Board of Scientific Counselors Program Review

1
Injury Data from NCHS
Lois A. Fingerhut, MA Special Assistant for
Injury Epidemiology
2
Where are we?
3
National Center for Health StatisticsMission

• Principal health statistics agency
• The Secretary, acting through the Center, shall
  conduct and support statistical and
  epidemiological activities for the purpose of
  improving the effectiveness, efficiency, and
  quality of health services in the United States.

4
Key NCHS Activities

• Research and methodology
• Analysis and epidemiology
• Classification and public health data standards
• International activities
• Dissemination and outreach

5
New NCHS Injury web pages
www.cdc.gov/nchs/injury.htm

• NCHS Injury Data and Resources
• The purpose of this Web site is to provide an
  overview of the sources of national level injury
  data available from the National Center for
  Health Statistics (NCHS) and to provide details
  on the injury morbidity and mortality data
  collection systems, surveys and coding schemes
  used to collect and categorize the data. Links to
  other injury data Web sites are provided.

6

• ICE on Injury Statistics
• The International Collaborative Effort (ICE)
  on Injury Statistics is one of several
  international activities sponsored by CDCs NCHS.
  The goal is to provide a forum for international
  exchange and collaboration among injury
  researchers who develop and promote international
  standards in injury data collection and analysis.
  A secondary goal is to produce products of the
  highest quality to facilitate the comparability
  and improved quality of injury data.

7
International Classification of Diseases (ICD)

• Recent Versions for Mortality
• ICD-9 (1979-1998)
• ICD-10 (1999)
• For Morbidity
• ICD 9-CM (Clinical modification)
• ICD-10-CM is not yet in effect (although draft is
  on the NCHS website)
• www.cdc.gov/nchs/about/otheract/icd9/icd10cm.htm

8
International Classification of Diseases (ICD)

• ICD Codes for injury
• Nature of injury codes (diagnoses, i.e.,
  fractures, head injuries, sprains, burns)
• External cause codes (mechanism/ cause and
  intent) of injury (unintentional fall, firearm
  suicide)

9

• Some of the differences between
• ICD-9 and ICD-10

• ICD-10
• E-codes are not for external causes
• V, W, X and Y codes are for external causes
• N codes are not for diagnoses
• S and T codes are for diagnoses
• ICD-10 codes for transportation related injuries
  are very different than they were in ICD-9

10
Differences (cont)

• ICD-10 has more extensive place and activity
  codes (although activity codes are rarely being
  used)
• ICD-10-CM T codes for poisoning morbidity will
  be different
• Injuries included in the WHO official leading
  cause list have changed

11
ICE on Injury Statisticsbegan in May 1994..

• Australia, Austria, Canada, Denmark, England,
  France, Greece, Israel, The Netherlands, New
  Zealand, Norway, South Africa, Spain, Sweden,
  Switzerland, Thailand, Trinidad and the US
• World Health Organization
• Pan American Health Organization

12
Active Projects as of 2003

• Indicators group
• Barell matrix- conversion to ICD-10
• Multiple injury profiles
• Main injury from multiple cause of death
• Household surveys
• Occupational injury
• Strategic Planning
• ICECI and WHO

13
Newer projects..

• Injury severity and administrative datasets
• Poisoning definition

14
AdvICE list
15

• NCHS data sources with injury information

• Mortality
• National Vital Statistics System
• Death Certificates
• Underlying cause
• Multiple Cause
• Morbidity
• National Health Care Surveys
• National Hospital Discharge Survey
• National Hospital Ambulatory Medical Care Survey
• National Health Interview Survey
• National Health and Nutrition Examination Survey

16
NCHS Data

• Strengths
• National coverage (census or sample)
• State and county for vital statistics
• Fully accessible, free of charge
• Well-documented
• ICD coded
• Limitations
• Lack of timeliness
• Lack of detail- ICD Coded
• Hard to get state/county level data for surveys

17
Burden of Injury, 2002

• Deaths 161 thousand
• Hospital Discharges 1.8 million
• ED Visits 33.0 million
• Other Outpatient visits 82.3 million
• All other???

• 1
• 11
• 205
• 511
• ???

18
Source of Mortality Data

• Annual data come from death certificates that are
  completed for each death occurring in the US
• Death certificates have two very different
  purposes
• A legal document regarding the facts and
  circumstances of the death
• A statistical resource providing information on
  the characteristics of the decedent

19

• Death certificates

• Special instructions for deaths from injury and
  poisoning
• Death certificates are generally revised every 10
  years to coincide with changes in the ICD
  revision
• Next revision is due to be implemented on a
  staggered basis beginning with data year 2003 PA
  has not yet set a date
• Moving toward an electronic certificate

20

• Quality of certification

• Precise wording on the certificate determines the
  cause of death.
• Concise description of how injury occurred
  improves ability to code cause of death.
• Lack of specificity translates into unspecific
  codes that hamper understanding and prevention of
  injury.

