3
Surveillance and Data

Surveillance may be defined as the systematic and ongoing collection of data. Surveillance serves at least four practical uses. First, surveillance describes the magnitude of a given type of injury relative to other types of injuries in the general population or in special populations. Thus, surveillance data may direct the priorities for injury research to the areas in greatest need of attention. Second, surveillance is used to monitor trends in specific areas of injury. Surveillance data may be analyzed to determine whether specific injury morbidity and mortality measures have increased, decreased, or remained stable over time. Surveillance systems are the primary means by which injury researchers and practitioners identify changes in the magnitude of a specific injury problem (see Fingerhut and Warner, 1997). Third, surveillance is used to identify new injury problems. Surveillance systems provide a means to identify injury risks that have been inadvertently created by the introduction of a new product or by a change in an existing product or process. For example, surveillance and follow-up investigations were instrumental in identifying the occupational hazards associated with skylights. The skylight itself was not dangerous; however, repair and installation presented a hazard because workers fell through the skylights and the openings cut in the roof for the skylight (NIOSH, 1989).

Fourth, surveillance is used as one way to evaluate injury prevention or intervention efforts. These can have differing outcomes: (1) a reduction in the incidence or severity of the target injury without resulting in an unintended adverse outcome; (2) a reduction in the incidence or severity of the target injury, but also resulting in an unintended adverse outcome; (3) little or no effect on the target injury and no unintended outcomes (other than the cost of implementation); or (4) little or no effect on the target injury, but an unintended adverse outcome. Two examples of the utility of surveillance data follow.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment 3 Surveillance and Data Surveillance may be defined as the systematic and ongoing collection of data. Surveillance serves at least four practical uses. First, surveillance describes the magnitude of a given type of injury relative to other types of injuries in the general population or in special populations. Thus, surveillance data may direct the priorities for injury research to the areas in greatest need of attention. Second, surveillance is used to monitor trends in specific areas of injury. Surveillance data may be analyzed to determine whether specific injury morbidity and mortality measures have increased, decreased, or remained stable over time. Surveillance systems are the primary means by which injury researchers and practitioners identify changes in the magnitude of a specific injury problem (see Fingerhut and Warner, 1997). Third, surveillance is used to identify new injury problems. Surveillance systems provide a means to identify injury risks that have been inadvertently created by the introduction of a new product or by a change in an existing product or process. For example, surveillance and follow-up investigations were instrumental in identifying the occupational hazards associated with skylights. The skylight itself was not dangerous; however, repair and installation presented a hazard because workers fell through the skylights and the openings cut in the roof for the skylight (NIOSH, 1989). Fourth, surveillance is used as one way to evaluate injury prevention or intervention efforts. These can have differing outcomes: (1) a reduction in the incidence or severity of the target injury without resulting in an unintended adverse outcome; (2) a reduction in the incidence or severity of the target injury, but also resulting in an unintended adverse outcome; (3) little or no effect on the target injury and no unintended outcomes (other than the cost of implementation); or (4) little or no effect on the target injury, but an unintended adverse outcome. Two examples of the utility of surveillance data follow.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Surveillance data played a critical role in the design and evaluation of a successful intervention to reduce burns and fatalities in residential fires in Oklahoma City. The city's surveillance system was used in designing the intervention, which consisted of providing homes with smoke detectors in neighborhoods with elevated rates of fires and then evaluating the intervention. Continued surveillance confirmed that the smoke detector program reduced the fire injury rate by 74 percent in areas targeted by the intervention, as compared with a small increase in the rest of the city. Additional information collected about each house and its occupants did not suggest any unexpected or undesirable outcomes as the result of the intervention (Mallonee et al., 1996). The mechanisms of airbag-associated deaths were elucidated through the Special Crash Investigation (SCI) program at the National Highway Traffic Safety Administration (NHTSA) (Winston and Reed, 1996). Through a voluntary national reporting network, the SCI program identified unusual crash circumstances in which minor- to moderate-severity collisions resulted in airbag-associated deaths. In-depth crash investigations were conducted in a timely manner across the United States, and the information was used by the automotive safety community to improve the performance of state-of-the-art safety systems. At the same time that data were collected regarding the role of airbags in fatal crashes, surveillance systems in trauma centers for the study of nonfatal injuries identified another unexpected outcome of the airbag intervention (Loo et al., 1996). In severe frontal crashes the occupants would survive but would sustain severe injuries to the lower feet, ankles, and lower legs. Although these types of injuries had always been present, they were of little consequence because the occupants usually died without the added protection of airbags.1 Surveillance data are needed at the national, state, and local (community) levels. National data are critical for drawing attention to the magnitude of an injury problem, for monitoring the impact of federal legislation, and for examining variations in injury rates by region of the country and by rural versus urban or suburban environments. They are also useful in aggregating sufficient numbers of rare cases of a particular type of injury to identify patterns of injury and mechanisms. State and local data better reflect injury problems in specific communities and are therefore more useful in setting program priorities and evaluating the impact of local policies and expenditures. Therefore, local data are usually needed to advocate effectively for the establishment of an injury-related policy or program at the local level. As discussed in this chapter, significant strides over the past decade have greatly enhanced our understanding of the magnitude and impact of injury as a 1 Despite these complications, analyses of surveillance data suggest that airbags reduce the risk of dying in a direct frontal crash by about 30 percent and had saved over 2,600 lives through November 1, 1997 (NHTSA, 1998a).

