Management of Health Services: Importance of Epidemiology in the Year 2000 and Beyond

Epidemiologic Reviews Copyright 2000 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 22, No. 1 Printed in U.S.A. Management of Health Services: Importance of Epidemiology in the Year 2000 and Beyond Haroutune K. Armenian 1 and Donald M. Steinwachs 2 In the last half of the 20th century, remarkable growth occurred in the preventive, diagnostic, treatment, and rehabilitative technologies shown to be efficacious in preventing and treating disease and injury. Management of health care services has become increasingly complex as the scientific basis of health care has grown. Complexity also arises from the multiplicity of decision makers involved in health services (1). Decisions about whether to seek preventive or symptomatic care generally are made by individual persons; the health care provider makes choices about diagnostic, treatment, and rehabilitative options; and, ultimately, the patient is expected to accept the provider's recommendations and to adhere to the prescribed regimen. In the broader context of health care systems, managers must make budget and staffing decisions that affect availability and access to services. Public and private purchasers have become a primary force in cost containment because of their decisions on the scope of coverage provided to their constituencies. Similarly, health insurers have to decide what services are covered, how much to pay for the services, and how much patients will pay out-of-pocket. Also, health insurers increasingly are being held accountable for the quality of care provided to their enrolled populations. These decisions affect access to services and the financial burden of health care on the family. In a perfect world, all decision makers would have access to complete information and could use this information to make "best" choices for themselves, their patients, and their communities. A central role for epidemiology in health services management is to provide the information needed by patients, providers, managers, insurers, regulators, and policy makers. A major challenge is to provide the information so it can Received for publication June 18, 1999, and accepted for publication December 9, 1999. 1 Department of Epidemiology, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD. 2 Department of Health Policy and Management, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD. be used by all decision makers, from the highly sophisticated to those with a more limited understanding of health care. This review highlights the role of epidemiology within the framework of decision making regarding management and delivery of health services. In addition, it projects some future options for integrating epidemiology into health services management. EPIDEMIOLOGY FOR HEALTH CARE EVALUATION: LEARNING WHAT WORKS An information science Epidemiology is an information science. Through a variety of approaches, epidemiology generates information for decision making by health professionals working at all levels of the health care system. (2). This information is relevant to persons and communities, to health care managers, and to policy makers. Although these three groups are exposed to different reports about epidemiologic research findings, they are less likely to read epidemiology and other science journals. Similarly, these groups are less likely to be exposed to and to understand the structured approach to making inferences and decisions, which is the foundation of epidemiology. In particular, epidemiology focuses attention on the population and those at risk. These concepts are central to the growing emphasis on population-based medicine that addresses individual and community health care needs. Until recently, epidemiology was used primarily to investigate problems. Epidemiologic investigation focused on identifying an etiologic process for use in both acute and chronic problem-solving situations and on evaluating interventions and programs. This process of problem investigation provides a discipline of operation and thought to epidemiologists that transcends the specifics of the various health problems they deal with. This discipline is the result of effective health care problem solving for over 150 years. Could such a discipline provide useful principles for health care management? 164

Management of Health Services 165 Tools for evaluation Evaluation and policy setting are important functions for any organization that is well managed. These functions are more relevant at an aggregate level of organizational structure. However, much of the interaction regarding delivery of health services occurs at the person or patient level. In addition to assessments of the efficacy of specific interventions that provide the information to make treatment decisions at the single patient level, epidemiology may help in patient management by providing case investigation methods (3, 4). These methods have been developed to explain disease occurrence at the individual patient level. Study methods, like the randomized trials and the case-control and prospective studies that have their origins in the social sciences, have been refined in the context of epidemiologic investigations of etiology. These methods have