilesLinked Birth and Infant Death Files
1983-1991 (birth cohort linkage)1983-1991 (birth cohort linkage) 1995-2004 (period and birth cohort linkages)1995-2004 (period and birth cohort linkages)
Division of Health Care StatisticsDivision of Health Care Statistics 2004 National Nursing Home Survey2004 National Nursing Home Survey
CMS Long Term Care Minimum Data Set on resident CMS Long Term Care Minimum Data Set on resident assessments and facility characteristicsassessments and facility characteristics
Division of Health Interview SurveyDivision of Health Interview Survey Medical Expenditure Panel Survey (MEPS) Medical Expenditure Panel Survey (MEPS)
Linkage FilesLinkage Files NHIS survey cohorts are linked by person NHIS survey cohorts are linked by person
identification numberidentification number
10
OAE Record Linkage OAE Record Linkage ActivitiesActivities
MortalityMortality National Death IndexNational Death Index
Retirement and DisabilityRetirement and Disability Social Security data from the Social Security data from the
Retirement, Survivors, Disability Retirement, Survivors, Disability Insurance (RSDI) and Supplemental Insurance (RSDI) and Supplemental Security Income (SSI) programsSecurity Income (SSI) programs
Medicare enrollment and paymentsMedicare enrollment and payments Enrollment and claims dataEnrollment and claims data
11
Summary Linked Mortality Summary Linked Mortality Data FilesData Files
12
Research Potential of Research Potential of Linked Linked
Mortality DataMortality Data
13
Linked Medicare FilesLinked Medicare Files
Medicare enrollment and claims data for Medicare enrollment and claims data for the years 1991-2000the years 1991-2000 Denominator fileDenominator file MEDPAR Inpatient hospitalizationMEDPAR Inpatient hospitalization MEDPAR Skilled nursing facility (SNF)MEDPAR Skilled nursing facility (SNF) Hospital outpatientHospital outpatient Home Health Agency (HHA)Home Health Agency (HHA) HospiceHospice Carrier (physician/supplier Part B file)Carrier (physician/supplier Part B file) Durable Medical Equipment (DMERC)Durable Medical Equipment (DMERC)
14
Research Potential of Research Potential of LinkedLinked
Medicare DataMedicare Data Examine risk factors for health Examine risk factors for health
conditionsconditions Examine reliability of survey dataExamine reliability of survey data
Compare survey reported Medicare Compare survey reported Medicare enrollments to Medicare claims recordsenrollments to Medicare claims records
Examine survey report of disability with Examine survey report of disability with program participation eligibility criteriaprogram participation eligibility criteria
Examine disparities in Medicare Examine disparities in Medicare service utilizationservice utilization
15
Research Potential of Research Potential of LinkedLinked
Medicare DataMedicare Data Examine risk factors for health Examine risk factors for health
conditionsconditions Examine reliability of survey dataExamine reliability of survey data
Compare survey reported Medicare Compare survey reported Medicare enrollment to Medicare claims recordsenrollment to Medicare claims records
Examine survey report of disability with Examine survey report of disability with program participation eligibility criteriaprogram participation eligibility criteria
Examine disparities in Medicare Examine disparities in Medicare service utilizationservice utilization
16
Publications & Current Publications & Current ProjectsProjects
Using Linked Medicare DataUsing Linked Medicare Data Publications:Publications:
Looker AC. Mussolino ME. Serum 25-Looker AC. Mussolino ME. Serum 25-hydroxyvitamin D and hip fracture risk in older hydroxyvitamin D and hip fracture risk in older U.S. white adults. Journal of Bone & Mineral U.S. white adults. Journal of Bone & Mineral Research. 23(1): 143-150, 2008 Jan.Research. 23(1): 143-150, 2008 Jan.
