BOX 4.1

Organization of Chapter

The Value Lost in Healthy Life Years

  • Estimating Health Capital

  • Imputing a Value for a Year of Life

  • Results

Quality of Life and Security for Families

  • The Value of Avoiding Risk and Uncertainty

  • Peace of Mind

Developmental Outcomes for Children

Uninsurance and Public Programs

  • Savings to Medicare

  • Disability Income Support

  • Reductions in Justice System Costs for Uninsured Persons With Severe Mental Illness

Workforce Participation, Productivity, and Employers

  • Workforce Participation

  • Employment-based Health Insurance

Health Systems Impacts

  • Access to and Quality of Health Care

  • Public Health System Capacity

Summary



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 62
BOX 4.1 Organization of Chapter The Value Lost in Healthy Life Years Estimating Health Capital Imputing a Value for a Year of Life Results Quality of Life and Security for Families The Value of Avoiding Risk and Uncertainty Peace of Mind Developmental Outcomes for Children Uninsurance and Public Programs Savings to Medicare Disability Income Support Reductions in Justice System Costs for Uninsured Persons With Severe Mental Illness Workforce Participation, Productivity, and Employers Workforce Participation Employment-based Health Insurance Health Systems Impacts Access to and Quality of Health Care Public Health System Capacity Summary

OCR for page 62
4 Other Costs Associated with Uninsurance In this chapter the Committee considers costs other than the direct costs of health care—opportunity losses—that uninsured individuals, their families, local communities, and the nation bear due to the lack of continuous and permanent health insurance coverage for the entire population. In contrast with the estimates of the costs of medical care provided to those without insurance, the costs considered in this chapter have not, in most cases, been studied systematically. The Committee has developed quantified estimates only for two kinds of internal or private opportunity losses: the worse health attributable to lacking coverage and the financial “exposure” faced by uninsured individuals that would be eliminated by health insurance. The Committee reserves one important, noneconomic opportunity loss for consideration in Chapter 6. In its conclusion to the report, the Committee examines the implications of extensive uninsurance nationally for American social and political values and ideals, including those of mutual caring and concern, equality of respect among members of a democracy, and equality of opportunity. This normative discussion is cast in terms of the benefits that could be expected if everyone had comparable financial access to care and the security afforded by health insurance. Box 4.1 provides a roadmap to the organization of the chapter. Following the schema of consequences of uninsurance that are presented in Figures 2.1 and 2.2, here the Committee relates its findings from previous reports and considers additional information from cost-of-illness and productivity studies to draw qualitative conclusions about certain economic impacts of lacking coverage and of relatively high rates of uninsurance within communities. The first two sections present quantified findings and the last four sections qualitative findings.

OCR for page 62
The first section focuses on the findings from Care Without Coverage and Health Insurance Is a Family Matter concerning health and health care deficits among uninsured adults and children and excess morbidity among uninsured adults resulting from a lack of health insurance. The Committee estimates the value of the increased lifespan and better health that uninsured persons could be expected to experience if they had continuous health insurance coverage over the course of their lives. This estimate is based on a commissioned analysis that is included as Appendix B of the report (Vigdor, 2003). The presentation of the qualitative findings notes when this quantified estimate of the probable value of healthier years of life forgone encompasses and subsumes the costs described qualitatively later in the chapter and when it does not. The second section considers the quality of family life and financial security, building on the Committee’s investigation of these issues in Health Insurance Is a Family Matter. Here the Committee presents its estimate of the monetary value of the financial risk protection that health insurance would provide to those now without coverage, both per capita and in the aggregate. The third section reviews findings from Health Insurance Is a Family Matter on children’s health and developmental outcomes as they depend on receipt of adequate services facilitated by insurance coverage. The fourth section looks at the implications of uninsurance for public program spending, including those that represent economic transfers as well as those that entail real resource costs. The fifth section considers productivity and workforce participation as related to health insurance status, primarily from the perspectives of the employer and the employee. Last, the sixth section summarizes findings from the Committee’s fourth report, A Shared Destiny, about health system and population health impacts of uninsurance at the community level. THE VALUE LOST IN HEALTHY LIFE YEARS Finding: The Committee’s best estimate of the aggregate, annualized economic cost of the diminished health and shorter life spans of Americans who lack health insurance is between $65 and $130 billion for each year of health insurance forgone. These are the benefits that could be realized if extension of coverage reduced the morbidity and mortality of uninsured Americans to the levels for individuals who are comparable on measured characteristics and who have private health insurance. These estimated benefits could be either greater or smaller if unmeasured personal characteristics were responsible for part of the measured difference in morbidity and mortality between those with and those without coverage. This estimate does not include spillover losses to society as a whole of the poorer health of the uninsured population. It accounts for the value only to those experiencing poorer health and subsumes the losses to productivity that accrue to uninsured individuals themselves.

