Overview of NCSA's and NVS Research Data Collection Systems Chip Chidester, Director Office of Data Acquisitions 1
Highway Safety Facts (2009) 33,808 Fatalities 30,797 1,517,000 1,929,000 Injury Crashes 3,957,000 4,348,000 Property Damage Only Property Damage Only 6,316,000 5,505,000 Police Reported Crashes Fatal Crashes >16,000,000 Crashes $230 Billion Societal Cost 2
Highway Safety Facts (2009) 33,808 Fatalities 38,309 30,797 309 1,929,000 1,816,000 Injury Crashes 4,348,000 4,304,000 Property Property Damage Damage Only Only 6,316,000 6,159,000 Police Reported Crashes Fatal Crashes FARS >16,000,000 Crashes >16,000,000 Crashes NASS - CDS 6,316,000 Police Reported Crashes 4,348,000 Property Damage Only NASS - GES 1,929,000 Injury Crashes State Data >16,000,000 Crashes $230 Billion Societal Cost Phone Surveys 38,309 309 3
NCSA s Data Collection Network 4
NHTSA Data Systems Detailed Investigations NMVCCS CDS Police Reports State Data GES Not in Traffic Surveillance Data 5
Investigation Based Data Programs CDS NMVCCS 6
National Automotive Sampling System (NASS) Provides nationally representative data on all crashes occurring in the United States Two Components Crashworthiness Data System (CDS) Evaluate Motor Vehicle Safety Countermeasures General Estimates System (GES) Track Motor Vehicle Crash Trends 7
National Automotive Sampling System (NASS) CRASHWORTHINESS DATA SYSTEM (CDS) Data on vehicle damage and the occupant outcome (crashworthiness) of towed light passenger vehicles Field Investigation Based Nationally representative Sample based 4000 to 5000 cases annually Over 600 CDS Data Elements Describe Crash Events Damage to Vehicle Crash Forces Involved Injuries to Victims Injury Mechanisms 8
NASS CDS Sample Design The selection of the NASS sample is accomplished in three stages. The first stage is the selection of geographical areas called primary sampling units (PSUs). The second stage is the selection of police jurisdictions. The third stage is the selection of the police traffic crash report. 9
NASS CDS Criteria Periodic Visits to Police Jurisdictions Crashes Listed & Stratified Towed passenger car, light truck, or van Selection Process Priority given to Late Model Year Vehicle Priority given to Injury Severity and Treatment Fatal, Hospitalized, Serious, Trans, None Sample Size based on the number of Researchers Sampling Weight (based on severity) Selected Proportional to Weight 10
National Automotive Sampling System (NASS) CDS Data Sources Overview of Scene Inspection Physical Plant Roadway Features, etc. Evidence Skid marks, etc. Scaled Diagrams 11
Overview of Vehicle Inspection National Automotive Sampling System (NASS) Exterior Crush Profile - Length, Depth, and Location Collision Deformation Classification (CDC) Principle Direction Of Force (PDOF) Tires Rollover (More Detail 2006) Side Impact (More Detail 2008) 12
National Automotive Sampling System (NASS) CDS Data Sources Overview of Vehicle Inspection Interior Ejection/Entrapment Intrusion Restraint Systems Occupant Contacts 13
National Automotive Sampling System (NASS) CDS Data Sources Overview of Vehicle Inspection Crash Severity Indicators WINSMASH -Delta V Barrier Equivalent Speed Visual Estimates - Speed Changes in Ranges - Minor, Moderate, or Severe Event Data Recorder Delta V over Time Air bag logic 14
CDS Data Sources National Automotive Sampling System (NASS) Overview of Medical Coding Injury Determination Medical Record Review Injury Coding NASS Injury Coding Based on AIS-90 (Update 98) through 2009 AIS 2005 (in 2010) Contact: AAAM, Suite 106, 2340 Des Plaines River Road, Des Plaines, IL 60018 - Added Coding Conventions - Added Aspect - Added Integumentary Injury Identification 15
National Automotive Sampling System (NASS) CDS Data Collection Hardware Laptop computers High resolution digital camera Event data recorder reader 16
National Automotive Sampling System (NASS) The Crashworthiness Data System (CDS) 1988-1996: Paper Form Data Collection Data collected on paper forms 35mm slides Electronic flat file database Hard copy storage 1997 -Electronic Data System (EDS) Pen-based computers to collect data Digital camera for images Oracle database Electronic storage Internet access to data 17
National Automotive Sampling System (NASS) Field Safety Notifications (FSN) Use of NASS infrastructure to identify potential irregularities in vehicles Report potential irregularities through the FSN program Establish link between field personnel and the Office of Defect Investigation (ODI) 18
Special Crash Investigations (SCI) Detailed data on new and rapidly changing technologies First Alert (Problem Identification) Elite Team of Investigators East, Central & West Most detailed investigations into: Unintended Acceleration Backover Avoidance Alternative fuel systems Air Bag systems: n Front, Side, Rollover Canopy School Bus Crashes Potential Vehicle Defects 19
Crash Injury Research and Engineering Network (CIREN) Detailed injury causation data in modern passenger vehicle crashes Multidisciplinary review, research and outreach Three Medical and three Engineering Centers State-of-the-art medical and engineering expertise DICOM repository Bio-Tab coding for all serious and disabling injuries Balances occupant data and vehicle data Validation of kinematics and injury causation Research outputs (recent) Knee-Thigh-Hip injury criteria NASS+CIREN data fusion Small overlap crash severity 315 cases annually / 4,100+ cases total 20