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment major public health problem (also see Chapter 2). As the availability, accessibility, and quality of the data have improved, they have played an increasingly important role in the development and evaluation of interventions at national, state, and local levels. However, significant impediments to effective injury surveillance remain, notably the high costs of development and maintenance of surveillance systems. Therefore, priority attention should be given to the improvement or expansion of existing data systems and to the development of efficient strategies for linking data across systems to gather additional and more complex information. Additionally, surveillance systems are dependent upon the quality of coded data. This chapter briefly discusses coding issues, and describes national, state, and local sources of injury data and points to areas where improvement is necessary. CODING ISSUES The World Health Organization's (WHO) International Classification of Diseases (ICD) is the most widely used system for coding and classifying the nature and external causes of injury (WHO, 1975). Originally developed in the late 1800s, the ICD is now in its tenth revision. Coding of U.S. mortality data will shift from ICD-9 to ICD-10 starting in 1999. To enhance the use of the ICD for coding nonfatal injury, the National Center for Health Statistics (NCHS) developed a clinical modification (CM) of the ninth revision of the ICD (ICD-9CM) that is the most commonly used classification system for morbidity reporting throughout the United States (U.S. DHHS, 1997). Work is currently under way to develop the clinical modification of ICD-10; its implementation is not expected until the year 2001. Use of the tenth revision of ICD and ICD-10CM will require major adjustments in the way nosologists and researchers approach data collection and analyses. In introducing ICD-10 and its clinical modification, it will be important to ensure that users are adequately trained to take full advantage of its added flexibility and specificity. Coding the Nature of the Injury The ICD-9 diagnostic codes characterize the nature of injury (e.g., open wounds and lacerations, fractures, sprains and strains, burns) as well as the affected region of the body. These codes are fewer in number than the codes in the ICD-9CM and historically have not been updated between revisions of the ICD. Consideration is being given for an updating process for ICD-10. For mortality, the nature of injury is considered to be a contributing rather than the underlying cause of death, which, by definition, is the external cause of the injury. Thus, to know the nature of the injuries contributing to the death, a researcher must have

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment access to the multiple cause-of-death data tapes. At this time, routine data on contributing causes of injury death are not published. For morbidity, the clinical modification of the ICD-10, currently under development, should improve its specificity regarding the extent and severity of the injury. These codes are widely used by health care providers throughout the United States, and are included in most administrative databases and registries. However, there has been little guidance from the injury field as to their optimal use in surveillance and research. The International Collaborative Effort for Injury Statistics (ICE) is currently developing recommendations for presenting injury morbidity data using the ICD nature-of-injury codes (L. Fingerhut, NCHS, personal communication, 1998). Coding the Causes of Injury The ICD-9 Supplementary Classification of External Causes of Injury and Poisoning (E codes) summarizes the circumstances causing the injury, including the intent (intentional and unintentional) and mechanism (e.g., falls, motor vehicle crashes, firearms) of the injury. For mortality data, coding guidelines indicate the use of an external cause-of-injury code as the underlying cause of death when the morbid condition is classifiable to an injury diagnosis. For morbidity data, the 1990s have witnessed major advances in coding because of the increasing availability of external cause-of-injury codes in health systems data. In 1991, the National Committee on Vital and Health Statistics issued a series of recommendations regarding the use of E-codes. These included a recommendation that ''whenever an injury is the principal diagnosis or directly related to the principal diagnosis for a hospitalized patient, there should be an external cause of injury recorded in the medical record" (NCHS, 1991). Since 1992, the Emergency Department Component of the National Hospital Ambulatory Medical Care Survey (NHAMCS) has routinely included E-codes (Burt and Fingerhut, 1998). The National Ambulatory Medical Care Survey (NAMCS) added E-codes in 1995 and the format of the redesigned National Health Interview Survey (NHIS) in 1997 will facilitate the data being E-coded. Currently, 23 states have mandated that E-codes be included in their hospital discharge data systems; 9 states have mandated E-coding of all emergency department encounters (Annest et al., 1998). A new framework for presenting E-coded injury data promises to expedite comparisons of injury profiles across populations and regions (CDC, 1997). Despite its usefulness for the general categorization of injuries by mechanism and intent, significant limitations of the ICD-9 E-code classification have been noted (Fingerhut and Cox, 1998; Smith and Langley, 1998). In response, the tenth revision of the ICD has made some notable improvements, including the addition of requisite codes for place of occurrence and activity in which the

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment person was involved when injured. The addition of those codes should facilitate the identification of occupational-, residential-, and recreational-related injuries. Currently, further efforts are under way by the WHO Working Group on Injury Surveillance Methodology to develop a more detailed classification of external causes that would be compatible with the ICD classification. A draft of the International Classification for External Causes of Injuries (ICECI) was released for comment at the fourth World Conference on Injury Prevention and Control in 1998. Field testing of the system will begin later in the year. The committee endorses this development and suggests broad-based field testing of the ICECI, recognizing, however, that its success will depend largely on its compatibility with the ICD. The effectiveness of external cause coding will be greatly enhanced by two additions to data bases: separate fields and free text. Without separate fields on discharge records, external cause codes may be dropped in preference to coding diagnoses, comorbidities, and complications. Allowance should be made for recording multiple codes, as a single code may be inadequate for describing the cause of injury. Free text, electronically accessible to searching, permits a fuller understanding of the circumstances surrounding an injury and can help identify specific injuries missed by conventional coding (Smith and Langley, 1998). The committee recommends that a high priority be directed at ensuring uniform and reliable coding of both the external cause and the nature of the injury using the ICD on all health systems data, particularly on hospital and emergency department discharge records. Special efforts should be directed at training to ensure optimal use of the tenth revision of the ICD. EXISTING DATA SOURCES There are numerous national, state, and local surveillance systems. They vary in scope and in the extent to which they provide information on mechanism and intent, nature and severity, risk factors, health services use and costs, and health outcomes. A brief review of federal (national), state, and local data systems is presented below. National Data Sources There are 31 federally funded national data systems that collect data on injury mortality, morbidity, and risk factors (Annest et al., 1996). A summary of these data sources is included in Table 3.1. Eight of the systems provide data on work-related injuries. Approximately one-half are ongoing surveillance systems; the remainder are periodic surveys. Many of the systems report information on the location of injury occurrence, the external cause, and the nature of the injury.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Counts of all fatalities in the United States are available from vital statistics, although these data are often limited in the information they provide about the exact nature and circumstances of the injury. The National Vital Statistics System is used to describe the epidemiology of injury mortality. While most NCHS data systems are sample based, the National Vital Statistics System is universal in its coverage. For injury-related deaths, the U.S. Standard Certificate of Death has a number of items including the date and time of injury, whether the injury occurred at work, a description of how the injury occurred, the place of the injury, and the actual street location. Clearly, the death certificate is a potentially rich source of statistical information on injuries and could be made more useful by including directives about the acceptable level of detail when describing an injury-related death and by including additional information. For example, questions about the role of drug and alcohol involvement in a death due to injury and, in the case of motor vehicle crashes, specific questions about whether the decedent was a passenger or the driver and the type of vehicle involved. Directives about what information to provide might reduce the current limitation in the vital statistics, particularly with regard to motor vehicle injuries (Robertson, 1998). Specifically, there is a 38 percent underestimation of fatal injuries associated with motorcycle crashes when death certificates rather than police reports are used as the source of data (Lapidus et al., 1994). More detailed information about the nature and cause of death is generally available from a medical examiner's or coroner's investigation and report (see discussion later in chapter). Uniform data on injuries resulting in hospitalization can be obtained from both the National Hospital Discharge Survey (NHDS) and the Health Care and Utilization Project (Version 3) (HCUP-3). HCUP-3 provides detailed information about the nature of injuries, treatment, and discharge disposition; however, both HCUP-3 and NHDS are limited in that not all states require external cause-of-injury coding on hospital discharge records. In 1994, only about half of the medical records for which an injury was the principal diagnosis had an accompanying E-code. This proportion has increased remarkably to 64 percent in 1996 as the number of states mandating external cause coding has increased. Although procedures exist for estimating distributions by mechanism and intent, given incomplete data, the lack of universal external cause coding of hospital discharges remains a significant impediment to the optimal use of these databases for studying the epidemiology of injury. National data on nonfatal injuries resulting in a visit to an outpatient setting (e.g., emergency department, clinic, physician office) are available from the National Ambulatory Medical Care Survey (NAMCS), and more recently, the National Hospital Ambulatory Medical Care Survey (NHAMCS). Both the NAMCS and the NHAMCS consist of data abstracted from injury-related visits to hospital emergency departments, hospital outpatient departments, and/or physician offices, whereas the National Health Interview Survey (NHIS) relies on self-reports of injury events. National data on injuries that require medical atten-

