Current Status of the National Violent Death Reporting System October 25, 2004 : Current Status of the National Violent Death Reporting System October 25, 2004 Malinda Steenkamp, M. Phil
Etiology and Surveillance Branch, Division of Violence Prevention
National Center for Injury Prevention and Control
Centers for Disease Control and Prevention
Agenda:
12:00-12:05 pm Introductions, announcements
12:05-12:55 pm Current status of NVDRS
Background – Q&A
About NVDRS – Q&A
Progress to date – Q&A
Results teaser – Q&A
The future – Q&A
12:55-13:00 pm Participant Evaluation Survey
Introduction : Introduction Malinda Steenkamp, M. Phil
Etiology and Surveillance Branch, Division of Violence Prevention
National Center for Injury Prevention and Control
Centers for Disease Control and Prevention
CDC : CDC DVP
James Mercy
Leroy Frazier
Marvin DeBerry
Kanika Shaw
Sanjeev Thomas
Len Paulozzi OSP
Lee Annest
Kevin Webb
Dionne White
Nimesh Patel
Karen Thomas States
MD, MA, NJ, OR, SC,VA
AK, CO, GA, NC, OK, RI, WI
CA, KY, NM, UT
+ others in the future Others
Academic community: NVISS
Organizations: CSTE, STIPDA, NAPHSIS
Federal Agencies (Non-CDC): Department of Justice
Practitioners: NAME
Many more … NIOSH NCHS EPO (MECISP) National Violent Death Reporting System (NVDRS): Partners
What Do We Know About Violence? : What Do We Know About Violence? 50,000 Americans died from violence in 2002
40,000 MVTA, 14,000 HIV
Suicide (32,000) > Assaults (18,000)
Homicide and suicide affect young persons
Homicide 2nd leading cause of death – persons 15-24 years
Suicide 2nd leading cause of death – persons 15-34 years
Elderly have high suicide rates
Violence greater impact on some populations groups
…But We Don’t Know Enough… : …But We Don’t Know Enough… What % of women killed in DV assaults had a restraining order against the offender?
How many drug-related homicides are there in the US each year?
How often do multiple homicides occur?
What % of youth suicide victims are intoxicated at the time they kill themselves?
How many homicides involves the use of an assault weapon?
What life crisis commonly precedes a suicide?
How many men kill themselves after killing their partner?
Public Health Approach: (1) Describe problem What data available for violent deaths? : Public Health Approach: (1) Describe problem What data available for violent deaths? Death Certificate data (NCHS)
Coroner/Medical Examiner data
Law Enforcement Data
Police Reports
Supplementary Homicide Reports (SHR)
Uniform Crime Reporting (UCR)
National Incident Based Reporting System (NIBRS)
Child Fatality Review Teams
Maternal Mortality Review Teams
Data Sources have strengths and limitations : Data Sources have strengths and limitations Data sources often not standardized, not computerized, and not linked
Law enforcement data - what happened before death
SHR cover homicides and have little information
NIBRS have trouble getting law enforcement buy-in
Coroner/Medical examiner files – rich information collected after death
Death certificates describe persons, not incidents, no mention suspects + cannot combine deaths in the same incident
What do we need? : What do we need? A national surveillance system that combines information from multiple sources to provide high quality deaths data that can be applied for the prevention of all types of violence
Features:
More timely
Able to provide information on incidents and victims
Able to characterize perpetrators and relationships
More detailed information
How did NVDRS come about? : How did NVDRS come about? 1999
Institute of Medicine Report recommendation
Early pilot of National Violent Injury Statistics System (NVISS) through private funding by 6 foundations
2000
NVISS gets under way at 12 sites, mostly universities
Harvard + Joyce Foundation convene expert meeting that suggests CDC direct a publicly funded system
