Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Introduction What are the health consequences for the roughly 30 million (one out of every seven) working-age Americans who do not have health insurance?1 If they were to gain health insurance coverage, would their health improve? This report of the Institute of Medicine (IOM) Committee on the Consequences of Unin- surance addresses these two questions with a critical review and assessment of research that has examined the question of how health insurance affects a range of health-related outcomes for adults.2 The Committee is charged with consolidating and critically appraising the evidence regarding the health, social, and economic impacts of uninsurance for individuals, families, communities, and the nation. In this, its second report, the Committee examines the effects on adultsâ health of lacking coverage by conduct- ing a systematic review of clinical and epidemiological literature that address this question. Finally, the Committee applies its findings from this review to reach conclusions about the likely impact that health insurance would have on the health of uninsured adults. A critical question for public policy concerning health insurance coverage and the financing of health care more generally is whether health insurance improves personal health outcomes. It is not, however, the only consideration that is rel- 1See the introductory section of Chapter 4 for a brief description of the social and demographic characteristics of uninsured adults. 2The Committeeâs review primarily focused on adults under age 65 and excluded studies that dealt exclusively with those over 65. 17
18 CARE WITHOUT COVERAGE: TOO LITTLE, TOO LATE evant to devising fair and compassionate public policies for health insurance and health care programs. Health insurance serves the dual purpose of promoting health by making routine health care services financially accessible and protecting individuals and families against large financial losses in the case of major illness or injury. Public policy recognizes both functions of health insurance in the direct governmental financing of Medicare, Medicaid, and the State Childrenâs Health Insurance Program and indirectly through the federal and state tax subsidies of employment-sponsored health insurance (IOM, 2001a).3 This report addresses the first purpose of health insurance only, the effective- ness of insurance coverage in promoting health among those covered. Most of us take for granted that medical services improve our health by relieving pain, reduc- ing functional impairments, curing disease, treating chronic conditions, and mak- ing available preventive measures that help us avoid becoming sick or that detect disease at early and readily treatable stages. Even if medical care were not effective in achieving all of these goals, contact with a physician or another health care provider when we are ill or injured, in pain, or not able to function well in everyday life may comfort and reassure us. Health insurance affords us these less tangible but widely appreciated benefits of our interactions with health care pro- viders and institutions. Finally, Americans value health care highly, as evidenced by the large share of the Gross Domestic Productâ14 percentâthat we devote to it (IOM, 2001a). Most Americans believe that essential care should be and is, in fact, provided to those who need it, regardless of a personâs financial resources or health insurance status (IOM, 2001a). The questions that this report attempts to answer are simple and direct. The evidence that bears on these questions, however, is complex and requires thor- ough analysis and careful interpretation. If health insurance is found to have a positive effect on personal health, it almost certainly must do so because medical careâthe services and access to a health care provider that are paid for and facilitated by health insuranceâpositively affects oneâs health. If research does not confirm that health insurance improves personal health however, it does not necessarily mean that medical care is not effective in improving health. Having insurance is neither a necessary nor a sufficient intermediate step in obtaining medical care. Other factors that enhance or diminish personal health, such as educational attainment, family income and disabilities are correlated with different forms of health insurance and may obscure any effects of health insurance on health outcomes. These factors that affect health status and their interactions with health insurance are discussed in Chapters 2 and 4 of this report. Studies of the impact of health insurance status (that is, whether or not one 3As discussed in the Committeeâs previous report, Coverage Matters, the value of this tax subsidy increases with income and thus disproportionately benefits higher-income households (IOM, 2001a). This may contribute to the disparity in health insurance coverage rates between higher and lower income groups.