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment TABLE 3.1 U.S. Federal Data Systems for Injury Surveillance, Research, and Prevention Activities Data System Federal Agency Description Census of Fatal Occupational Injuries (CFOI) BLS Compiles a count of work-related injury fatalities using multiple cross-referenced data sources. Survey of Occupational Injuries and Illnesses (SOII) BLS Provides national and state data on work-related injuries and illnesses as reported by a sample of employers via an annual mail-out survey. National Crime Victimization Survey (NCVS) BJS Measures the number, nature, and characteristics of specific types of crime and victims of crime. A sample of households is screened for victimization and then interviewed. National Ambulatory Medical Care Survey (NAMCS) NCHS Collects data on visits by patients to nonfederal office-based physicians. Samples of surveys completed by a physician or staff are collected regarding patient visits and ambulatory care during a randomly assigned 7- day period per year. National Hospital Ambulatory Medical Care Survey (NHAMCS) NCHS Collects data on visits to U.S. short-stay hospital emergency and outpatient departments from a set of primary sampling units. National Hospital Discharge Survey (NHDS) NCHS Measures inpatient care and hospital utilization by data collected from NCHS or manually collected. National Health Interview Survey (NHIS) NCHS Annual household interviews are conducted to measure the health status of the noninstitutionalized population. Supplemental questionnaires focus on specific health issues (HIV/AIDS, disability, etc.). National Mortality Followback Survey—1993 (NMFS93) NCHS Data from personal interviews or medical records are used to examine socioeconomic differentials in mortality. Aims to evaluate risk factors, prevention, and effect of health care for individuals in their last year of life.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Data System Federal Agency Description National Vital Statistics System—Current Mortality Sample (NVSSS) NCHS Collects mortality data monthly on cause of death, race, age, location, and date from a sample of death certificates filed at state vital statistics offices. National Vital Statistics System—Final Mortality Data (NVSSF) NCHS Compiles demographic and causal data on all death certificates filed in the United States. Behavioral Risk Factor Surveillance System (BRFSS) CDC Monthly household phone surveys on behavioral risk factors are conducted and analyzed for reference when evaluating current and prospective public health programs. Youth Risk Behavior Surveillance System (YRBSS) CDC School-based surveys are conducted every two years to monitor risk behaviors for pupils in grades 9–12 regarding causes of morbidity and mortality. National Traumatic Occupational Fatality Surveillance System (NTOF) NIOSH Census of occupational injury death data for all workers in the United States. National Electronic Injury Surveillance System (NEISS) CPSC Data regarding injuries associated with consumer products or recreational activities are collected from a sample of hospital emergency departments. Law Enforcement Officers Killed and Assaulted (LEOKA) FBI Data from the FBI Uniform Crime Reporting (UCR) System is compiled to monitor law enforcement officer deaths and assaults. National Incident Based Reporting System (NIBRS) FBI Compiles data from all crimes and arrests through the Uniform Crime Reporting System for use by law enforcement, legislators, criminologists, and the public.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Data System Federal Agency Description Nationwide Personal Transportation Survey (NPTS) FHWA Phone surveys are conducted to determine the amount of household travel within a year. Health Care Financing (HCFA) Administration 5% Sample Standard Analysis File (SAF) and Medicare Provider Analysis and Review (MEDPAR) File HCFA Automated Medicare claims are gathered to provide Medicare claims data files for administrative and research purposes. Indian Health Service—Ambulatory Care System (IHSACS) IHS Data from the IHS and tribally operated hospitals are compiled to reflect the- ambulatory care issued to Native Americans by the IHS. Indian Health Service—Inpatient Care System (IHSICS) IHS Data from the IHS and tribally operated hospitals are compiled to reflect the direct inpatient care provided to Native Americans by the IHS. National Child Abuse and Neglect Data System (NCANDS) DHHS Data from automated state child abuse registries or the automated state child welfare social services information systems are compiled, analyzed, and made available as state child abuse and neglect reporting information. National Incidence Study of Child Abuse and Neglect (NIS) DHHS Child protective services (CPS) agencies and non-CPS professionals submit data aimed at estimating the number of abused children and monitor changes. Fatality Analysis Reporting System (FARS) NHTSA Details on motor vehicle crashes involving a fatality are collected within 30 days to provide a general measure of highway safety and problems. National Automotive Sampling System—Crashworthiness Data System (NASSCDS) NHTSA Collects statistical data on motor-vehicle-related crashes, including police reports, medical records, physical evidence, and interviews.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Data System Federal Agency Description National Automotive Sampling System—General Estimates System (NASSGES) NHTSA Samples from police department traffic crash reports are coded and used to estimate national characteristics of police-reported motor vehicle crashes. National Occupant Protection Use Survey (NOPUS) NHTSA A sample of direct observations from observers is used to evaluate passenger vehicle occupant shoulder-belt use, motorcycle helmet use, and child occupant restraint use. Monitoring the Future Study (MTFS) NIDA Surveys in schools are issued to assess the attitudes of youth in grades 8, 10, and 12 regarding drugs, alcohol, tobacco, and their relation to motor vehicle accidents, delinquency, and victimization. Drug Abuse Warning Network (DAWN) SAMHSA Emergency department and medical examiner records regarding drug misuse are compiled in annual reports indicating the scope of substance abuse problems. Census of Agriculture—1992 (BCCOA) Bureau of the Census Questionnaires are sent to all farm operators to gather data on injuries or deaths connected with farm or ranch work in 1992. National Fire Incident Reporting System (NFIRS) NFA Local fire departments voluntarily report fire incidents and civilian and fire service casualties. Health Care Cost and Utilization Project (HCUP) DHHS Compiles data from a sample of more than 900 hospitals regarding inpatient stays; detailing use and cost of medical services so as to better facilitate health services research and policy analysis nationwide. NOTE: BJS = Bureau of Justice Statistics; BLS = Bureau of Labor Statistics; CDC = Centers for Disease Control and Prevention; CPSC = Consumer Product Safety Commission; DHHS = Department of Health and Human Services; FBI = Federal Bureau of Investigation; FHWA = Federal Highway Administration; HCFA = Health Care Financing Administration; IHS = Indian Health Service; NCHS = National Center for Health Statistics; NFA = National Fire Administration; NHTSA = National Highway Traffic Safety Administration; NIDA = National Institute on Drug Abuse; NIOSH = National Institute for Occupational Safety and Health; SAMHSA = Substance Abuse and Mental Health Services Administration. SOURCES: Annest et al. (1996); AHCPR (1998).