proven to be very valuable in addressing some of today's important questions about the organization and delivery of health services. Whether it is the impact of managed care, the increasing concern with quality, the changing patterns of institutional care, the effect of information technology, or the issues related to care of the aged and disadvantaged, epidemiologic methods can provide very useful investigative tools to study these problems and identify solutions for them. For example, Selby et al., who analyzed a cohort of the Kaiser Permanente health maintenance organization (HMO) in California, demonstrated that "among members of an HMO, the introduction of a small co-payment for the use of the emergency department was associated with a decline of about 15 percent in the use of that department, mostly among patients with conditions considered likely not to present an emergency (5, p. 635)." In another example of the use of epidemiologic methods, Vardi et al. (6) carried out a controlled intervention aimed at reducing unjustified hospital stays at the Department of Pediatrics of the Tel Aviv University Medical School (Israel). Preset criteria for hospitalization were applied to 155 pediatric inpatients in one ward; these patients were then compared with 248 controls in another ward. The study demonstrated a significant reduction in unjustified stay for the intervention group, with no reduction in subsequent out-of-hospital mortality. These two examples illustrate the growing literature regarding use of epidemiologic methods to address issues related to health services management. Epidemiology for better management? The aim of health services is to have a positive impact on health status outcomes. Information about etiology, efficacy, and effectiveness is what epidemiology provides best. But beyond playing such an information science role, can epidemiology help us to be more efficient in achieving health-services organizational goals? Can epidemiology help us to become better managers of health services resources? In a review on the quality of health care, Blumenthal highlighted clinical epidemiology as "one of the most important new sciences of quality measurement and improvement (7, p. 1147)." Others have presented information on health services functions where the role of epidemiology is paramount (8, 9). These functions have been limited primarily to evaluation and information for policy setting. However, can epidemiologic approaches be used to manage health care organizations more effectively at an operational level? Approaches that epidemiology provides and that are relevant for effective management are based on welltested methods of investigating health problems. Wellestablished information systems for surveillance of outcomes, rigorous processes of data inference, and continuous reassessment of knowledge are valuable characteristics that epidemiology can offer in a health care management environment. EVOLVING INTERFACE OF EPIDEMIOLOGY AND HEALTH SERVICES MANAGEMENT Health services research is the field of inquiry that provides information to providers, managers, patients, policy makers, and researchers on the effectiveness of health care organization and finance, patterns of service use, quality of care, health outcomes, and costs of care (10). It draws on the theories, methods, and applications of many disciplines, including epidemiology. Health services research and epidemiology share some common roots in medicine and public health. Early in the 19th century, P. C. A. Louis introduced the quantitative and comparative method to medical care evaluation and, through his students, strongly influenced the development of public health and epidemiology (11). Health services research provides a framework for describing management problems to which epidemiologic methods might be applied. Two examples are discussed here. Managers make choices regarding allocation of resources and may decide what services are covered by insurance and how they can be accessed. An important source of information for making these choices is the cost and effectiveness of the alternatives being considered. Decisions regarding which among comparable medications are to be included in the formulary or which hospital and surgeon should be used for a specialized procedure should be made by considering both

166 Armenian and Steinwachs cost and outcomes. Today, information on the costs of care is more widely available and is thought to dominate many decisions, in part because of the limited availability of information on comparative effectiveness. Epidemiology provides many of the tools and the perspective for this critical missing piece of information for managers. Also frequently missing is a comprehensive assessment of costs, extending to the health care organization and including the societal costs of lost productivity, out-of-pocket costs to patients, and others. Incorporating cost considerations in epidemiologic studies of comparative effectiveness provides an important dimension that will improve health services decision making. One of the risks contributing to poor outcomes can be delayed detection, diagnosis, and treatment. Evidence that delay contributes to poorer outcomes can argue for screening programs (e.g., mammography) that can be shown to be