Using Linked Data to Examine Injury and Disability

1
Using Linked Data to Examine Injury and Disability

• Beth Rasch and Chris Cox
• National Center for Health Statistics

2
Background

• Research is limited on the
• injury disability pathway, particularly
  large scale studies
• Determining the temporal relationships between
  injury and disability is challenging

3
Temporal relationships

• Ideally, we would like to know about
• the level of functioning prior to the injury
  event
• the injury event
• the level of functioning following the injury
  event

Pre-injury functioning
Injury
Post-injury functioning over time
4
Data gap

• In the U.S., national data sources that include
  longitudinal information on injury and disability
  are limited
• Data availability may be different in other
  countries
• Record linkage can begin to address this gap

5
Purpose of presentation

• Identify data needs for studying injury /
  disability
• Begin to examine the potential to study injury /
  disability using longitudinal survey data and
  linked administrative data sources

6
What is record linkage?

• Record linkage combines information for the same
  entity or individual from a variety of sources
• Respondent information from health surveys can be
  linked to administrative records (e.g. vital
  statistics, receipt of disability benefits, etc.)

7
Benefits of record linkage

• Increases analytic potential of health surveys
  for epidemiologic research
• Increases accuracy and detail of data collected
• Augments available information
• Provides longitudinal component to survey data
• Reduces cost (eliminates re-contact)

8
Data needs
Pre-injury functioning
Injury
Post-injury functioning over time
9
Example Linked survey data

• NHIS and MEPS are health surveys
• MEPS is a nationally representative subsample of
  NHIS households
• Annual files to link NHIS to MEPS available on
  request from AHRQ
• Longitudinal weight files available on AHRQ
  website
• http//meps.ahrq.gov/Puf/PufSearch.asp?

10
Key analytic variables MEPS and NHIS

• Functioning
• Self-reported function in ADLs, IADLs, movement,
  seeing/hearing, cognition, work, social
  activities
• Use of assistive equipment
• Time lost from usual activities

• Injury
• ICD diagnosis code
• Onset
• Location
• Cause of injury
• Received treatment
• Recovery (MEPS)
• Seriousness of injury (MEPS)

11
MEPS sample design
1995 NHIS
1996
1997
1998
MEPS, Panel 1, 1996-1997
Jan
Jan
Jan
Dec
Round 1 Round 2 Round 3 Round 4 Round
5
MEPS, Panel 2, 1997-1998
1996 NHIS
Round 1 Round 2 Round 3
12
Examining disability after injury

• Identify functioning (adults) in NHIS
• Identify injuries in Round 1 of MEPS
• Determine change in functioning in Rounds 2-5 of
  MEPS
• All indicators are not available for each round

13
Example Linkage with administrative data

• 1994-1998 NHIS linked with SSA Master Beneficiary
  Records from 1962-2003
• Access to restricted files through NCHS Research
  Data Center
• Note NHIS injury data differs before and after
  1997

14
Key analytic variables SSA

• Benefit history
• Applications for disability benefits
• Program eligibility
• Benefit amount
• Payment status
• Disabling condition

15
Temporal relationships
Disability status after injury
History of disability
Injury event
Application for benefits and program
eligibility (1962-2003)
Application for benefits and program
eligibility (1962-2003)
1994-1998 NHIS
Yes
Yes
Yes
No
No
No
16
Temporal relationships
Disability status after injury
History of disability
Benefit status
Injury event
Application for benefits and program
eligibility (1962-2003)
1995 NHIS
1996 MEPS Round 1
1996 MEPS Rounds 2-5
Yes
Yes
Yes
Yes
No
No
No
No
17
Summary

• Linked files provide a longitudinal aspect to
  survey data, offering additional options to study
  the injury disability pathway
• Users need to be aware of the unique analytic
  issues encountered when using linked files