Click for next page ( 129


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 128
6 Information Systems OVERVIEW The single most troublesome problem encountered in attempts to conduct epidemiologic studies of illnesses among Persian Gulf War (POW) veterans has been the inability to retrieve information on medical care events such as hospitalizations, outpatient visits, and diagnoses and treatments from Department of Defense (DoD) and Department of Veterans Affairs (DVA) medical records in a uniform and systematic manner. Lack of uniform and retrievable medical information concerning reserve, National Guard, active, and separated forces has greatly inhibited systematic analysis of the health effects of mobilization. DoD and DVA have different and only partially automated inpatient hospital record systems. Neither DoD nor DVA has automated outpatient record keeping, although the committee has recently learned that a database with outpatient records will be available in the near future from DVA. Current systems are fragmented, disorganized, incomplete, and therefore poorly suited to support epidemiologic and health outcomes studies. As an interim measure to obtain information about exposures, health, and medical care among POW veterans, DoD and DVA established four independent programs: the DVA Persian Gulf Health Registry (PGHR), DoD Comprehensive Clinical Evaluation Program (CCEP), DoD Troop Exposure Assessment Model (TEAM), and DoD Registry of Unit Locations (RUL). The committee finds that the PGHR and CCEP are useful for clinical evaluation of the health problems of POW veterans but cannot be utilized for 128

OCR for page 128
INFORAt4 TION SYSTEMS 129 research because they include only self-selected individuals who volunteer to participate in these programs. TEAM and RUL also will have limited utility for epidemiologic studies since they provide information at the unit level rather than at the individual level. Whereas no system of medical record keeping can or should be designed to provide the information needed to address every unanticipated issue regarding the health consequences of either military service in general or a specific military conflict, health information systems can be established to facilitate epidemiologic studies of such service. The committee has identified several changes in health information systems for military personnel that will enhance the capability of the military to evaluate the health consequences of future deployments and service. These include creation of a uniform medical record (UMR), including data from civilian providers; full implementation of the Defense Medical Epidemiological Database (DMED) system; and completion of the Army's Patient Accounting and Reporting Real-Time Tracking System (PARRTS), including expansion to the other branches of service. Medical care and health surveillance (for persons who may need medical attention now) and epidemiologic evaluation of potential threats to the health of service personnel (for research to prevent future problems) will be greatly strengthened by the development of a system that provides access to the entire medical history of each member of the armed services and facilitates linkage to other sources of data. Such a system would provide substantial benefits to the service member and veteran, to future service persons whose health will be better protected, and to DoD or any agency that needs healthy personnel. As far back as World War I, and perhaps antiquity, every war has left a proportion of service personnel and veterans with serious medical complaints that are not explainable on the basis of known health hazards or identified physical illnesses. This pattern is so consistent, and the health problems are so important, that databases and health information systems should be designed and implemented now to deal with and mitigate similar problems that are likely to arise in future conflicts. This chapter addresses the committee's charge to "assess the effectiveness of actions taken by the Secretary of Veterans Affairs and the Secretary of Defense to collect and maintain information that is potentially useful for assessing the health consequences of . . . military service" and to "make recommendations on means of improving the collection and maintenance of such information" (see Appendixes A, B). The chapter focuses on two categories of health and exposure information systems: (1) those established in response to health concerns related to service in the POW and (2) those developed to improve the future capability to evaluate military-service-related health issues. Several systems exist for collecting health and exposure information. Some are relevant to clinical evaluations, others are relevant to research, and some are