OCR for page 62
The Committee approached placing a value on the difference in health outcomes that health insurance would make for those who lack it using the concept of health capital. As described in Chapter 2, health capital is the value to the individual of the “stock” of health that he or she can expect to have over the remaining course of life. Health capital encompasses and subsumes the market-based valuation of an individual’s productive capacity over the years of labor force participation that is represented by the notion of human capital. The capacity to be economically productive and to earn income is some, but not all, of the reason that we value good health. In order to account more fully for the value of better health, the Committee chose to base its estimates on the broader metric of health capital. The value ascribed to health capital, the present value of the stock of health that one will experience over the remainder of one’s lifetime, is the value the individual herself places on life in particular states of health. Most of the benefits that can be expected to accrue to uninsured individuals themselves if they were to gain health insurance, including increased productivity and labor force participation, and improved developmental outcomes among children (with whom the formation of human capital begins), are represented in the single estimate of gains in health capital. It does include, but goes beyond, the value we attach to being alive rather than being dead. It does not, however, include the value that others may ascribe to an individual’s particular state of health. The benefits realized by families, such as greater financial security and less stress and worry about health care and coverage, may well be additional to those that are accounted for in the aggregate estimate of gains in healthy life years because these benefits are interpersonal, not individual. The Committee commissioned an analysis by economist Elizabeth Richardson Vigdor to estimate the value of diminished health and longevity among the 40- some million persons who lack health insurance.1 Vigdor measures health capital empirically by combining data on length of life, the prevalence of adverse conditions among those alive, and the health-related quality of life conditional on having these conditions for insured and uninsured populations. She used the Current Population Survey (CPS) for determining the size and demographic composition of the uninsured population and the National Health Interview Survey (NHIS) for morbidity information. Relative mortality rates for insured and uninsured populations were taken from the Committee’s systematic literature review of health outcomes as a function of health insurance status presented in Care Without Coverage and Health Insurance Is a Family Matter (IOM, 2002a, 2002b). Based on its earlier work, the Committee chose to use a point estimate of a 25 percent greater mortality risk for 1   This analysis used the latest Current Population Survey (CPS) estimates of uninsurance available at the time, which were for 2000, reported in September 2001 (Fronstin, 2001b). For 2001, CPS reported increased numbers and age-specific proportions of uninsured Americans (Fronstin, 2002; Mills, 2002).

OCR for page 62
uninsured individuals between ages 1 year and 65 years compared with those insured. This value reflects the results of two observational longitudinal studies (Franks et al., 1993; Sorlie et al., 1994) that adjusted for demographic and socioeconomic characteristics (and in the former for multiple health status and health behaviors as well). The confidence intervals around the mortality differentials found in these two studies are large; hence the point estimate of 25 percent is uncertain. However, these two studies’ results are reinforced by multiple crosssectional studies of disease and condition-specific mortality rates as a function of health status that are also part of the literature review, and the Committee believes this assumption is reasonable. Estimating Health Capital In her analysis, Vigdor uses alternative assumptions that bound the range of likely values of health capital. In the first set of estimates, the years of life (YOL) approach, everyone who is alive is assumed to be in perfect health. Thus the difference in health capital for the insured and the uninsured is due exclusively to the differences between their adjusted mortality rates, as reported previously by the Committee in Care Without Coverage (IOM, 2002a). The second approach incorporates morbidity information to determine quality-adjusted life years (QALYs), but assumes that the insured and uninsured populations have the same disease prevalence and health-related quality of life (HRQL). This approach provides a lower bound of the loss in morbidity-adjusted health capital because, in fact, the uninsured are likely to have additional morbidity as a result of lacking coverage. Vigdor derived HRQL weights empirically from population-wide prevalence estimates for 15 conditions (from the NHIS and the Surveillance, Epidemiology and End Results databases). The third approach took into account variations in disease prevalence and HRQL by insurance status. Because the uninsured differ from the insured across several characteristics, however, the observed difference in morbidity is unlikely to be entirely caused by lack of health insurance. Even after controlling for differences between the two populations in terms of age, gender, race and ethnicity, and educational attainment, unobserved differences that are correlated with higher morbidity among the uninsured likely remain. Therefore this estimate serves as an upper bound to the possible health gains from insurance. In this upper-bound approach, age and sex-specific prevalence estimates were obtained from the NHIS, controlling for insured status, race, ethnicity, income, urbanicity, and region of the country. In age-sex groups where the insurance status coefficient was not statistically significant at p = 0.10, no difference in prevalence by insurance status was assumed for that condition. The HRQL weights were also allowed to vary by insurance status, although most of these interactions were not significant. Notably, being uninsured had a significant negative effect on HRQL even after controlling for other factors. This result is reflected in the health capital estimates as well. After calculating health capital as a function of insurance status under each set