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment tion or restricted activity are available from the NHIS. Historically, however, the NHIS has not been a rich source of data on cause of injury because of the lack of detailed information on the mechanism and circumstances of the injury. Beginning in 1997, the core questions of NHIS were redesigned and an entire section on injury was added, including the narrative text of how the injury occurred. Additionally, the definition of injury was changed to include only those injuries that resulted in medical attention (either in person or via telephone). Data have yet to be released from these new questions, although they hold promise for enhancing our understanding of minor injuries. One of the advantages of the NHIS is the ability to link injury data with other kinds of health, social, and demographic information. Routine ICD external cause-of-injury coding on these health systems surveys now permit researchers to make national estimates of health care utilization for less severe injuries according to external cause of injury. An important source of information on product-related injuries is the National Electronic Injury Surveillance System (NEISS) maintained for 25 years by the Consumer Product Safety Commission (CPSC). NEISS obtains statistical information through surveillance of 101 hospital emergency departments and through follow-up studies. Information collected from the NEISS database includes age and gender of the patient, nature of the injury, body part affected, disposition of the patient, product involved, and location of the incident. Also, the address and phone number of the injured person are included, permitting limited follow-up investigations about the nature and cause of the injury. CPSC uses the data for a variety of purposes, including national incidence estimates, priority setting, development and evaluation of product standards, and the identification of products to be banned or recalled. Although there are limitations to NEISS, from CPSC's perspective it provides the data necessary to carry out its mission. Limitations of NEISS, as it is currently configured, include inadequate sampling of hospitals, because the current sampling system is so limited that the data cannot be used for state or even regional estimates; the lack of sufficient detail regarding the nature of the injury (ICD diagnostic codes are not incorporated); and the paucity of detailed data on circumstances surrounding the product's involvement (GAO, 1997). Despite these limitations, NEISS serves as one model for surveillance of injuries that do not necessarily result in death or hospitalization. For these reasons, the National Center for Injury Prevention and Control (NCIPC) recently collaborated with the CPSC in a pilot study to assess the feasibility of expanding NEISS to an all-injury data system. NEISS was expected to capture all causes and types of injuries treated in hospital emergency departments. Preliminary findings from the pilot study are encouraging. NEISS coders recorded about 80 percent of injury-related cases identified by an independent review of a representative sample of emergency department records from 6 of the 21 pilot hospitals. Additionally, about one-half of the 26 million trauma cases treated in emergency departments involved products under the jurisdiction