OCR for page 128
130 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR relevant to both. Not all of these information systems are appropriate for use in research activities, nor do they have to be. Some of these systems, such as inpatient hospitalization data, were available at the time of the PGW; others, such as the PGHR, were established shortly thereafter; still others, such as PARRTS, have been developed or extended since the PGW. Some of these systems will be useful for collecting data that strengthen future military health preparedness to address research questions. CRITERIA FOR A RESEARCH-ORIENTED HEALTH INFORMATION SYSTEM Information systems should include data items that are selected to meet a clearly articulated purpose. A clear statement of objectives for an information system is essential to selecting appropriate and necessary data items for inclusion; determining the level of detail required; and assessing data completeness, accuracy, and utility. If information systems for DoD and DVA are to support the delivery of high-quality medical care during mobilization and the evaluation of health consequences associated with the conflict, information should be collected for each individual or for well-defined subpopulations and should include at least the following: 1. a list of all individuals participating in activities related to the conflict, to establish the population at risk; 2. complete and accurate information pertaining to the experiences, exposures, and environment of the individuals comprising the population at risk; 3. health experiences prior to mobilization, during mobilization, and after the conflict; and 4. complete and accurate data on personal risk factors for adverse health outcomes and prior health history. Health information systems should be complete, accurate, cost-effective, and readily usable for practical applications. Therefore, each data item should clearly support the stated objectives of the information system, provide useful insight into health concerns within acceptable time frames, and facilitate decision making for prevention and health care programs. Continual evaluation is essential. It is not possible to design and implement an information system that can anticipate every question of current or future concern, but systems can be developed to maximize the opportunity to detect trends, define areas of concern, direct corrective actions, and identify needs for supplementary data collection. Here we discuss systems developed specifically to address PGW issues and

OCR for page 128
INFORAl4 TION SYSTEMS 131 systems being developed to meet future needs for improved military health information systems. Appendix G includes a partial listing of existing databases that could provide information for a Medical Health Surveillance System (MHSS) for the armed forces and others that may be useful sources of supplementary information. The large number of existing databases are acm~n~sterea independently and are rarely linked to each other, which highlights the problems of fragmentation and disorganization. PERSIAN GULF WAR HEALTH INFORMATION SYSTEMS Several information systems have been developed in response to concerns about the health consequences of the PGW. Those designed to obtain health information include DVA's PGHR and DoD's CCEP. Other systems are designed to obtain exposure information, including TEAM and RUL. The PGHR includes active duty, retired, reserve, and National Guard veterans of the PGW who are self-referred to obtain a medical examination and appropriate follow-up and to provide registry information. Recently, the committee has learned that provisions are being made to include spouses and offspring of PG veterans in this database. The PGHR contains identifying and demographic information, history of tobacco use, exposure to selected potentially hazardous substances and experiences, diagnosed diseases and conditions, and self-reported symptoms (DVA, 1995b). The CCEP includes self-referred individuals who are experiencing illnesses that may be related to their service in the PGW and are currently on active duty. Eligible family members are also examined. This program was designed to evaluate and treat the health problems of these individuals. Therefore, data contained in this system include demographic and identifying information, medical history data, self-reported symptoms, physical examination data, laboratory test results, diagnostic data, reported workdays lost, and family member data (DoD, 1996; IOM, 1996~. Other information systems were designed to gather or use exposure data. TEAM was presented to the committee as including National Oceanic and Atmospheric Administration (NOAA) models for the entire period of the oil well fires, troop unit locations and movement data, satellite imagery to determine the daily geographic extent of oil fire plumes, U.S. Army Center for Health Promotion and Preventive Medicine (IJSACHPPM) air pollution data, and other relevant exposure and toxicologic information when and if it becomes available. Recently, the committee learned that data on some individuals who had received the botulinum toxoid or the anthrax vaccine have been included in the database. If TEAM is to be useful for research, it must contain accurate