OCR for page 62
of assumptions, Vigdor determined how much health would be gained if the current population of uninsured individuals had coverage. The analysis assumes that those without coverage acquire the same mix of insurance policies as those currently insured and that the previously uninsured maintain coverage until turning 65 and becoming eligible for Medicare. The coverage scenario presented in this chapter assumes that each individual faces the average probability of being insured each year until reaching age 65. This results in a conservative estimate because it assumes (1) that there is no correlation between current insurance status and insurance status next year, and (2) that the overall rate of uninsurance does not increase in the future. If being uninsured currently increases the probability of being uninsured in the future, and if length of uninsured spell has an adverse impact on health outcomes, this scenario underestimates the potential gains from health insurance. The actual mortality differential between insured and uninsured individuals could be either less than or greater than the estimated 25 percent, and the economic value estimated for health insurance is approximately proportional to this estimate. If unmeasured personal characteristics account for some of the mortality differential, this would lower the estimate of health capital attributable to health insurance. On the other hand, several disease- and condition-specific outcomes studies, including studies of breast and prostate cancer and HIV infection report higher mortality differentials (Ayanian et al., 1993; Lee-Feldstein et al., 2000; Roetzheim et al., 2000a,b; Goldman et al., 2001). Imputing a Value for a Year of Life Each of the benchmark estimates of health capital assumes a value of $160,000 per year of life in perfect health. As noted in Chapter 2, the value of a life year is taken from a survey of the literature by Hirth and colleagues (2000) and is the mean value of the estimates they obtained from a number of contingent valuation studies. Contingent valuation is a survey-based methodology that captures the value of intangible benefits (such as the intrinsic value of having good health) and benefits that are not traded in a competitive market. The results from contingent valuation studies are not driven solely by the respondent’s ability to pay. Vigdor notes that ascribing a value of $160,000 to a year of life in perfect health corresponds to an average value of $4.8 million for a statistical life, assuming a 3 percent discount rate and a life expectancy at birth of 76 years. This places the $160,000 value in the mid-range of values ($3 to $7 million for a statistical life) used by federal regulatory agencies to evaluate the economic costs and benefits of risk reduction and life-saving interventions (Vigdor, 2003). In this analysis, contingent valuation measures the entire health benefit to an individual (and thus incorporates gains in productivity or productive capacity that accrue to the individual), but does not capture any external effects that an individual’s improved health might have for society. The analysis gives equal value to every life year across people, ages, and time. Thus, by assuming that a year of