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment of the CPSC. Other kinds of injury were also included, classified as due to automobiles (13 percent); work related (13 percent); intentional injuries, including firearms (6 percent); food and drugs (1 percent); and other (18 percent) (CPSC, personal communication, 1998). National estimates of the number of injury-related emergency department visits, based on the pilot study, are comparable to those reported from other databases such as the NHAMCS. However, there has been some evidence to suggest that the sensitivity of case identification was somewhat lower for intentional injuries as opposed to unintentional injuries. This discrepancy needs to be addressed. Based on the promising results of this pilot and discussions with staff at both the NCIPC and the CPSC, the committee recommends an expansion of NEISS data collection to include all injuries treated in emergency departments to increase knowledge of the causes and severity of nonfatal injuries. Furthermore, an expanded NEISS could greatly benefit the injury field because it would provide a new and important tool for gathering national estimates and monitoring national trends in injury morbidity, for identifying emerging problems, for evaluating interventions through follow-up studies, and for providing data for policy decisions. If NEISS is expanded to collect all injury data, the committee believes that the system should remain at CPSC since the system is vital to its mission. Additionally, this expansion will be more cost-effective if it remains at CPSC because the agency has already developed the hospital sampling frame, contracted with hospitals, established relationships with hospital personnel, trained hospital coders, and developed procedures for collecting data and for quality assurance. Thus, the work and cost of expanding the current system to all causes of injury would be minimized if CPSC continues to develop the NEISS. The committee recommends that CPSC expand its NEISS system to gather nationally representative data on all injuries treated in emergency departments to increase knowledge of the causes and severity of nonfatal injuries. To ensure the success of an expanded NEISS, the CPSC should convene a steering committee (with representation from CPSC, NCIPC, the Department of Justice, and other relevant federal and state agencies) to set policies and procedures for the expanded NEISS and its uses. Additionally, a comprehensive evaluation of the system's cost-effectiveness should be performed, by an independent body, to determine the utility and future of the effort. The committee stresses the need to ensure that the sampling frame is adequate for intentional injuries. Additionally, training sessions for hospital personnel and onsite NEISS coders will be important, not only in improving the information gathered related to the history of the injury, but also in identifying injuries related to violence since case identification appears to be lower for this type of injury. Finally, an

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment independent evaluation of the expanded system should be conducted annually or biennially, to determine its utility to the field and its ultimate future. In addition to NEISS, there exist several other sources of national data specific to a particular injury mechanism or intent. Examples include the Fatality Analysis Reporting System (FARS) and the National Automotive Sampling System (NASS) for motor vehicle-related injuries; the National Traumatic Occupational Fatality Surveillance System (NTOF), the Census of Fatal Occupational Injuries (CFOI), and the Survey of Occupational Injuries and Illnesses (SOII) for occupational injuries; the National Crime Victimization Survey (NCVS) and the Uniform Crime Reporting (UCR) System for intentional injuries associated with criminal conduct. These data systems are particularly useful for monitoring trends in injury rates specific to certain mechanisms and for identifying risk factors associated with their occurrence. This information, in turn, has been useful in setting national priorities for research and program implementation and in developing and evaluating national policies. For example, data from FARS was able to document the benefits of legislation that raised the minimum purchase age for alcoholic beverages (Chapter 5). The committee considered the merits and feasibility of establishing a comprehensive fatal injury surveillance system that would collect detailed data on all fatal injuries not currently included in existing fatality surveillance systems and that would be coordinated with existing systems. However, the magnitude of that task would be enormous, given the approximately 140,000 injury deaths each year, of which approximately 42,000 are entered into the FARS database (NHTSA, 1998b) and approximately 6,200 are occupational fatalities (BLS, 1996). Therefore, the committee considered ways to accomplish that goal incrementally by identifying a subset of deaths (e.g., homicides and suicides). The committee noted that an ongoing federally sponsored system of surveillance for all intentional injuries (homicides and suicides) is conspicuously absent from the array of data systems available on a national level. While the UCR System, maintained by the FBI, does provide some information on homicides, detailed information about the type of weapon involved is missing. Since the UCR is a voluntary reporting system, it tends to underestimate the actual incidence of homicide (as compared with the record of vital statistics). Moreover, suicides (which outnumber homicides) and unintentional firearm injury deaths are not included in the database. Given the success of FARS, in monitoring motor vehicle fatalities, and the utility of occupational surveillance systems (e.g., CFOI), it seems reasonable to consider a system for recording detailed data on injury deaths that are ordinarily the subject of police investigations, such as suspected homicides and suicides. However, suicide is not a crime in all jurisdictions, and police do not necessarily investigate deaths that are clearly self-inflicted. Therefore, a system depending on police reports for case identification could miss a large proportion of the 31,000 suicides annually.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Injury researchers and practitioners have suggested the development of a firearm surveillance system (Teret et al., 1992). Firearms are the second leading mechanism, after motor vehicles, of injury deaths; better data on firearm deaths would be helpful in regulating firearm commerce and use. However, limiting a new system to firearms seem unnecessarily restrictive and would preclude a number of useful investigations involving other mechanisms on which surveillance data should be collected. It would also appear limiting because the same sources of information that would be tapped for a firearm surveillance system could provide information on homicides and suicides committed by other means, and they may all be amenable to appropriate and specific prevention interventions. Furthermore, a firearm-only system would preclude investigations into weapon substitution in homicides and suicides and provide an unrepresentative sample for investigations of crosscutting issues such as alcohol and other drug abuse. In addition to police reports, medical examiner's and coroner's investigations and reports are another source of data about the nature and cause of death. The medical examiner and coroner systems vary among jurisdictions (whether state, county, district, or city). A medical examiner is usually a licensed physician, whereas a coroner need not be a physician and is often an elected official. A medical examiner system exists in 22 states, a coroner system exits in 11 states, and a ''mixed" medical examiner-coroner system exists in 18 states (Combs et al., 1995). Since the judgment of the medical examiner or coroner ultimately determines whether a death is a homicide, a suicide or an "accident" according to vital statistics, this system could be considered a potential starting point for a fatal intentional injury surveillance system. Medical examiner's and coroner's reports are particularly valuable when they include a full autopsy with blood screens for drugs and alcohol, and information from police reports and forensic scene investigation. However, the completeness of reports varies widely from jurisdiction to jurisdiction, as does the extent to which data are maintained in a centralized data system. Local funding for the medical examiner generally determines the degree of completeness and number of autopsies performed. In 1987, the CDC's National Center for Environmental Health (NCEH) established the Medical Examiner and Coroner Information Sharing Program, to improve the quality of data on death certificates and to increase the availability of those data for scientific research (NCEH, 1998). The committee recommends the development of a fatal intentional injury surveillance system, modeled after FARS, for all homicides and suicides. The committee urges the CDC (specifically NCIPC, NCHS, and NCEH) in collaboration with the National Institute of Justice (NIJ) and NHTSA to conduct a feasibility study for estab-