CDC starts planning
NVDRS History: 2002 to Present : NVDRS History: 2002 to Present Feb 2002 – 1st Appropriation by Congress ($1.5 million)
Sep 2002 – CDC funds 6 of 17 applicants : MA, MD, NJ, OR, SC, VA – data from 1/1/2003
Feb 2003 – 2nd Appropriation ($1.5 million)
Aug 2003 – CDC funds 7 of 16 applicants : AK, CO, GA, NC, OK, RI, WI – data from 1/1/2004
Feb 2004 – 3rd appropriation ($725,000)
Sep 2004 – CDC funds 4 of 10 applicants: CA, KY, NM, UT – data from 1/1/2005
FY 2005 – Senate included $700,000 in their budget House $0 for NVDRS in their budget
Slide11 : NVDRS States as of September, 2004
FY 02 (6 states) FY 03 (7 states) FY 04 (4 states)
About NVDRS : About NVDRS
National Violent Death Reporting System : National Violent Death Reporting System Mission:
To collect high quality detailed, timely information on all resident and occurrent violent deaths in the US
Design:
Population-based surveillance system
Data from multiple sources
Provides more detailed data on incidents, victims, perpetrators (suspects), weapons, relationships/connections
Data collected by states
Coordinated by CDC
Incremental funding of states
Scope of NVDRS : Scope of NVDRS Intentional Injury Deaths:
Suicide (including terrorism)
Homicide (including terrorism)
Legal intervention (excluding legal executions) Unintentional Injury Deaths
NVDRS Deaths of Undetermined Intent Unintentional
Firearm
Deaths
Operational Definition: ICD-10 Codes that Define NVDRS Cases : Operational Definition: ICD-10 Codes that Define NVDRS Cases
NVDRS - Eligible Deaths in US in 2002 : NVDRS - Eligible Deaths in US in 2002
NVDRS Data Sources : NVDRS Data Sources Primary sources:
Death Certificates (DC)
Coroner/Medical examiner (C/ME) records
Police Records (PR)
Crime lab data (Lab)
Data abstractor input
Secondary, optional sources:
Child Fatality Review Team data (CFR)
Supplementary Homicide Reports (SHR)
Hospital data
Alcohol Tobacco, Firearms and Explosives (ATF) trace information on firearms
Slide18 : Flow of Information for NVDRS Med. Exam/
Coroner State Health
Dept State agencies
Other Fed. agencies
General public
Researchers Law enforcement Crime
Lab Case Info
(- identifiers)
Death certificate NVDRS (CDC) State VDRS CFR Team Analysis Files (+)
Data Re-release policy
Features of the Data Structure : Features of the Data Structure Relational data base (25 data tables)
Incidents are unit of surveillance
Complex incidents can be described, most are not
Many have 1 victim and 1 weapon (e.g. suicide)
Majority of homicides – 1 victim, 1 suspect, 1 weapon
Suspects can be deceased or alive
Connections between persons + weapons captured
Each death can still be unit of analysis
Death rates
Not all data sources available for all persons/weapons
Live suspects have no death certificates
Guns often have no crime lab information
Data Source Concept : Data Source Concept Software allows for data entry by source
DC data entered into the DC source screens
CME data go into CME source screens, etc.
Variables not unique to one source, available from > 1source
e.g. age + sex in DC, CME and PR screens
Data items have same labels and response options irrespective of source (some exceptions)
e.g. place of injury same for CME and PR
Screen shot of NVDRS software : Screen shot of NVDRS software
Sorting Data by Source : Sorting Data by Source Allows for:
independent entry at source - no need to pool sources before entry
Review of each source’s contribution
Feedback to sources of what records held
Eventual direct importation of each source’s data without one overwriting the other (e.g. DC)
Increases validity of comparisons by year + state
Cost – ↑ number of variables – analysis
NVDRS >600 variables
Discrepancies between sources (e.g., age, sex, manner, etc.)