Current Projects:Current Projects: Assessing the Economic Burden of Chronic Assessing the Economic Burden of Chronic
Kidney Disease in the United StatesKidney Disease in the United States The Association of Obesity and Overweight with The Association of Obesity and Overweight with
Higher Medical Care Costs in Medicare Higher Medical Care Costs in Medicare BeneficiariesBeneficiaries
Comparing Self-Reported Chronic Conditions Comparing Self-Reported Chronic Conditions with Medicare Claims Datawith Medicare Claims Data
17
Linked Social Security Linked Social Security FilesFiles
Retirement, Survivor, & Disability IncomeRetirement, Survivor, & Disability Income Master Beneficiary Record (MBR), 1962-2003Master Beneficiary Record (MBR), 1962-2003
Program eligibility, benefit amount, payment status, Program eligibility, benefit amount, payment status, dual entitlementdual entitlement
Payment History Update System (PHUS), 1984-Payment History Update System (PHUS), 1984-20032003
Benefit payment amounts, including withholding Benefit payment amounts, including withholding information for Medicare Part B premiumsinformation for Medicare Part B premiums
Supplemental Security IncomeSupplemental Security Income Supplement Security Record (SSR), 1974 to 2003Supplement Security Record (SSR), 1974 to 2003
Program eligibility, benefit information, and payment Program eligibility, benefit information, and payment statusstatus
18
Social Security LinkageSocial Security Linkage
19
Research Potential of Research Potential of Linked Social Security DataLinked Social Security Data
Examine reliability of survey information for Examine reliability of survey information for SSA program participation and benefitsSSA program participation and benefits
Compare the health characteristics of early Compare the health characteristics of early retirees (age 62) to those who postpone retirees (age 62) to those who postpone benefitsbenefits
Policy and analysis using validated survey dataPolicy and analysis using validated survey data Predicting the number of people who will become Predicting the number of people who will become
disabled based upon survey reported health disabled based upon survey reported health conditionsconditions
Determining whether current disability entitlement Determining whether current disability entitlement funding levels will be adequate as the population funding levels will be adequate as the population agesages
20
Publications & Current Publications & Current ProjectsProjects
Using Linked Social Security Using Linked Social Security DataData
Publications:Publications: Riley GF. Health Insurance and Access to Care Riley GF. Health Insurance and Access to Care
among Social Security Disability Insurance among Social Security Disability Insurance Beneficiaries during the Medicare Waiting Beneficiaries during the Medicare Waiting Period. Inquiry 43:222-230, 2006 Fall.Period. Inquiry 43:222-230, 2006 Fall.
Current Project:Current Project: The Importance of Objective Health Measures The Importance of Objective Health Measures
in Predicting Exit From the Labor Force Via in Predicting Exit From the Labor Force Via Early OA, DI, and SSI ProgramsEarly OA, DI, and SSI Programs
Where Are They Now? The Subsequent Labor-Where Are They Now? The Subsequent Labor-Force Participation and Health Status of Force Participation and Health Status of Rejected Disability ApplicantsRejected Disability Applicants
Concordance between Self-Reports of SSDI Concordance between Self-Reports of SSDI Application & ReceiptApplication & Receipt
agreementsagreements Balancing limited resourcesBalancing limited resources Improving data accessImproving data access
22
Informed ConsentInformed Consent
Satisfying institutional Satisfying institutional requirements for permission to requirements for permission to linklink
Communicating the importance Communicating the importance of record linkage to survey of record linkage to survey respondents and gaining their respondents and gaining their cooperationcooperation
Tested two options for obtaining consent Tested two options for obtaining consent to record linkageto record linkage Treatment 1: Ask permission to link survey Treatment 1: Ask permission to link survey
data with health-related records of other data with health-related records of other government agenciesgovernment agencies
If no, endIf no, end If yes, ask for last four digits of SSNIf yes, ask for last four digits of SSN
Treatment 2: Ask last four digits of SSN; Treatment 2: Ask last four digits of SSN; consent to link embedded in the questionconsent to link embedded in the question
If partial SSN reported, endIf partial SSN reported, end If partial SSN not reported, as permission to link If partial SSN not reported, as permission to link
Treatment 2 was associated with Treatment 2 was associated with significantly higher odds of consent significantly higher odds of consent overall, and consent with or without overall, and consent with or without an SSN, compared to treatment 1an SSN, compared to treatment 1
Treatment 2 substantially increased Treatment 2 substantially increased the percentage of sample adults the percentage of sample adults consenting to record linkage consenting to record linkage compared to 2005-06compared to 2005-06
Design and test matching Design and test matching algorithms that utilize last four algorithms that utilize last four digits of SSNdigits of SSN
Work with other agencies who Work with other agencies who currently require all 9 digits of currently require all 9 digits of SSN for linkageSSN for linkage
28
Identification DataIdentification Data
Incomplete or inaccurate Incomplete or inaccurate identification data from the identification data from the survey interview can lead to survey interview can lead to potential biases in linked data potential biases in linked data filesfiles NamesNames AddressesAddresses
29
NamesNames
Issues with collection and Issues with collection and cleaning of survey participant cleaning of survey participant namesnames Created standardized procedure to Created standardized procedure to
identify non-names in survey dataidentify non-names in survey data Developed nickname to proper Developed nickname to proper
name conversion tablename conversion table
30
AddressesAddresses
Important to keep address information Important to keep address information currentcurrent Improve linkage accuracyImprove linkage accuracy Particularly important for common surnamesParticularly important for common surnames
Conduct passive trackingConduct passive tracking National change of address matchesNational change of address matches Standardize and update addressesStandardize and update addresses Address updates current as of June 2007Address updates current as of June 2007 Now geo-coding new address dataNow geo-coding new address data
31
Interagency AgreementsInteragency Agreements
Complexity in drafting agreementsComplexity in drafting agreements Agency differences in legislative Agency differences in legislative
mandates and requirements to mandates and requirements to protect data and survey respondent protect data and survey respondent confidentialityconfidentiality
Resolving issues of data ownership Resolving issues of data ownership and access to linked data files, e.g.and access to linked data files, e.g. Can a public-use file be created?Can a public-use file be created? If restricted access required, where will If restricted access required, where will
data reside? Who will control access?data reside? Who will control access?