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment lishing such a system as an extension of the medical examiner and coroner systems. The study should examine the medical examiner and coroner system for ways to standardize, computerize, and centralize data; examine policies and practices of police investigations of both homicides and suicides to maximize the collection of pertinent data; and make realistic estimates of the costs in time and money to establish such a system. The development of a fatal intentional injury surveillance system based on the medical examiner and coroner systems would have to address the variability in the completeness, quality, and reporting of death investigations and concerns about the underreporting of certain types of injury deaths in medical examiner reports (Dijkhuis et al., 1994). It would have to develop a centralized data system that would collect all the pertinent information on homicide and suicide cases from the medical examiner's and coroner's reports, information that often remains only in hard-copy form in the medical examiner's or coroner's office. In an effort to develop as complete a picture as possible of each such event, the proposed system should, at a minimum, include information about (1) the time and place of the incident and of the actual death; (2) characteristics of the injury or wound; (3) characteristics of the victim and, if relevant, the perpetrator(s) and the relationship of the victim to the perpetrator; (4) the motivation and circumstances related to the death; (5) detailed characteristics of the mechanism or weapon; (6) key circumstances related to the death, including the possible role of alcohol and drugs; and (7) details on the location of the injury. Clearly, since much of this information will be supplied to the medical examiner or coroner by the police departments, it is essential that various offices within the Department of Justice and the NIJ participate fully in the development of this system. Unique challenges will be presented in the collection of data for this system, compared to the collection of data about motor vehicle crashes, because some of the desired information, particularly about the perpetrator, motivations, and weapons may not be known in every case. The committee believes that the development of a fatal intentional injury surveillance system is essential for a nationwide effort in reducing fatal intentional injuries. It will identify common mechanisms and situations resulting in such deaths and will enable researchers to develop preventive interventions. Additionally, it will be a positive step in strengthening the medical examiner and coroner systems, which could ultimately lead to the goal of a comprehensive fatal injury surveillance system. State and Local Sources of Injury Data Although national data are useful in monitoring nationwide trends and evaluating national policies, they are often insufficient for identifying injury patterns or evaluating programs and policies specific to local areas. Indeed, all

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment surveillance systems used to evaluate interventions must be relevant to the scope of the intervention in terms of the geography and the population involved. If the intervention is federally mandated (e.g., a driver's side airbag), then a national-level injury surveillance system is appropriate. If, however, an injury problem and its associated risk factors are initially identified by national-level surveillance but the intervention is conducted locally, then state and local data systems are needed to determine the effectiveness of an intervention. Use of local surveillance data, for instance, played a critical role in the design and evaluation of the intervention to reduce burns and fatalities from residential house fires in Oklahoma City (Mallonee et al., 1996). The type and availability of state and local data vary substantially by area and locale. Vital statistics data are available for all injury-related deaths. As discussed above, however, these data are limited in the information they provide about the nature and circumstances of the injury and risk factors associated with the death, and potentially helpful medical examiner and coroner reports vary in their completeness and quality. A potentially powerful source of data for states and local communities in identifying injury risks and priorities for prevention is routine review of selected deaths by multiagency teams of experts, generally referred to as death review teams. Multiagency review of unexplained deaths among children began in Los Angeles County in 1978. In the mid-1980s, in response to the increasing awareness of severe violence against children in the United States, more teams began to emerge across the country. By June 1997, there were multiagency state and/or local teams in every state and the District of Columbia. A challenge facing many death review teams is the management of the data collected during review sessions, which are essential for establishing prevention goals and for future evaluation of the initiative. The Center on Children and the Law of the American Bar Association has provided leadership in developing guidelines and training modules for child death review teams. These guidelines should be reviewed as a possible basis for establishing similar programs for reviewing deaths other than those of children. Surveillance systems could be supplemented with the detailed data gathered from child death review teams and similar death review teams. State and local data on injury hospitalizations are generally available from two principal sources: trauma registries and uniform hospital discharge data. The scope and quality of hospital-based trauma registries have improved significantly over the past several years, and nearly one-half of all states now maintain such registries on a statewide basis (Shapiro et al., 1994; Rutledge, 1995). They provide detailed information about the nature and severity of the injury, its treatment, and the status of the patient on discharge from the hospital. Although basic information about the cause of the injury is always included, most registries do not collect detailed information about the event or the circumstances surrounding the event. Although inclusion criteria vary from hospital to hospital and from

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment state to state, most registries include only "major" trauma patients and generally exclude patients who survive and stay in the hospital less than three days. Further, some trauma registries do not include drowning, poisoning, and/or suffocation. Typically, however, deaths are included regardless of whether they occur in the emergency department or after admission to the hospital. Most statewide registries collect data only from trauma centers. The American College of Surgeons has developed a minimum data set recommended for use by all trauma centers in developing their registries. Efforts are also under way to establish a National Trauma Data Bank (NTDB) that will collate data from trauma centers and trauma systems across the country (G. Strauch, American College of Surgeons, personal communication, 1998). When fully operational, the NTDB, will be a valuable resource to injury researchers. In general, trauma registries can provide useful information, especially for continuous quality improvement initiatives and clinical research (Rutledge, 1995). However, it is important to remember that not all are population based, and some typically exclude data on "nonmajor" trauma and on trauma not treated at trauma centers; thus they are not as useful for evaluating the epidemiology of trauma or trauma systems evaluation (see Chapter 6). Perhaps the most limiting characteristic of most trauma registries is the lack of information about outcomes other than hospital morbidity and mortality. However, if registry data are kept for several years, over time, outcomes could be examined. A recurrent theme of this report is that priority should be directed to the development of better information on the epidemiology, treatment, and outcomes of nonfatal injuries. As more lives are saved due to more effective prevention and regionalization of care, attention is shifting from a singular focus on survival as the criterion of success, to a detailed consideration of nonfatal outcomes as well. Extending the current concept of a trauma registry to include information on longer-term and nonfatal outcomes provides an excellent opportunity to begin developing this information for trauma patients. Considerable effort has been focused, over the past several years, on developing shorter, less time-consuming instruments to gather the necessary data. It will be important to evaluate these newer instruments for their sensitivity and responsiveness to a broad range of trauma patients. Further, effective and efficient systems for tracking patients are needed so that, at specified times following an injury, an assessment of outcomes can be made for patients who received or did not receive acute care. Guidelines for developing and maintaining these registries are critical and should include recommendations regarding specific outcome measures, supplementary data needed to interpret outcomes, methods for routinely tracking and assessing patients, the optimal timing of assessments, and approaches for summarizing and using data on outcomes. Information on long-term outcomes has to be uniformly collected. Attention to issues of data quality, privacy, and confidentiality must be carefully addressed. The committee strongly urges extending the use of established trauma registries to monitor outcomes following injury. A consensus panel should be con-