effective in reducing delay. Delay and the occurrence of poor outcomes may not be distributed randomly among the population at risk. Epidemiology provides the critical information on differences in risk and the effectiveness of screening in improving outcomes. Managers and medical directors may need to decide at what patient ages and time intervals screening should be recommended or covered by insurance and, among those at higher risk, whether to provide more intensive monitoring or outreach programs. Screening can involve a substantial investment of resources to cover the costs of both screening and resolution of false-positive results. Optimally designed screening programs may improve patient outcomes and reduce the need for the highercost medical interventions associated with late diagnosis. The tools of epidemiology provide the means to assess the potential value of screening in improving health outcomes, the effectiveness of specific screening technologies and intervals, and the identification of persons at higher risk who may benefit from more aggressive screening and outreach. The recent development of case-control and other case-based study designs to evaluate screening programs is a good example of how epidemiologic methods can be useful when it is very difficult to use the better-established "gold standard" of the randomized trial. When these study designs are combined with information from the social and behavioral sciences on the willingness of persons to be screened routinely and information from health economists on the societal costs of screening, the manager is provided with the information needed to make informed choices. This information is also highly relevant for persons at risk, purchasers of health care, and policy makers. FUTURE CHALLENGES FOR EPIDEMIOLOGY AND HEALTH SERVICES MANAGEMENT The growing emphasis on health outcomes as the metric for assessing the value of health services and the orientation toward rigorous, evidence-based guidelines for medical practice elevate the need for solid epidemiologic methods of evaluation. At the same time, constraints on resources and rising health care costs have focused attention on the need to demonstrate the value of services paid for and provided. Value is measured as the ratio of outcomes (quality) divided by costs (12). Therefore, value can be increased by improving outcomes, reducing costs, or both. A new cadre of epidemiologists Opportunities for epidemiologic thinking to permeate to a broader level among health professionals and managers are unlikely to be realized unless a new cadre of epidemiologists seeks to integrate their expertise regarding problem solving in health services management. Unfortunately, a groundswell of interest from epidemiologists has not occurred over the past couple of decades. The results of treatment efficacy trials provide general guidance on what treatments to recommend. However, who takes the next step to help the physician and manager decide how to match the treatment to an individual patient (13)? The following are examples that can involve epidemiology going beyond a judgment on efficacy. In some trials, 30 percent or more of participants may respond positively to the placebo; ideally, they should not receive the full treatment but only enough of it to gain the placebo benefit. In most trials, some fraction of participants fails to respond positively to the efficacious treatment; ideally, this group should be identifiable and should not receive this treatment. Since multiple treatments may be available, the comparative efficacy of treatments makes treatment matching to the individual patient more complicated and more critical to achieving good outcomes at reasonable costs. To address these and similar issues, we need to train or redirect epidemiologists for a different level of involvement when making decisions at the patient care level. Assessing the effectiveness of new systems While medicine and public health have contributed substantially to understanding the efficacy of treatment, much of the medical care provided today does not have a rigorous evidence base on which to justify intervention. This shortcoming is even more evident when it comes to recommending system changes on the basis of proven effectiveness. For example, in

Management of Health Services 167 1978, the World Health Organization and the United Nations Children's Fund (UNICEF) introduced the primary health care approach to the world, in Alma- Ata, Kazakhstan, although very few documented studies demonstrated the effectiveness of the approach in affecting outcomes (14). Experimental trials for new systems of health care are rare, expensive, and difficult to mount; thus, in many circumstances, analytical epidemiologic studies such as the case-control method provide a good alternative to randomized trials (15, 16). Epidemiology within the organization In the industrial model of system and organizational performance evaluation, results and outcomes are assessed primarily in comparison to an external standard. Case fatality rates, immunization levels, and readmission rates are judged in comparison to such rates from similar organizations and to past trends. However, such an assessment does not help the manager explain what really underlies these rates in this particular