INTRODUCTION 19 has coverage and, if so, what kind of coverage) on health care and health must take a wide variety of personal and environmental factors into account in order to isolate the effect of coverage. Before presenting findings, the Committee intro- duces its conceptual framework for identifying these factors and effects related to health insurance coverage, and describes the approach to collecting, evaluating, and interpreting clinical and population survey research. The remainder of this introductory chapter describes the scope of the report, provides basic definitions and classification criteria, and presents the Committeeâs conceptual framework. SCOPE OF THE REPORT This report examines the quantifiable effects, for adults of working age (18 to 65), of having or lacking health insurance on measures of health status (including self-reported health and functioning, mortality, and condition-specific morbidity indicators such as blood pressure for hypertension and blood glucose levels for diabetes) and on measures of the quality and processes of health care. The Com- mittee addresses childrenâs health and health care utilization and perinatal care and outcomes in a forthcoming report that examines them within the dynamic context of the family and of family-level effects of uninsurance. This report also excludes studies limited to people 65 years and older because virtually all elderly have Medicare or comparable insurance coverage.4 The health-related effects for adults who lack any form of health insurance are the primary focus of this report. The scope of health insurance benefit packages and service exclusions (for example, whether and the extent to which a health plan covers prescription drugs, mental health services, or screening services) is a sec- ondary focus that emerged in the course of the Committeeâs review. Studies that compared uninsured adults with those who had various forms of coverage with different benefits, health care utilization, and outcomes had findings that appeared to depend in part on the nature of the benefit package. Furthermore, in a few instances, studies explicitly compared health plansâ benefit structures (for example, in the cases of preventive services and mental health care, as discussed in Chapter 3). Thus, the Committeeâs findings and conclusions reflect the evidence available about the effects of particular benefits and health plan characteristics and compare these for insured and uninsured persons. Last, this report considers only health-related outcomes associated with health insurance status. It does not address the financial impacts on individuals and families of having or lacking coverage, impacts on health care providers and institutions, or societal economic costs. These impacts will each be considered in subsequent Committee reports. 4While Medicare covers hospitalizations, physician and other outpatient services, including reha- bilitative therapies and home health care, it does not cover most outpatient prescription drugs nor does it cover nonrehabilitative long-term care.
20 CARE WITHOUT COVERAGE: TOO LITTLE, TOO LATE DEFINITIONS The Committeeâs definitions of health insurance and uninsured are consistent with those adopted in its previous report, Coverage Matters: Insurance and Health Care, reflecting both a policy focus on people who lack any form of financial coverage for basic hospital and physician care and the practical limitations of how information about health insurance status is presented in health services research. âHealth insuranceâ is defined by the Committee as financial coverage for basic hospital and ambulatory care services provided through employment-based in- demnity, service-benefit, or managed care plans; individually purchased health insurance policies; public programs such as Medicare, Medicaid and the State Childrenâs Health Insurance Program, and other state-sponsored health plans for specified populations.5 âUninsuredâ refers to persons without any form of public or private coverage for hospital and outpatient care for any length of time. The Committee did not set out to examine specific insurance benefits or levels of cost sharing, both of which are relevant to questions about the adequacy of health insurance coverage, or âunderinsurance.â6 As discussed immediately above, the research results dictated that the Committee consider evidence and findings re- lated to the scope of benefits covered by various health insurance plans and programs. Durations of time are important in measuring both health insurance coverage and its effects. Virtually all studies in this field have limitations with respect to measuring periods of coverage and changes in insurance status over time. Studies usually report health insurance status at the time of initial contact with the health care system and enrollment in the study. Any exception to this practice is noted in the discussion of a particular study. Surveys that measure health insurance cover- age rates for populations may use any of the following definitions of uninsured: â¢ uninsured at the time of the interview, â¢ uninsured that month, or â¢ uninsured over a past period of a given length (e.g., uninsured for the entire previous year).7 5CHAMPUS, now TriCare, the health insurance program for dependents of active and retired military personnel is classified in some studies as employment-based coverage, and grouped with private plans, and in others (less commonly) as public insurance, and grouped with Medicaid and Medicare. The Committee does not classify Department of Veterans Affairs (VA) health care as health insurance because it provides direct health care to veterans entitled on the basis of disability and serves as a provider of last resort for other veterans. Studies evaluating health insurance status effects most commonly classify VA health care as a separate category or exclude it in recognition of its unique features. Some studies include persons served by VA facilities as publicly insured. 6See the Committeeâs first report, Coverage Matters (pp. 27â28) for a brief discussion of underinsur- ance. 7See Appendix B of Coverage Matters (IOM, 2001a), for a description of the largest and most frequently cited surveys of health insurance coverage.