21
Issues to consider with injury mortality data

• Know what you want to analyze
• Underlying cause of death external cause of
  death codes motor vehicle crashes, firearm
  deaths
• or
• Multiple causes of death injury diagnosis
  codesinternal organ injuries, penetrating
  wounds, poisoning substances or agents

22
More issues to consider

• Timeliness of the data
• What are the most recent data years available?
• Final mortality data are often not available
  until 18 months after the data year ends.
• Preliminary data, available sooner, are accurate
  but are pre-tabulated.

23
Limitation of death certificates

• Little or no information on
• Details on the circumstances,
• Other persons involved,
• Drug and alcohol involvement,
• Weapon type (if applicable)

24
Data dissemination

• Annual published reports
• Deaths Injuries (2001 was 1st year)
• Deaths Final data for (year)
• Deaths Leading causes for (year)
• Health US Chartbooks
• Tabulated data
• Interactive data sources
• WISQARS, WONDER
• Public Use data

25
What can you get from routine published injury
mortality statistics?

• Underlying cause of death
• Age, sex, race and/or ethnicity
• Time trends
• Some geographic detail- State
• Pre-classified external cause of injury
  categories

26
Tabulated data

• Mortality tables by NCHS, Division of Vital
  Statistics available on the web
• Detailed tables include data on age, race, sex,
  cause-of death.
• Includes some state-level data
• Size of tables not constrained www.cdc.gov/nchs/da
  tawh/statab/unpubd/mortabs.htm

27

• What can you get from annual public use data
  files?

• Underlying and multiple cause of death
• 1 cause vs. up to 19 contributing causes (but for
  injury usually only 1 or 2 cc)
• For injury the underlying cause is the external
  cause while the contributing causes are generally
  injury diagnoses
• Detailed demographic geographic variables
• Cause of death either by each cause code or by
  NCHS defined recodes

28
Public use data

• Available on CD-ROM from NCHS at no charge.
• ASCII text file
• Need to have software to interface with the data

29
Frameworks for Presentation

• External Cause of Injury Matrix
• Collaboration between Injury ICE and APHA Injury
  Section-began in late 90s.
• Classifies external causes by cause and intent
• For morbidity and mortality
• ICD-9, 9-CM and ICD-10

30
Recommended External Cause of Injury Matrix

• Exclusion criteria
• place of occurrence
• second-hand tobacco smoke
• misadventures to patients during surgical and
  medical care
• drugs, medicinal, and biological substances
  causing adverse effects in therapeutic use

31
External Cause of Injury Matrix (in part)
Intent of injury Intent of injury Intent of injury Intent of injury Intent of injury
Mechanism Unintentional Suicide Homicide Undetermined Legal/Other
Cut
Firearm External cause codes ICD-9, 9-CM ICD-10 External cause codes ICD-9, 9-CM ICD-10
Poisoning External cause codes ICD-9, 9-CM ICD-10 External cause codes ICD-9, 9-CM ICD-10
Struck by/ against
Suffocation Etc..
32
Injury deaths according to matrix US, 2002
Source National Vital Statistics System
33
Injury death rates, lt20, 2001-02
Source National Vital Statistics System
34
Injury death rates, ages 15-19 yrs, PA and US
1999-2002
35
Injury death rates by intent, lt20, 2002
Source National Vital Statistics System
36
Mortality sex ratios by intent, lt 20, 2002
Homicide line is broken due to unstable numbers
(lt20)
Source National Vital Statistics System
37
Injury death rates by mechanism of injury, lt 20
years 2002
Log scale
Source National Vital Statistics System
Lines are broken when rate would have been based
on fewer than 20 deaths.
38
The Barell Matrix

• Classifies body region by nature of injury
• Developed for morbidity (ICD-9-CM)
• Reinvented for mortality (ICD-10)

39
The Barell Matrix (in part)
40
Barell Matrix ICD-10 Mortality

• Similar structure to morbidity matrix
• Was difficult to revise because of changes made
  in ICD-10 dx codes
• Mortality issues
• Any mention vs total mention of an injury

41
Total and any mentions of injuries on death
certificate by site, lt 20 yrs 2002
Total number of injury deaths17,589
Source National Vital Statistics System
42
Total and any mentions of injuries on death
certificate by nature of injury, lt20 yrs 2002
Source National Vital Statistics System
43
Leading site by nature of injury groups, (total
mentions) lt 20 yrs 2002
Unspecified site
Thorax
Source National Vital Statistics System
44

• But.....
• deaths are only the tip of the injury problem

45
Morbidity data from NCHS
46

• NCHS data sources with injury information

• Morbidity
• National Health Care Surveys
• National Hospital Discharge Survey
• National Hospital Ambulatory Medical Care Surveys
• National Health Interview Survey
• National Health and Nutrition Exam
  Survey-Interview Component

47
National Hospital Discharge Survey (NHDS)

• Provides data on hospitalizations, not persons
• Collected annually 1965-present
• National probability sample of short stay
  non-Federal hospitals

48
NHDS - Variables

• Patient demographic characteristics
• ICD-9-CM diagnoses (up to 7)
• Inpatient procedures (up to 4)
• Days of care
• Month of admission/discharge
• Discharge disposition
• Weight to account for national sample