OCR for page 128
132 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR environmental measurements, plausible computer estimates for the time periods of greatest concern, and the ability to link these data systems to each other. One component of TEAM is RUL, which is an example of a geographical information system (GIS). RUL establishes the latitude and longitude of each military unit at the company level from January 15, 1991, to the date that unit left the Gulf region. TEAM therefore excludes some troops involved in Operation Desert Shield who left the Gulf prior to January 15, 1991. Locations will be recorded for the unit rather than the individual soldier. For example, a time series of geographic locations of troop units might be useful to distinguish between units that were in the vicinity of Kuwait oil well fires and units that were in the area but not downwind from the plume. These data could be a useful resource for general exposure information for units of POW troops, but they are likely to be subject to misclassification of exposures at the individual level, particularly for individuals with assignments out of their units for particular segments of time. Some investigators hope to use the RUL database to assess other exposures, such as proximity to depleted uranium (DU) weapons during tank battles. Such a population might have greater potential exposure to DU than other troop units. Investigating possible disease clusters is another potential use of this database- for example, studying diseases or symptoms among units closest to damaged Iraqi chemical weapons depots. Unit location data will be linked to models of oil well fire location, oil fire plume location, and air pollution data collected by the Army beginning in May 1991. Plans include the development of individual exposure information matrices. Risk values for putative carcinogenic and noncarcinogenic exposures will be determined from sampled as well as modeled grid data. Risk information is to be provided to individual veterans, but this step should be carefully planned and should include explanations of the limitations of interpolating unit data to individuals. Unit-level data are "ecologic" in nature because each person in a given unit will be assigned the average or aggregate exposure for the entire unit. With ecologic data, information about the joint distribution of exposure and health outcome at the individual level is unknown. This can give rise to the "ecologic fallacy," in which the true exposure-response relationship at the individual level is biased by the grouping of data and possibly by uncontrolled confounding at the individual level. This bias could result in either overestimating or underestimating the risk being considered. Other limitations affect one-time ecologic summaries of environmental conditions such as oil fire plumes. Exposures that have varying intensity or are characterized by a "pulsing" or short duration of exposure may be inadequately measured. Late placement of health professionals to document these changing conditions in-theater has probably led to mismeasurement of exposures; this may be especially serious for troops or units with the highest levels of exposure.

OCR for page 128
lNFORM'4 TION SYSTEMS 133 Inadequate data will severely limit the ability of analyses to establish a connection between a service-related exposure and health outcomes or to demonstrate a dose-response relationship. Merely confirming the presence of a potential hazard in the combat environment does not indicate whether the exposure itself caused the adverse effect or whether there was a level of exposure below which no adverse effect occurred. These questions must be assessed accurately to respond to service members' questions and to evaluate the effectiveness of current preventive medicine practice and of protective equipment and programs. HEALTH INFORMATION SYSTEMS FOR THE FUTURE New information systems are being planned and developed to improve military readiness to respond to future health concerns of military service and deployments. These include the UMR, DMED, and Army PARRTS. The committee has identified several changes in systems and practices for collection of information on the health and service-related exposures of military personnel that will increase the ability of the military services to pursue epidemiologic investigations and health outcomes studies. These changes will increase the capacity of the services to evaluate the efficacy of mobilization- supporting health services, including prevention programs; Remobilization, mobilization, and demobilization services; and routine military medical care. The most important of the proposed changes is the creation of a uniform, continuous, and retrievable medical record. The UMR system should include a minimal data set for all service personnel, encompassing personal and demographic descriptions; health- and service-related exposures; illnesses, injuries, and medical conditions that occur during military service; hazardous and potentially hazardous exposures, job assignments, and locations throughout military service; and periods on temporary duty assignment for training and during deployment for military action, particularly to overseas locations. It also should include information about medical contacts that occur after military discharge through the DVA or other government medical providers and, wherever possible, private providers. This information should be collected according to standardized procedures and maintained in a computer-accessible format. The primary value of a uniform and centralized record system used by all military services will be an improvement in the health protection and treatment of individuals. Another important benefit will be the support of epidemiologic investigations. A UMR would facilitate selection of appropriate comparison groups as well as linkage to civilian data sources, such as cancer registries, mortality files, and birth defects registries for health outcomes assessment. A UMR also could facilitate follow- up and follow-back by linkage with such electronic data sources as the Health