organization and at this time. Epidemiology provides the investigative tools to identify the determinants of outcomes from within the organization. These epidemiologic tools and principles must be organized and systematized so they become more accessible to the largest number of health care providers and managers. If the objective is to enable managers, providers, patients, policy makers, and other stakeholders to use epidemiologic information, the field will need to reach out to these stakeholders and to understand how they use information and how advances in epidemiology can have the maximum opportunity for improving health care delivery and the health outcomes of patients and populations. Although a number of health systems and managed care organizations have established health services research units that study important problems faced by the organization, what is proposed here is a more integrated role for epidemiology within the day-to-day operational decision-making process of health care managers (17). By making epidemiologic tools and decision processes more systematic and accessible to these managers, both effectiveness and efficiency should improve. Decision making by consumers The information revolution we are now witnessing will greatly expand the availability of information for consumers, health care providers, managers, and others. Thus, it is likely to contribute to new relationships between consumers and health care providers. Consumers will become more empowered in the decision process because of their improved access to information. However, the empowered consumer may be in danger of making inappropriate decisions if knowledge is inadequate regarding how to use available data to make personal and health care decisions. Developing the approaches necessary to bring some of the premises of epidemiology-based decision making to the level of the individual consumer is another challenge that epidemiology and health services research may have to face in the year 2000 and beyond. CONCLUSION As an information science, epidemiology cannot be indifferent to the historic changes we are witnessing in information technology. With massive changes occurring in access to information, epidemiology needs to be concerned with not just the quality of the information that it generates but also the needs of a wider base of users of such information. We need to develop effective strategies and approaches for decision making by these users, whether they are health professionals or others dealing with health problems. Over the past two centuries, epidemiologic methodology has been enriched by incorporation of new tools into the discipline whenever epidemiologists decided to tackle a new set of health problems. Addressing current and future issues related to health services management provides such an opportunity to expand epidemiologic methodology beyond its current boundaries. REFERENCES 1. Steinwachs DM. Management information in decisionmaking. In: Armenian HK, Shapiro S, eds. Epidemiology and health services. New York, NY: Oxford University Press, 1998:41-59. 2. Armenian HK. Problem investigation in epidemiology. In: Armenian HK, Shapiro S, eds. Epidemiology and health services. New York, NY: Oxford University Press, 1998:3-13. 3. Armenian HK. Case investigation in epidemiology. Am J Epidemiol 1991 ;134:1067-72. 4. Armenian HK. Case investigation. In: Armenian HK, Shapiro S, eds. Epidemiology and health services. New York, NY: Oxford University Press, 1998:107-33. 5. Selby JV, Fireman BH, Swain BE. Effect of copayment on use of the emergency department in a health maintenance organization. N Engl J Med 1996,334:635^1. 6. Vardi A, Modan B, Blumstein Z, et al. A controlled intervention in reduction of redundant hospital days. Int J Epidemiol 1996;25:604-8. 7. Blumenthal D. Quality of health care. Part 4: the origins of the quality-of-care debate. N Engl J Med 1996;335:1146-9. 8. Shapiro S. Epidemiology and public policy. Am J Epidemiol 1991;134:1057-61. 9. Ibrahim MA. Epidemiology and health policy. Rockville, MD: Aspen Publications, 1985. 10. Steinwachs DM. Health services research: its scope and significance. Am J Pharmacol Ed 1991(Fall);55:274-8. 11. Lilienfeld AM, Lilienfeld DE. A century of case-control stud-

168 Armenian and Steinwachs ies: progress? J Chronic Dis 1979;32:5-13. 12. Tengs TO, Adams ME, Pliskin JS, et al. Five-hundred lifesaving interventions and their cost-effectiveness. Risk Anal 15. 1995; 15:369-90. 13. Steinwachs DM, Wu AW, Cagney K. Health policy issues: 16. outcomes research and quality of life. In: Spilker B, ed. Quality of life and pharmacoeconomics in clinical trials. 2nd ed. Philadelphia, PA: Lippincott-Raven Press, 1995:747-52. 17. 14. Armenian HK. Model systems for primary health care in developing countries. In: Human resources for primary health care in the Middle East. Beirut, Lebanon: American University of Beirut, 1980:194-205. Selby JV. Case-control evaluations of treatment and program efficacy. Epidemiol Rev 1994;16:90-101. Armenian HK. Case-control methods. In: Armenian HK, Shapiro S, eds. Epidemiology and health services. New York, NY: Oxford University Press, 1998:135-55. Gazmararian JA, Lewis N. Health services research in the Prudential Health Care System. Med Care Res Rev 1996;53: S82-S91.