National Motor Vehicle Crash Causation Survey (NMVCCS) On scene data on the factors or events that led up to a crash (primary prevention) of towed light passenger vehicles Nationally representative Sample based On scene field investigation Pre-crash events and associated factors related to how crashes occur n Starts at the critical event Data collection Period n 2005 2007 5,000 cases investigated Preliminary reports on distraction NMVCCS 21
NMVCCS What Happened? Physical Evidence, Witness interviews, etc. were all utilized by the researcher to determine the events and factors that led up to the crash. 22
NMVCCS Human Driver / witnesses Assess condition (better to converse with a conscious person!) Contact all drivers n n n Target interviews Prioritize / order Obtain contact info for all Conduct interviews n Separate from other witnesses / drivers n Follow Interviewing techniques n Document follow-on data 23
NMVCCS Environment Physical evidence Plant layout FRP & POI Digital Images Line of sight Obstructions Relevant signage POI and FRP 24
NMVCCS Vehicle Physical evidence Damage Tires Crash avoidance features Event Data Recorder 25
State Record Based GES 26
Fatality Analysis Reporting System (FARS) The source of information on fatal traffic crashes in the U.S. Source of national statistics on police reported fatalities Annual file Census ~ 33,000 fatalities/year Operated Cooperatively with States n Covers All 50 States, DC, and Puerto Rico n Each crash in FARS has more than 100 different coded data elements that characterize the crash, the vehicles and the people involved. 27
Building the FARS Case Speeding Restraint Use Crash Report Supplement (PJ) Crash Report (PJ) EMS Times (EMS agency) Road Classification (state DOT) FARS Case Toxicology Report (ME) Alcohol Related Driver Records (DMV) Vehicle Registration (DMV) Death Certificate (Vital Records) Race & Ethnicity 28
National Automotive Sampling System (NASS) Nationally representative data Two Components General Estimates System (GES) Tracks Motor Vehicle Crash Trends Crashworthiness Data System (CDS) Evaluate Motor Vehicle Safety Countermeasures 29
National Automotive Sampling System (NASS) GENERAL ESTIMATES SYSTEM (GES) National estimates on all police reported crashes KABCO Nationally representative Sample based Collected at 410 Police Jurisdictions in 60 locations across the US ~55,000 crashes annually Infrastructure NiTS Special Studies 30
Building the GES Case 31
State Data Files Provides computerized state crash data files Crash data from 33 States Voluntary participation Requires conversion for NHTSA use While each States share many common variables and values Each State s variables and values are unique 32
Crash Outcome Data Evaluation System (CODES) Provides States the resources and guidance to perform data linkage on their existing files Links crash and patient data Medical and financial costs Data from 17 states Resource Centers States use this linked data to justify support of safety legislation such as: Mandating Motorcycle Helmet Use Primary Enforcement of Safety Belt Use 33
National Emergency Medical Services Information System Nationwide data on emergency medical services in the U.S. Evaluate patient and EMS System outcomes Develop nationwide training curricula Better injury surveillance data to inform highway safety decisions Address resources for disaster and domestic preparedness 34
Not in Traffic Surveillance (NiTS) Provides data on non-traffic crashes & non-crash incidents Information on deaths and injuries Not previously collected by NHTSA Crashes off the trafficway (road environment) n n Occur in parking lots, on private property, etc. Special emphasis on crashes involving vehicle backing into pedestrian or pedalcyclist Noncrash incidents involving motor vehicles n n n n Hyperthermia / hypothermia Trunk entrapment Power window strangulation Carbon monoxide poisoning A virtual system using multiple existing systems NHTSA databases (GES, FARS, SCI, State Data Files) CPSC s NEISS CDC s NVSS 35
Unintended Acceleration Information in the NCSA Data Systems NASS CDS 2004-2010 CDS file was queried to look for Toyota vehicles potentially involved with a UA incident A total of 145 cases were identified. None had any indication of a UA incident. 36
Unintended Acceleration Information in the NCSA Data Systems Special Crash Investigations 28 active cases When completed, they will be available through the NHTSA website 37
Unintended Acceleration Information in the NCSA Data Systems NMVCCS The entire file was queried to look for a likely case in a UA incident A total of 113 cases were identified. Three UA cases found n 2003 Chevrolet two wheel drive Blazer 18 y/o driver male n 1965 Ford Mustang 2-door coupe 18 y/o driver female n 1994 Chevrolet Corvette coupe 29-y/o male 38
Unintended Acceleration Information in the NCSA Data Systems State Crash Record Based Systems Source records do not specifically identify UA Information captured on the police crash reports tends to be person-oriented Reliant on police-reported information Law enforcement officers may not differentiate person s action vs. vehicle defects Narrative is the primary source 39
Transparency and Data Availability Reports State Traffic Safety Information Data Files All systems FARS Query System NASS Case Viewer SCI Case Viewer Through www.nhtsa.dot.gov NCSA web page 40