OCR for page 62
each person’s life is equally valuable, this approach builds in a substantial element of equity. Following the recommendations of the Panel on Cost-Effectiveness in Health and Medicine (Gold et al., 1996b), the analysis uses a 3 percent annual discount rate for costs and benefits occurring in future years. Tables in Appendix B also present results using 0 and 6 percent annual discount rates, producing a wider range of estimated values. Results Table 4.1 presents the summary results of the commissioned analysis. Using the benchmark assumptions just described, the approximate present value of future forgone health to an uninsured 45-year-old male is $10,300 using the YOL approach, and between $8,300 and $12,600 when morbidity is incorporated (QALY approach, upper and lower bounds). For a male newborn, the figures are $7,700 under the years of life method, and $6,600 to $15,600 using the QALY approach. The value of future forgone health to an uninsured 45-year-old female is $7,800 for the YOL method, and between $6,200 and $10,500 once morbidity is incorporated. An uninsured baby girl forgoes health worth $4,600 under the YOL method and $3,900 to $11,600 under the QALY method. The differences in estimates by sex reflect differences in life expectancy and morbidity, not differences in earnings. These numbers add up to a very large aggregate cost across the currently uninsured population of approximately 40 million, with estimates of the total discounted present value of this cohort’s forgone health ranging from $250 billion to $500 billion. (See Table B.11 in Appendix B.) These estimates range over a wide interval and depend critically on the assumptions used in the analysis. It is not possible to select a single set of assumptions to represent the adverse health consequences of uninsurance. The lower-bound number likely underestimates the size of the effect that health insurance status has on health outcomes and the higher-bound number likely overstates this effect. These estimates can also be construed in terms of the expected gain in value of statistical healthy years per year of additional insurance provided. Because health insurance is an investment in future health, and most adverse health outcomes occur at older ages, the benefit per year of insurance provided rises quite sharply with age. In this exercise, the value of future health gains is discounted to the present annual value. The benefit per year of insurance for a 45-year-old male ranges from approximately $3,100 to $4,800 under the different scenarios, while for a 45-year-old woman it ranges from $2,100 to $3,600. In contrast, the health capital gain per year of additional insurance for a newborn boy (discounted to present value at 3 percent per year) ranges from $1,200 to $2,800, and for a newborn girl the gain is between approximately $750 and $2,300.2 It is important 2   These values are presented in Tables B.9 and B.10 in Appendix B.

OCR for page 62
to remember that these annualized values are predicated on continuous health insurance coverage. Based on its previous review of studies, the Committee found that health measures and outcomes for those with intermittent coverage are more similar to those for uninsured than to insured counterparts (Schoen and DesRoches, 2000; Baker et al., 2001; IOM, 2002a). Using the actual age distribution of those who reported being uninsured in 2000, the average value of health capital (i.e., quality-adjusted years of life) that could be gained with a year of insurance for a member of this population ranges from $1,645 to $3,280. (See the next-to-last row in Table 4.1.) Aggregating over the population of uninsured individuals (40 million individuals × $1,645 or $3,280), the annualized value of health capital lost through uninsurance is in the range of $65–$130 billion. (See the last row in Table 4.1.) QUALITY OF LIFE AND SECURITY FOR FAMILIES Finding: Uninsured individuals and families bear the burden of increased financial risk and uncertainty as a consequence of being uninsured. Although the estimated monetary value of the potential financial losses that those without coverage bear is relatively small (compared to the full cost of their services) because of uncompensated care, the psychological and behavioral implications of living with financial and health risks and uncertainty may be significant. Health insurance confers health and financial benefits on families. It also improves their well-being in ways that extend beyond their individual health status and family finances. Having family members insured reduces the extent of tradeoffs that families must make between health care and other uses of their money. Even if all members are healthy today, health insurance reduces the stress and uncertainty about future medical care needs and financial demands that can accompany the lack of coverage. Families in which all members have health insurance do not experience the worries, demands, and indignities that accompany illness without coverage (Kaiser, 2000; Andrulis et al., 2003). The lower incomes of families who are uninsured further constrain their financial choices. As incomes rise, families have more to spend both on necessities and on discretionary purchases. The lower income among families with uninsured members means the payment for a doctor visit represents a larger share of income than it does for the typical family where all members have health insurance. Gaining health insurance would relieve some of the impact of health expenses on family budgets. The Value of Avoiding Risk and Uncertainty Health insurance reduces families’ risks and uncertainty regarding future health care costs. People with health insurance benefit from less unpredictability and