32
Interagency AgreementsInteragency Agreements NCHS taking leadership role in working NCHS taking leadership role in working
across federal agencies through Federal across federal agencies through Federal Committee on Statistical Methodology Committee on Statistical Methodology (FCSM)(FCSM) 2006: NCHS & Census presentation highlighting 2006: NCHS & Census presentation highlighting
difficulties in developing agreementsdifficulties in developing agreements Included a recommendation that FCSM facilitate the Included a recommendation that FCSM facilitate the
development of a IAA templatedevelopment of a IAA template 2008: FCSM convenes sub-committee on usage 2008: FCSM convenes sub-committee on usage
of administrative recordsof administrative records Defining best practices across federal agenciesDefining best practices across federal agencies Developing an interagency agreement template for Developing an interagency agreement template for
Limited resources to both assist data Limited resources to both assist data users and conduct new linkagesusers and conduct new linkages Developing user documentation, web Developing user documentation, web
pages, analytic guideline & other user pages, analytic guideline & other user toolstools
Transforming administrative data into Transforming administrative data into analytic dataanalytic data
Providing technical assistance to data Providing technical assistance to data usersusers
34
Data User ToolsData User Tools
File layouts & detailed notesFile layouts & detailed notes Sample SAS & STATA input statements Sample SAS & STATA input statements
for public-use linked mortality filesfor public-use linked mortality files Dummy dataDummy data Matching methodology reportsMatching methodology reports Linkage rates for SSA & CMS linked Linkage rates for SSA & CMS linked
datadata Analytic guidelinesAnalytic guidelines
Weighting and variance estimation (NHIS)Weighting and variance estimation (NHIS)
35
Data User Tools (cont.)Data User Tools (cont.)
Summary Medicare and SSA filesSummary Medicare and SSA files Feasibility data files for SSA & CMS Feasibility data files for SSA & CMS
Files – Download from webFiles – Download from web Comparative analysis of the public-Comparative analysis of the public-
use and restricted-use linked use and restricted-use linked mortality datamortality data
Compare the mortality experience of Compare the mortality experience of the NHIS linked mortality cohorts to the NHIS linked mortality cohorts to U.S. populationU.S. population
36
Data AccessData Access
Expand data access for restricted filesExpand data access for restricted files Expand RDC locations (e.g. NCHS Expand RDC locations (e.g. NCHS
restricted data now accessible through 9 restricted data now accessible through 9 Census RDCs)Census RDCs)
Create designated agentsCreate designated agents Create public use file, possible but Create public use file, possible but
difficultdifficult Assess disclosure riskAssess disclosure risk Develop synthetic public-use micro data Develop synthetic public-use micro data
files that are analytically useful and validfiles that are analytically useful and valid
Risk of survey participant re-identification Risk of survey participant re-identification required restricted-use designation for the required restricted-use designation for the most recent mortality updatemost recent mortality update
Great degree of motivation to create public-Great degree of motivation to create public-use linked mortality files to enhance use linked mortality files to enhance utilizationutilization Developed perturbation strategy to ensure Developed perturbation strategy to ensure
protection of respondent identity, while protection of respondent identity, while maintaining statistical validitymaintaining statistical validity
Conducted comparative study to determine Conducted comparative study to determine whether findings based upon the public-use files whether findings based upon the public-use files reproduce those using the restricted-use files.reproduce those using the restricted-use files.
38
Selected Variables on the Selected Variables on the NHIS Linked Mortality FilesNHIS Linked Mortality Files
The public-use linked mortality files The public-use linked mortality files (NHIS, NHANES III, and LSOA II), (NHIS, NHANES III, and LSOA II), with a limited amount of perturbed with a limited amount of perturbed data and reduced number of data and reduced number of mortality variables, yield similar mortality variables, yield similar results as the restricted-use dataresults as the restricted-use data
Findings from the NHIS comparative Findings from the NHIS comparative analyses forthcoming in the analyses forthcoming in the American Journal of EpidemiologyAmerican Journal of Epidemiology
Multi-agency collaborative project exploring Multi-agency collaborative project exploring differences in survey and program estimates of differences in survey and program estimates of Medicaid enrollmentMedicaid enrollment
2001-2002 NHIS linked to MSIS data files2001-2002 NHIS linked to MSIS data files Comparison of NHIS survey estimates of Medicaid Comparison of NHIS survey estimates of Medicaid
enrollment with counts from CPSenrollment with counts from CPS Expansion of partnership planned to include collection Expansion of partnership planned to include collection
of linked Medicaid data for NHIS and NHANESof linked Medicaid data for NHIS and NHANES
State based linkagesState based linkages State Cancer RegistriesState Cancer Registries
National Program of Cancer Registries (NPCR)National Program of Cancer Registries (NPCR) SEER registriesSEER registries