49
NHDS - Publications

• Series Reports
• National Hospital Discharge Survey 2001 Annual
  Summary with Detailed Diagnosis and Procedure
  Data (Series 13, No. 156)
• Advance Data
• 2002 National Hospital Discharge Survey (No.342)
• Hospitalizations for injury United States, 1996
  (No. 318)
• All of these can be downloaded via web

50
NHDS - Public use data

• Website
• CD-Rom
• Multiple year data file (1979-2002)
• Single year

51
Future publication. . .

• National Trends in Injury Hospitalizations
  1979-2001
• Presents data using new tools available to
  categorize and analyze injury hospitalization
  data
• STIPDA recommendations
• Barell matrix
• External cause matrix

52
Injury hospital discharge rates, children lt 15
yrs 1979-2001
Males
Females
.
Source National Hospital Discharge Survey
53
Fracture discharge rates, lt15 yrs 1979-2001
Source National Hospital Discharge Survey
54
Percent of injury hospitalizations with at least
one valid external cause of injury code, 1979-2001
Source National Hospital Discharge Survey
55
Percent distribution of injury discharges by
external cause of injury codes, 2001
Source National Hospital Discharge Survey
56
National Hospital Ambulatory Medical Care Survey
(NHAMCS)

• Emergency department component
• Outpatient clinic component

57
NHAMCS - ED Data

• Provides data on visits, not persons
• Collected annually 1992-present
• Data abstracted from medical records
• National probability sample survey of visits to
  hospital emergency departments of non-Federal,
  short-stay, general hospitals in the United States

58
NHAMCS (ED)- Variables

• Patient demographic characteristics
• Physicians diagnoses (3)
• Diagnostic services (check box or write in up
  to 21)
• Procedures (check box or write in up to 13)
• Medication (up to 6)
• Disposition
• 1995-2002 variance information available
• Weights

59
Injury specific variables 1995-2002

• Related to injury or poisoning?
• Place of injury occurrence
• Intent of injury
• Work-related
• Cause of injury - verbatim
• Cause of injury (up to 3 ICD-9-CM or External
  cause codes)

60
NHAMCS (ED) Public use data

• Website
• CD-Rom
• Single year 1992-2002
• SAS SPSS input statements

61
ED Visit rates for those lt 18 yrs 2001-02
Source NHAMCS, ED component
62
Injury visit rates, lt 18 yrs 1998-2002
Source NHAMCS, ED component
63
Extremity injury visit rates by nature of injury,
lt18 1998-2002
Source NHAMCS, ED component
64
National Health Interview Survey (NHIS)

• Provides data on nonfatal, intentional and
  unintentional injuries for all family members
• Medical treatment or advice
• Collected annually since 1958 with the most
  recent redesign in 1997
• National probability sample

65
NHIS - Variables

• Demographics for all family members
• Injury episodes (3)
• Verbatim about injury/poisoning event
• External cause of the injury/poisoning episode
  (3)
• Injury/poisoning ICD-9-CM diagnosis (4)

66
NHIS Variables (cont.)

• What person was doing at time of injury/poisoning
• Where person was (home, school, etc.)
• Detailed follow-up questions related to certain
  external cause groups
• Motor vehicle, falls, etc.
• Seat belt, safety seat, helmet use
• Treatment locale (hospitalized, office visit, ed,
  phone call, etc.)
• Missed work and/or school days

67
NHIS Public use data

• Website
• CD-Rom
• Single year from 1969-2002
• Files are in ASCII format and can be read
  directly from CD Rom or downloaded directly from
  web
• SAS and SPSS program input statements available
  on the web

68
Injury episode rates among children and
adolescents NHIS, 2002
Source NHIS
69
Burden of injuryChildren and Injury
70
Injury rates Visits, Discharges and Deaths,
2001-2002
Log scale
ED visits
Hospital discharges
Deaths
Source NCHS- NVSS, NHDS,NHAMCS
71
Tools to Access NCHS Data
72
CDC WISQARS Web-based Injury Statistics Query
and Reporting System

• Leading causes of death reports
• Fatal injury report
• Years of potential life lost (YPLL)
• Injury Maps
• Nonfatal injury data from NEISS-AIP

73

• CDC WONDER
• Compressed mortality file-all causes

• Data from 1979 through 2001
• National, State and County level data
• Each death defined by ICD-9 or ICD-10 code
• Age in 5 - and 10- year age ranges
• Gender
• Race (white, black, other)
• Counts, crude rates, age-adjusted rates (with
  choice of standard populations)
• wonder.cdc.gov/mortSQL.html

74
Future

• Encourage you to look on the web before you begin
  analyses..
• http//www.cdc.gov/nchs/
• Data warehouses are appearing on all different
  topics some include injury data for example,
  womens health, aging, mortality statistics
• Much more attention to multiple cause of death
  data

75
Thanks please feel free to email me

• LFingerhut_at_cdc.gov

www.cdc.gov/nchs/injury.htm