INTRODUCTION 21 The majority of published research literature consists of cross-sectional analyses (i.e., studies comparing different groups at a single point in time) that have limited usefulness in understanding the effects of health insurance status longitudinally or over time (e.g., whether health outcomes differ for persons who lack coverage for 6 months compared with those who lack coverage for 18 months), or the effects of transitions or interruptions in coverage over time. This report uses the terms needed or necessary health care, appropriate health care, and, less frequently, guideline-concordant care to refer to health care and medical services that, with reference to some explicit professional consensus- or evidence- based utilization or practice guidelines, are considered effective and necessary, not only to treat emergency and acute conditions but also to manage chronic disease and sustain or enhance health in those who are well. As discussed in the following section, the Committee has focused its evaluation of the evidence about health insurance on those studies that have used explicit criteria for the appropriate use of health care services. A CONCEPTUAL FRAMEWORK FOR UNDERSTANDING THE EFFECTS OF HEALTH INSURANCE STATUS Measuring the unique, or independent, effect of having or lacking health insurance is difficult at best, because investigations of this question are not con- ducted as randomized, controlled experiments, as are studies of the effect of a drug at a given dosage. Unlike clinical studies of the efficacy of a particular therapeutic intervention (e.g. aspirin after myocardial infarction), studies of the effects of health insurance on health-related outcomes do not randomly assign participants to experimental and control groups, where one group receives the âtherapeutic interventionâ of health insurance and the other group is denied coverage. Re- search ethics, over and above practical considerations, preclude this kind of pro- spective, experimental design in which benefits are withheld.8 Furthermore, health insurance or its absence is a characteristic that is closely intertwined with other features of individuals and population groups and with their physical and social environments. Specifically, it is intertwined with the very characteristicsânamely, education, income, employment and health statusâthat influence health insurance coverage (IOM, 2001a). Thus, it is impossible to ad- 8The notable exception to this generalization is the RAND Health Insurance Experiment, con- ducted between 1974 and 1982, which randomly assigned roughly 2,000 families in six different sites to one of 14 experimental health insurance plans that varied in their cost-sharing arrangements. Cost sharing ranged from none in the free care plan to 95 percent for all health services, limited to a maximum of $1,000 per family per year (in then-current dollars) with reduced amounts for low- income families (Keeler et al., 1985; Newhouse et al., 1993). The study was designed so that no participating family could be made worse off as a result of their participation than they would have been if they had not participated. This social experiment did not, however, include a study group without any health insurance at all.
22 CARE WITHOUT COVERAGE: TOO LITTLE, TOO LATE minister and measure a âpureâ dose of health insurance (and nothing else) and difficult to isolate its effect. The conceptual framework introduced in the Committeeâs previous report is adapted here to highlight and explicate the mechanisms by which health insurance influences the kind of health care received and health outcomes. Appendix A presents the Committeeâs general conceptual framework that was introduced in Coverage Matters (IOM, 2001a) as additional background to the discussion here. Figure 1.1 expands the relevant sections of the general framework and focuses on the relationship between health insurance status, the process of obtaining health care, and personal health-related outcomes. Coverage serves as a means to the end of obtaining services that may induce physiological (and other kinds of) changes that are reflected in various measures of health. Having health insurance does not affect oneâs health by eliciting a measur- able physiological response in those who possess it except, perhaps, by reducing stress and worry about covering medical expenses.9 Intermediate measures of the effect of health insurance on health, and thus mediating factors, include â¢ having a regular or usual source of care, â¢ the receipt of âappropriate care,â as validated by professional opinion and clinical effectiveness research, and â¢ the quality of care received. Health insurance plays both a critical and, at the same time, partial role in facilitating the receipt of appropriate, high-quality health care (Bindman et al., 1996; Eisenberg and Power, 2000). Because many characteristics other than health insurance status (including genetic factors, education, income, social class, personal health practices, beliefs about health care, and social and physical environment) affect the ultimate outcome of interestâhealth statusâdirectly, intermediate mea- sures are often used as proxies for health outcomes in studies of health insurance effects. These intermediate measures are represented by the boxes entitled âPro- cess of Careâ and âIntermediate Outcomesâ in Figure 1.1. The Committee considers both intermediate measures and final health out- comes as health-related outcomes to determine the impact of health insurance status on personal health. Health-related outcomes include self-reported overall health sta- tus and measures of evaluated health such as mortality, functional status, disability, stage of disease at the time of diagnosis, and a number of condition-specific physiological and behavioral indicators, such as blood pressure control for hyper- tension and blood glucose levels for diabetes.10 Figure 1.1 also includes a commu- 9There is a research literature on the relationship between psychological stress and healing; see Kiecolt-Glaser et al. (1998) for a description of the theory and evidence. 10Household or general population health surveys often include questions about participantsâ over- all health status. These subjective measures have been validated as corresponding well to health outcomes such as mortality (Mossey and Shapiro, 1982; Idler et al., 1990; Franks et al., 1993b).