OCR for page 128
134 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR Care Financing Administration, the National Death Index, and other data files. These sources are utilized for vital status follow-up in longitudinal studies. Although attractive scientifically, follow-up methods, linkages, and information collected require careful attention to infonned consent and privacy protection. A UMR would solve certain problems in the current medical information systems as well. Most current medical information is in the traditional paper format. Electronic data are limited primarily to hospital inpatient discharge summaries, utilizing international diagnostic codes for predominant medical conditions and the more common medical procedures. Information from mobilization stations and mobile field medical facilities is most likely to be incomplete and to be separated from other military medical records. A single computerized record that follows each individual throughout all facilities would solve many of these problems. Furthermore, such a UMR system would eliminate the need to establish conflict-specific registries for current and future deployments. In the construction of the UMR and related databases and their application to military health issues, care must be taken to safeguard the privacy and confidentiality of military personnel and their family members encompassed by these systems. The committee has identified three aspects of current policies and practices that must be modified to support the completion and implementation of integration of medical record systems and their coordination with civilian medical records. The first is the policy of assigning separate responsibility to each military service for the medical records of active personnel and separate responsibility to DVA for the medical records of veterans in its facilities. Second, medical records established and maintained by the reserves and National Guard are kept separate from each of the above and are not routinely linked. As a result of these practices, the content of the medical record differs among medical services. Third, data recorded by the reserves, National Guard, DVA, and DoD are not linked to data from civilian physicians and facilities. Therefore, one cannot obtain a comprehensive profile of the health of service personnel discharged from active duty. The committee concludes that the branches of the military service, the reserve and National Guard organizations, and the DVA must work together in the development of standardized and uniformly applied practices regarding the collection, recording, and maintenance of service health records. Medical care of the individual, the efficiency and effectiveness of the medical care system, health surveillance, and epidemiologic evaluation of potential threats to the health of service personnel will be greatly strengthened by the development of a system that provides access to the entire medical history of each member of the armed services. The committee is mindful of the need for meticulous attention to many difficulties that will arise in the last step of this proposed data system the

OCR for page 128
INFORMi4 TION SYSTEMS 135 linkage to civilian sources. Those difficulties will include concerns about privacy and confidentiality, costs imposed on the provider of the information (individually small perhaps, but large in the aggregate), barriers to integrating information from civilian sources that is not provided in a UMR format, the need for very rapid response regarding individual medical problems, and the sheer size of the proposed system. We believe that these problems can be solved, but solutions will take time. Delay in integrating data from the civilian sector must not be allowed to retard progress in integrating data from all segments of the military. The committee is encouraged that efforts are being made toward unifying medical records among the service branches, and computerizing parts of this record in a uniform way is being discussed. It is understandable that each service will have aspects of the medical record that are specific to its mission; however, uniform collection of core data items is important. The complexity of medical information systems and the problems encountered in obtaining such data are exacerbated when investigations focus upon reproductive outcomes. Therefore, problems related to health studies in this area highlight issues critical for developing a military medical information system. Discharge summaries, records of infant births, and records of congenital malformations recognized at the time of a newborn's discharge from the hospital are collected in a standardized manner throughout DoD medical treatment facilities worldwide. However, only limited data are available from insurance records such as CHAMPUS (Civilian Health and Medical Program of the Uniformed Services a civilian community-managed care plan for DoD beneficiaries) or from TRICARE (regionalized tri-service health care system). These data may be useful for studying the effects of service on late-gestational pregnancy losses, prematurity, birth weight, and major malformations identifiable within the first days of life. Data are less likely to be available for infertility, pregnancy loss prior to the third trimester, delayed growth and development of offspring, or any condition dealt with by civilian health care providers. The committee has been told that service members on active duty often choose to obtain reproductive care from civilian sources, thus highlighting the need for linkage of this important area of military medical research with civilian medical records. Although a uniform medical record would substantially strengthen the ability of the military services to evaluate the health of service personnel and the efficacy of military health programs and doctrine, it cannot provide all of the information that might be needed to respond to unanticipated health problems arising after a deployment. Several supplementary databases are described below. The committee has heard presentations on the DMED system under development by the three services and the DVA. This system will contain