Need to be reconciled
Reconciling Sources : Reconciling Sources Primacy: pick best based on a judgment of validity
(+ assessment in future)
CDC will use the same primacy ranking for all states
States can set own primacy for variables
For Analysis - different primacy in different parts of state or at different times
Missing data defaults to next lower primacy
Combining: most useful for circumstance variables
Data abstractor Input : Data abstractor Input Abstractor determines key features
# of people and weapons in incident
One source may report more than the other
Death Type, i.e., manner of death
Handles the occasional discrepancies
Applies the standard conceptual definition
Weapon Type, e.g., poison or knife
Handles discrepancies among sources
Data Elements : Data Elements Early required elements
Almost all DC variables + some from abstractor
Most within 6 months (except ICD-10 codes)
Demographics, Cause and Manner of death, Multiple cause of death codes, etc.
Late required variables
Generally from C/ME or law enforcement sources
Most within 18 months
Demographics, data on incident, suspects, victim, weapon
Optional variables
Less available information or only for some cases (CFR)
Some more related to treatment
Progress to date : Progress to date
Progress to Date : Progress to Date Implementation Working Group and Stakeholder Forum met twice
17 states funded:
First 6 states completing 2003 data
13 states entering 2004 data
17 States to collect 2005 data
Software developed and used by funded states
Version 2.01 of software
Ongoing technical support through Helpdesk
Users’ manuals developed - on the web at www.cdc.gov/ncipc
Coding Manual Version 2 (not on web yet)
Implementation and coding training done
Reverse site visits conducted, in addition to individual site visits
Draft data release and publication plan developed
Progress to Date (cont.) : Progress to Date (cont.) QA Framework developed
Quantitative QA measures
Ongoing coding support (Helpdesk, Consultation, Coding question database)
Technical and coding assistance conference calls with states each month
Monthly broadcast messages
First Preliminary Data Analysis File on VICTIMS for 2003 being finalized
Ongoing monitoring of 2004 data
Quantitative QA measures
Ongoing coding support
Technical and coding assistance conference calls with states each month
Monthly broadcast messages Conference calls
Broadcast messages
Cheat sheets
First MMWR on preliminary results being finalized
Aim to release before end of 2004
Results Teaser : Results Teaser
Preliminary NVDRS Results (as of October 22, 2004) : Preliminary NVDRS Results (as of October 22, 2004)
Preliminary NVDRS Results (as of October 22, 2004) : Preliminary NVDRS Results (as of October 22, 2004) *Legal Intervention
Preliminary NVDRS Results (as at October 22, 2004) : Preliminary NVDRS Results (as at October 22, 2004)
2005 and beyond … : 2005 and beyond …
Next steps : Next steps 2005
Improving infrastructure
Implementation and coding training for 4 new states
Reverse site visit combined with stakeholder forum
Finalize Data Release and Publication Plan
Preliminary and Final Files
MMWR on 2004 data (annual publication of latest data)
Presentations and publications
Additional Preliminary 2003 Data Analysis Files (Incidents, Suspects, Weapons)
Start with development of Final Data Analysis File for 2003 data year
Develop Preliminary Victims Data Analysis File for 2004 data
First Public Use Dataset made available
Looking further … : Looking further … Restricted Access Data Analysis Files and Processes
More states on board – funding
Ongoing and improved QA
Dissemination of findings – more timely in various ways
Application of findings
Incorporation into Public Health Information Network (NEDSS)
Final remarks : Final remarks Timely and detailed data on violent deaths There are challenges:
Data sources (buy-in, data flow, quality)
Diversity and differences among states and states’ VDRS
Coding and Analysis
Communications
Perceptions
Dissemination of results
Funding
Data are useful and being used by states
Violent Death Reporting is possible and desirable!
Contact Information : Contact Information Malinda Steenkamp, M. Phil
Etiology and Surveillance Branch, Division of Violence Prevention
Msteenkamp@cdc.gov
National Center for Injury Prevention and Control
Centers for Disease Control and Prevention