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment vened to develop the guidelines for collecting information on outcomes. The consensus panel, at a minimum, should include NCIPC and the Health Resources and Services Administration, in collaboration with the American College of Surgeons' Committee on Trauma, the American Academy of Physical Medicine and Rehabilitation, and the State and Territorial Injury Prevention Directors' Association (see Chapter 6). Uniform hospital discharge data, currently maintained by 39 states (Annest et al., 1998), are another source of information about injuries resulting in hospitalization, and are most informative when external cause coding is appended to every discharge record. Hospital discharge data systems are particularly well suited for examining the epidemiology of injury-related hospitalizations and for system evaluation, because they maintain information on all hospitalizations regardless of where the patient is treated. Increasingly, they are being used to evaluate the performance of inclusive trauma systems (MacKenzie et al., 1990a; Mullins et al., 1994) and to examine the epidemiology of injury. However, to use these data to estimate the true incidence of injury resulting in hospitalization, one must avoid double counting those patients transferred from one facility to another and those admitted multiple times for treatment of the same injury. Although the percentage of patients transferred is generally low (e.g., less than 3 percent in Maryland), transferred patients represent an important subgroup (MacKenzie et al., 1990b). In most instances, a readmission should be assigned a principal diagnosis that reflects the complication or need for further treatment, with an additional code to indicate the late effect of a particular injury. Increasingly, databases are being constructed to facilitate the identification of readmissions either through the incorporation of an additional field for readmission data or by facilitating the ability to link information on multiple discharges referring to the same person. The committee strongly suggests that additional fields be added to hospital discharge systems to indicate that a patient transferred from another hospital; that the patient is being readmitted for a previous injury; and for the date of the injury. Although statewide uniform hospital discharge data do not include information on injury deaths that occur in the field or during transport, they have become a valuable source of information for states and local communities. In particular, when linked to vital statistics or medical examiner data, they provide a complete picture of all trauma severe enough to result in death or hospitalization. The epidemiology of less severe injuries requires data from emergency departments, hospital clinics, and physicians' offices. Nine state have recently established E-coded statewide emergency department record systems, an encouraging trend that can be assisted by the Data Elements for Emergency Department Systems, Release 1.0 (NCIPC, 1997). This new system contains uniform specifications for data entered in emergency department patient records, including external cause-of-injury coding, and incorporates national standards for electronic data interchange. Other state and local data that can be useful in studying

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment the epidemiology of less severe injury include routinely collected information from emergency medical services, police and fire departments, and poison control centers. DATA LINKAGES The utility of existing data can be enhanced significantly by linkages across jurisdictions, which overcome the limitations of separate databases and go far in developing comprehensive information about an event, its circumstances, the occurrence and severity of the injury, the type and cost of treatment received, the outcome in terms of both mortality and morbidity, and the administrative or legal outcome. NHTSA has fostered development of linked databases by funding several states to develop Crash Outcome Data Evaluation Systems (CODES) (Johnson and Walker, 1996). CODES were initially designed to develop comprehensive data for determining the impact of safety-belt and motorcycle helmet use on the incidence and severity of injuries, health care costs, and outcome. The implementation of CODES required the linkage of police crash reports with death certificate or medical examiner data and health care data (including emergency medical services data, emergency department data, hospital discharge data, and occasionally data from the insurance claims). CODES databases are now being used to address a variety of motor vehicle injury research and evaluation questions at the local, state, and national levels. Similar linkages will be needed for other types of injuries. Significant barriers exist to successful linkage, however, and are related to (1) limited access to databases (in some cases relevant data are collected but not computerized and, if computerized, are not readily available because of data release policies, concerns about confidentiality, and interagency politics); (2) high costs and limited resources for developing and maintaining databases; and (3) technical difficulties. When the databases to be linked use similar unique identifiers, linkage is relatively easy. However, for confidentiality reasons, most databases do not contain unique identifiers. Probabilistic matching software has been developed and used to link databases when unique identifiers are not available and to deal with inevitable discrepancies related to spelling, data entry errors, or similar problems (Johnson and Walker, 1996). SUMMARY In summary, although significant advances have been made in the development of effective strategies for injury surveillance, there is still much to be accomplished. The development of information to monitor trends in nonfatal injury (nationally and locally), to determine the place of occurrence of injuries, to determine trends and mechanisms for intentional injury deaths, to identify new or