OCR for page 128
136 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR standardized data elements from each service's epidemiologic database, including demographic data (sex, race, ethnicity, date of birth, marital status, education), personnel data (personal identification number, rank, duty station, unit identification code, unit zip code, DoD occupation group, length of service, dates of active service), and medical events (hospitalization dates, disposition, up to eight diagnoses and eight procedures, cause of injury, sick days per episode, medical treatment facility, autopsy). Future expansion of DMED may include deployment data, medical readiness data (vaccinations, examination status), temporary duty stations, reserve and National Guard data, and outpatient data. Many important research investigations could be undertaken with such information. The fully developed system will, however, have other research capabilities. First, it can be used as an index to persons who have specific features, for example, all those who may have had military occupational exposure to some solvent or all those who have developed some specific form of cancer. Such persons can then be studied in greater depth from the original health records, personal interviews, or other information sources. Second, a complete listing can be used as a "sampling frame" for detailed study of a random sample of persons with specific characteristics, such as a sample of all those who served in the POW. This may be critical when the intensity of analysis precludes study of all persons who have specific features (e.g., the 697,000 who served in the POW). The committee believes that mechanisms should be established to collect more extensive data during periods of deployment and combat. A presentation and demonstration to the committee of the Army's PARRTS indicated that significant efforts have been made to collect in-theater hospitalization data. Inclusion of information from other services and casualties and addition of information on ambulatory care will strengthen the ability of this system to provide real-time data on medical conditions in-theater. PARRTS demonstrates that the Army has initiated real-time electronic submission of data describing health conditions that may compromise the success of a mission and events that may reflect a breakdown in the prevention of illness. These data will have to be linked to individual health information in other databases or in the UMR to be of value beyond the reporting of aggregate combat field hospitalizations. A central group of civilian experts, military specialists, and major operational commanders should review and evaluate the program periodically as one means to advance its mission. The current surveillance of reportable illnesses and the publication of the Medical Surveillance Monthly Report by the Army provide timely data for the entire Army, not just the deployed forces. Similarly, it is important that real- time data collection be monitored to document the numbers of key adverse health effects and the characteristics of the population from which they derive. Medical profiles should be updated periodically to reflect current experiences.

OCR for page 128
INFORAL4 TION SYSTEMS 137 For each military operation, prior or concurrent identification of plausible infectious threats or environmental hazards to health will assist in determining what additional data should be collected for specific in-theater exposures. Additional information needed will depend on geographic area, endemic diseases, nature of the conflict, expected duration and intensity of exposures, and perhaps other factors specific to the conflict. Forward planning for a range of future conflicts will be required, along with ongoing revision during mobilization and deployment, during the conflict itself, and during the postwar period. Establishment by the Army of the Theater Area Medical Laboratory (TAML) for the purposes of identifying and evaluating medical problems and conducting studies during deployments will improve the capability to investigate potential health problems and disease outbreaks while troops are still in the field. However, the success of this concept in the future will depend on the commander's support and commitment to utilize the unit early in deployment. The expertise in epidemiology and the clinical and laboratory diagnostic capabilities offered by TAML could provide immediate and useful guidance and capability to collect information in the field when problems or unusual exposures are identified. In combat situations, military success is, of course, all important. Our recommendations in no way suggest or endorse compromising the military mission for the purpose of improving health data collection. Questions about possible acute or delayed health effects of military service must not interfere with operational activities in any way that could degrade effectiveness in successfully fulfilling the primary military mission. Rather, there should be prior and concurrent review and planning by experts who understand both the military imperatives and the health consequences of service to establish the appropriate mechanisms to collect these data. Two examples from the POW where improvements in data maintenance would have been valuable are the preservation of predeployment immunization records and a more full, informative, and nonthreatening health assessment at the time of demobilization. CONCLUSIONS The purposes of medical records and research records differ, but there is great benefit in collecting as much information as possible in a structured format. This structure will reduce data errors, be compatible with computerized clinical data systems, and be available for research studies. A medical record system for patient care should be constructed with major input from physicians, nurses, and administrators and should be oriented largely toward the care of the individual patient. There is a need for a detailed record of personal, family, and medical history; symptoms at the time the patient is first seen and later; physical findings and how they change; results of each laboratory test and radiological