Obtaining Access to Health Care Health-Related Outcomes Individual and Family Level Resources (e.g., health Decision-making Process of Care Intermediate Health Outcomes insurance status, income) Outcomes â¢Subjective health â¢ Individual, â¢ Presentation of Characteristics about use of illness or â¢ Timeliness of status Need health services condition diagnosis â¢ Functional/ â¢ Provider (e.g., â¢ Prevention â¢ Severity of disability status differential â¢ Management illness at â¢ Clinical markers management) diagnosis of disease for specific conditions Community Level â¢ Patient-provider communication â¢ Quality, â¢ General and Resources (e.g., health intensity, and disease-specific insurance coverage rates, site of care mortality safety net services) Characteristics Needs Effects on Communities: Health Disparities FIGURE 1.1 A conceptual framework for assessing the effect of health insurance status on health-related outcomes for adults. 23
24 CARE WITHOUT COVERAGE: TOO LITTLE, TOO LATE nity-level outcome, which emerged as an important finding in the course of conducting the research literature review, namely, the impact of health insurance on the extent of health disparities among racial and ethnic population groups. The Committee considers measures of the process of health care delivery, such as quality of care and the number or frequency of health care services, as health-related outcomes only if they have been endorsed as widely recognized professional guidelines or validated by science-based effectiveness studies as deter- minative of final health outcomes. For example, the recommendations of the U.S. Preventive Health Services Task Force regarding the frequency of Pap smears to detect cervical cancer establish the appropriateness criterion that the Committee uses in evaluating evidence from studies that compare the receipt of cervical cancer screening by insured and uninsured women. (See Box 3.2 in Chapter 3.) It is important to look at intermediate health outcomes such as receipt of cancer screening tests at the recommended intervals because many final health outcomes (e.g., diagnosed cancer or death) are not evident or realized for many years and very few studies of health insurance status extend over a long enough time to capture these outcomes. ORGANIZATION OF THE REPORT The Committeeâs critical assessment of the research literature is described, its findings are presented, and the implications of its findings and conclusions are discussed in the remaining chapters of this report. Chapter 2 outlines the reportâs analytic design and methods. First, it details the Committeeâs conceptualization of how health insurance status affects access to health care services, which in turn affects health outcomes. Next it identifies important covariates of health insurance, such as health status, race and ethnicity, and socioeconomic status and considers analytic approaches to disentangling their effects on health-related outcomes. Chapter 3 presents the Committeeâs findings based on its literature review. These findings are summarized within categories that reflect both type of service (e.g., preventive services, disease management for chronic conditions, hospital- based care) and condition-specific outcomes (e.g., stage of disease at detection and mortality rates for cancers), in addition to overall measures of health and mortality for large population samples. Chapter 4 employs the findings enumerated in Chapter 3 to draw conclusions about the effects of coverage on health outcomes for specific population groups that are at particular risk of adverse effects from lacking coverage and about the features of health insurance that account for its effects on health outcomes. Last, the Committee considers the potential benefits of extending health insurance to the U.S. population of uninsured adults.