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment changing injury problems, and to evaluate promising new interventions will require a long-term funding commitment to develop new surveillance systems or to support and expand existing systems. Particularly promising efforts include (1) ensuring uniform and reliable coding with the ICD; (2) extending the current NEISS data system to include all injuries treated in the emergency department; (3) the development of a fatal intentional injury surveillance system, based on the successful FARS data system; (4) extending the use of trauma registries to monitor long-term outcome following injury; and (5) the further development and broader application of techniques for data linkage. All of these efforts require substantial coordination and cooperation among multiple agencies and professional groups. It will be important to identify and engage the relevant stakeholders early in the process to ensure successful implementation of each of these initiatives. REFERENCES AHCPR (Agency for Health Care Policy and Research). 1998. Health Cost and Utilization Project. [World Wide Web document]. URL http://www.ahcpr.gov/data/hcuppkt.htm (accessed August 1998). Annest JL, Conn JM, James SP. 1996. Inventory of Federal Data Systems in the United States for Injury Surveillance, Research and Prevention Activities. Atlanta, GA: National Center for Injury Prevention and Control. Annest JL, Conn JM, McLoughlin E, Fingerhut LA, Pickett D, Gallagher S. 1998. How States Are Collecting and Using Cause of Injury Data. Report of the Data Committee of the Injury Control and Emergency Health Services Section of the American Public Health Association. BLS (Bureau of Labor Statistics). 1996. BLS [World Wide Web document]. URL http://www.bls.gov.oshcftab.htm (accessed July 1998). Burt CW, Fingerhut LA. 1998. Injury visits to hospital emergency departments: United States, 1992–95. NCHS Vital Health Statistics 13(131). CDC (Centers for Disease Control and Prevention). 1997. Recommended framework for presenting injury mortality data. MMWR Recommendations and Reports 46(No. RR-14):1–32. Combs DL, Parrish RG, Ing RT. 1995. Death Investigation in the United States and Canada. Atlanta, GA: Centers for Disease Control and Prevention. Dijkhuis H, Zwerling C, Parrish G, Bennett T, Kemper HC. 1994. Medical examiner data in injury surveillance: A comparison with death certificates. American Journal of Epidemiology 139(6):637–642. Fingerhut LA, Cox CS. 1998. Poisoning mortality 1985–1995. Public Health Reports 113(3):219–233. Fingerhut LA, Warner M. 1997. Injury Chartbook. Health, United States, 1996–97. Hyattsville, MD: National Center for Health Statistics. GAO (General Accounting Office). 1997. Consumer Product Safety Commission: Better Data Needed to Help Identify and Analyze Potential Hazards. Washington, DC: GAO. GAO/HEHS-97-147.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment Johnson SW, Walker J. 1996. NHTSA Technical Report: The Crash Outcome Evaluation System (CODES). Washington, DC: NHTSA. DOT HS 808 338. Lapidus GD, Braddock M, Schwartz R, Banco L, Jacobs L. 1994. Accuracy of fatal motorcycle-injury reporting on death certificates. Accident Analysis and Prevention 26(4):535–542. Loo GT, Siegel JH, Dischinger PC, Rixen D, Burgess AR, Addis MD, O'Quinn T, McCammon L, Schmidhauser CB, Marsh P, Hodge PA, Bents F. 1996. Airbag protection versus compartment intrusion effect determines the pattern of injuries in multiple trauma motor vehicle crashes. Journal of Trauma 41 (6):935–951. MacKenzie EJ, Morris JA, Smith GS, Fahey M. 1990a. Acute hospital costs of trauma in the United States: Implications for regionalized systems of care. Journal of Trauma 30:1096–1103. MacKenzie EJ, Steinwachs DM, Ramzy AI. 1990b. Evaluating performance of statewide regionalized systems of trauma care. Journal of Trauma 30(6):681–688. Mallonee S, Istre GR, Rosenberg M, Reddish-Douglas M, Jordan F, Silverstein P, Tunell W. 1996. Surveillance and prevention of residential-fire injuries. New England Journal of Medicine 335(1):27–31. Mullins RJ, Veum-Stone J, Helfand M, Zimmer-Gembeck M, Hedges JR, Southard PA, Trunkey DD. 1994. Outcome of hospitalized injured patients after institution of a trauma system in an urban area. Journal of the American Medical Association 271(24):1919–1924. NCEH (National Center for Environmental Health). 1998. CDC Medical Examiner and Coroner Information Sharing Program. [World Wide Web document]. URL http://www.cdc.gov/nceh/programs/mec/mec.htm (accessed August 1998). NCHS (National Center for Health Statistics). 1991. Report on the Need to Collect External Cause of Injury Codes in Hospital Discharge Data. Hyattsville, MD: NCHS. NCHS Working Paper Series No. 38. NCIPC (National Center for Injury Prevention and Control). 1997. Data Elements for Emergency Department Systems (DEEDS) [World Wide Web document]. URL http://www.cdc.gov/ncipc/pubres/deedspage.htm (accessed December 1997). NHTSA (National Highway Traffic Safety Administration). 1998a. AirBags. [WorldWide Web document]. URL http://www.nhtsa.dot.gov/airbags (accessed August 1998). NHTSA (National Highway Traffic Safety Administration). 1998b. Traffic Safety Facts, 1997. [World Wide Web document]. URL http://www.nhtsa.dot.gov/people/ncsa/pdf/Overview97.pdf (accessed August 1998). NIOSH (National Institute for Occupational Safety and Health). 1989. Preventing Worker Deaths and Injuries from Falls Through Skylights and Roof Openings. Cincinnati, OH: NIOSH. DHHS (NIOSH) Publication No. 90-100. Robertson LS. 1998. Injury Epidemiology. New York: Oxford University Press. Rutledge R. 1995. The goals, development and use of trauma registries and trauma data sources in decision making in injury. Surgical Clinics of North America 75:305–326. Shapiro MJ, Cole KE Jr, Keegan M, Prasad CN, Thompson RJ. 1994. National survey of state trauma registries, 1992. Journal of Trauma 37:835–840. Smith GS, Langley JD. 1998. Drowning surveillance: How well do E codes identify submersion fatalities. Injury Prevention 4:135–139. Teret SP, Wintemuite GJ, Beilenson PL. 1992. The firearm fatality reporting system: A proposal . Journal of the American Medical Association 267:3073–3074.

OCR for page 60
Reducing the Burden of Injury: Advancing Prevention and Treatment U.S. DHHS (Department of Health and Human Services), Public Health Service, Health Care Financing Administration. 1997. International Classification of Diseases, 9th Revision, Clinical Modification, 6th Edition. Washington, DC: U.S. Government Printing Office. DHHS Publication PHS 96-1260. WHO (World Health Organization). 1975. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death, 9th Revision. Vol. 1. Geneva: WHO. Winston FK, Reed R. 1996. Airbags and children: Results of a National Highway Traffic Safety Administration special investigation into actual crashes. In: 40th Stapp Car Crash Conference Proceedings. Albuquerque, NM: Society of Automotive Engineers. SAE Publication 962438. Pp. 383–389.