OCR for page 62
TABLE 4.1 Increases in Health Capital with Health Insurance for Those Currently Uninsured, by Sex and Agea   Change in Health Capital   Benefit per Year of Insurance Average pr (ins) Until 65b   Average pr (ins) Until 65b YOL Approach Lower-Bound QALY Approach Upper-Bound QALY Approach   YOL Approach Lower-Bound QALY Approach Upper-Bound QALY Approach   (dollars)   Men   0 7,716 6,567 15,572   1,408 1,198 2,842 18 10,715 8,942 19,136 1,841 1,536 3,288 25 10,652 8,789 17,680 2,197 1,812 3,646 35 10,581 8,612 15,362 2,952 2,403 4,286 45 10,268 8,271 12,638 3,872 3,119 4,766 55 8,433 6,732 9,075 4,466 3,565 4,806

OCR for page 62
Women 0 4,577 3,864 11,646 893 754 2,273 18 5,874 4,826 13,018 1,120 920 2,482 25 6,427 5,241 12,180 1,427 1,164 2,705 35 7,258 5,864 11,528 2,009 1,624 3,191 45 7,772 6,225 10,544 2,661 2,131 3,610 55 7,076 5,618 8,186 3,267 2,594 3,780 Overall average per uninsured personc 2,014 1,645 3,280   (dollars in millions)   Total for 40 million uninsured 80,560 65,800 131,200 aCalculations assume a value of a life year of $160,000 and a real discount rate of 3 percent. bAssumes that uninsured individuals face the average probability of being uninsured until reaching age 65. cCalculated by summing the average difference in health capital by age and sex over all uninsured people under age 65 in the Uni ted States in 2000. SOURCE: Vigdor, 2003.

OCR for page 62
variability in their out-of-pocket spending for health care. Lower variability itself is a benefit to people who do not like to face financial risks. Those who are “risk averse” care both about the expected cost of a risky event as well as about how variable those expected costs are. The value of the reduction in the risk of financial loss that health insurance provides is calculable empirically (Buchanan et al., 1991). The Subcommittee on the Societal Costs of Uninsured Populations followed the methodology of Buchanan and colleagues to construct an estimate of the value of the risk borne by the uninsured.3 The insurance value of coverage was estimated by applying a constant relative risk aversion parameter (0.00024) to the reduction in the variance of out-of-pocket spending obtained through insurance coverage.4 While risk reduction is one consideration in valuing health benefits, it turns out to have a small value for those without health insurance because of the kinds of out-of-pocket costs the uninsured actually pay. (See the discussion in Chapter 3 and Figure 3.1 on out-of-pocket costs among uninsured individuals.) There is much more variance in the total expenses incurred by persons without health insurance than there is in their out-of-pocket payments. In the sample of respondents to the Medical Expenditure Panel Survey (MEPS), the highest total annual health expense observed for an uninsured individual was nearly $500,000. The highest amount reported paid out of pocket by an uninsured individual, however, was $26,000 (although less than 1 percent of those without insurance spent $5,000 or more).5 If buffering mechanisms such as charity care, bad-debt writeoffs, and bankruptcy did not exist, and each uninsured person was compelled to pay for the health care he or she used (through loans, for example), the per capita value of the risk reduction from having health insurance would be more than $2,100. Because out-of-pocket expenses for the uninsured show less variation than their total expenses, the value of the reduction in variance is proportionately lower. The predominant source of variation, inpatient costs, is not being paid out of pocket by uninsured individuals. They tend, instead, to pay out of pocket for the smaller elements, and inpatient costs are borne by a combination of public and charitable support and absorbed as hospital bad debt. In effect, a large share of the overall “financial exposure” has been shifted to those who ultimately pay for uncompensated care. For this reason, the estimated value of the risk that uninsured individuals bear (and that they would not have if insured) is between $40 and $80. This modest amount represents an annual aggregate cost borne by families with uninsured members of $1.6 to $3.2 billion. 3   Calculations by Sherry Glied, Columbia University, from merged MEPS files for 1996, 1997 and 1998 prepared by Jack Hadley, Urban Institute. 4   This value represents the midpoint of the range identified by Szpiro (1986) and is comparable to the estimate used by Manning and Marquis (1996). 5   These annual expenses measured by MEPS do not include fees and interest charges that may be added by collection agencies for late payment. See Lagnado (2003a) for a recent account of such cumulative billings.