OCR for page 128
138 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR procedure; detailed records of findings at surgery and from pathology examination; reasons for and results of specialist referrals; daily or even hourly nurses' notes; and condition at discharge, with copies of any instructions provided to the patient regarding activity level, drugs, return visits, etc. Considerable information must be in free-form text, but much can be collected in a structured format. Although patient care data should be collected in as structured a format as possible, the requirements of research studies are somewhat more stringent. A research protocol generally requires additional structure. For example, the investigator may need to collect the same data items for each patient in the same way at the same time point in the medical process. Requirements for precision of observations may also be more rigorous (e.g., special devices and techniques to measure liver function with more precision than is needed for clinical care), and some observations may have to be made with increased frequency. Thus, data items of research interest tend to be more numerous, more precise, and more patterned than those for patient care only. Importantly, most of the text in a clinical record will be of little value to the researcher in that format, although it may have a very important role in helping to understand a complicated medical situation and in the completion and accurate coding of structured items. Computerizing the entire medical record may also reduce the cost of research that utilizes the data, since access to such information will be more efficient. Because of the need for meticulous attention to research needs for specific data in a sometimes unavoidably chaotic medical setting, investigators often find that they must take the lead in screening and securing the data they require, and the data must be available in their laboratories or offices. However, a combined system of records for patient care and research will be increasingly feasible, and the collection of structured data should be maximized. The committee considers these four steps the development of a uniform medical record, (2) the improvement of data collection on exposures and health status of deployed service personnel, (3) the provision of supplementary data on occupational and environmental exposures, and (4) the inclusion of early detection medical teams during major deployments to be important elements of an MHSS that would increase the nation's capacity to address questions about the acute and chronic health consequences of deployments of U.S. military service personnel. In summary, an MHSS would establish the capacity to respond to questions such as the following: 1. What are the baseline personal characteristics and medical status of military service personnel? To what extent are these baseline or preexisting characteristics correlated with the risk of future illnesses and adverse health conditions? Is any correlation likely to be one of cause and effect?

OCR for page 128
INFORMA TION SYSTEMS 139 2. For each activity in, or in support of, combat-related military duty, what assignments and exposures are potentially hazardous to health? How great is the hazard? 3. What is the incidence of illnesses, injuries, and medical conditions occurring during routine active military duty? 4. What incident illnesses, injuries, and medical conditions occur during combat-related military duty? Does their frequency or severity change because of specific deployment activities? 5. What hazardous exposures and assignments experienced during active military duty can be linked to specific health outcomes, particularly military- unique exposures or exposure conditions that are significantly different from civilian settings? 6. What preventive measures can be taken prior to or during known exposures to specific hazardous substances or conditions? Are important positive synergies or adverse interactions anticipated among the multiple prevention approaches employed (multiple vaccines, chemopreventives, uniform repellents, area spraying, etc.~? The development of an MHSS should focus on several issues. Data quality should reflect attention to case and item definitions and ease of input. Data systems should serve a shared purpose among all participating services, agencies (the DoD, DVA, Department of Health and Human Services, Environmental Protection Agency, etc.), and components (active, reserve, and National Guard troops). A lead agency should be identified as the government's proponent and authority for maintaining the MHSS. Recent military deployments have raised questions of service-connected adverse health effects of delayed onset. These often will be identified and treated in DVA or civilian settings after active military demobilization. As a result, it will be extremely important for DoD to ensure that active military health data systems facilitate efforts to address questions that arise months or years after personnel leave active service or that occur among their family members. Such a proposal would require a research capability and supporting health information system that do not exist today in either DoD or DVA. Cooperation of agencies within the government toward this objective is essential to establishment of an effective MHSS. Questions about the frequency of health events also occur independently of deployment. Therefore, national data systems maintained by government agencies such as the Department of Health and Human Services or the Centers for Disease Control and Prevention should be analyzed to obtain baseline or comparison data for referent populations. Participation and oversight of experts external to the government will increase the operational effectiveness of such health information systems.

OCR for page 128
140 HEALTH CONSEQUENCES OF THE PERSIAN GULF WAR Any impediment to health care access by service members decreases the ability of health information system to recognize that health events are occurring and to assess their service-connectedness in a timely manner. Such obstacles to access to health care after the POW resulted in amplification of concern, exaggerated community and political response, and well-intended but occasionally unwise and potentially uninformative government-funded activities, as described throughout this report. Information systems developed immediately after the POW are limited in scope and disconnected from each other. Systems under development to ensure future medical readiness and to enhance epidemiologic capabilities have great potential for producing a seamless medical record that can be linked to other information systems, and thus meet the important military medical objectives of prevention, providing effective and appropriate medical care, and facilitating epidemiologic research.