OCR for page 62
BOX 4.5 Recent Trends in Employer Health Insurance Coverage Even though the number of firms offering health insurance to at least some of their workers has increased in recent years, the proportion of workers with an offer of coverage has decreased. Between 1979 and 1998, the proportion of workers with insurance provided by their own employer fell from 66 percent to 54 percent (Medoff et al., 2001). This decline accompanied a shift in that 20-year period from manufacturing jobs to service jobs and from large employers to small employers, who in either case are less likely to offer insurance to their employees. Shifts in industry and occupation accounted for about 30 percent of the decline in employment-based coverage. The rest was the result of a widespread drop in coverage in nearly all industries. See Figure 4.1 for changes in employment-based coverage over time. The decline in coverage among wage earners is not distributed uniformly across the labor force, however. Over the period 1979–1998, the percentage of private-sector workers aged 21–64 with insurance from their own employer fell from 72 to 60 percent.1 The percentage of workers with such insurance in the highest income quintile fell from 90 to 80 percent. The percentage of workers with such insurance in the lowest 20 percent of the wage distribution fell from 42 to 26 percent. As posited by economic theory, the decline in the insurance status among workers is larger for the low-wage workers whose productivity increases are likely to be smaller than the increase in premiums. Several studies have confirmed that declining take-up rates of employment-based coverage are a major component of overall declines in coverage (Cooper and Schone, 1997; Farber and Levy, 2000; Cutler, 2002). Examining the change in employment-based coverage over the past decade, David Cutler (2002) found that the proportion of the U.S. population under age 65 with such coverage fell from 71 percent in 1987 to 68 percent in 2000, while the proportion of the same population that was uninsured increased by 3 percentage points or 7.2 million persons. While the share of workers in firms offering insurance to at least some of its workers remained roughly constant between 1988 and 2001 at about 80 percent, and the eligibility for coverage in firms offering benefits declined only slightly from 93 to 91 percent over the same period, the take-up rate among eligible workers declined from 88 to 85 percent. Among full-time, full-year male workers, for whom offers of and eligibility for coverage are higher than for other workers, the take-up rate among those eligible declined even more over this period, from 94 to 90 percent. Cutler’s analysis attributes 61 percent of decline in workers’ coverage from their own employer to changes in take-up and the remainder to changes in eligibility. Among full-time male workers, changes in take-up accounted for 80 percent of the decline in own-employer coverage. Using data from employer surveys over this period to model the change in take-up rates as a function of the health insurance premium price faced by the employee, Cutler concluded that nearly all of the decrease in the take-up of insurance coverage between 1988 and 2001 could be attributed to increases in employee share of premiums over this same period. 1   This calculation does not take into account those with health insurance through a family member’s employment benefit.

OCR for page 62
FIGURE 4.1 Percent of private-sector employees with insurance from own employer, by hourly wage quintile, 1979–1998. NOTE: In 1998, the lowest 20 percent earned less than $7 per hour and the highest 20 percent earned more than $21 per hour. SOURCE: Medoff et al., 2001. Data from the Current Population Surveys: May 1979, 1983, 1988; March 1996, 1999. In their review of the history of health insurance in the United States, Currie and Madrian (1999) observe the following: . . . the genesis of employer-provided health insurance is rooted in employment-based programs implemented precisely because health impacts labor market activity and labor market activity impacts health (p. 3365). Although employees have reasons to want health insurance as part of their compensation package and employers also have reasons to provide this workplace benefit, the extent to which employers realize tangible financial benefit from having an insured workforce is not well documented. Two recent surveys of health services and economics research identify the possible reasons that employers might be willing to accept higher production costs in order to provide health insurance to their employees and review the evidence for this proposition (Buchmueller, 2000; O’Brien, 2003). These review articles encompass and expand on much of the discussion that follows. Having health insurance as part of the offered wage may help employers to attract employees more easily. However, employment situations that provide health insurance benefits may attract workers in relatively worse health or those with sick dependents, which can increase an employer’s group premium rate. Still, even healthy employees may value health insurance at more than the forgone wages. Second, health insurance may also help employers to retain workers once they are recruited, although insurance may not be more effective in retaining employees than the equivalent wage (Gruber and Madrian, 2001). Because chang-

OCR for page 62
ing jobs often entails an interruption in or loss of coverage during which the employee or his family may be financially exposed to uninsured health expenses, an aversion to this risk may help keep workers in a firm. This phenomenon, called job lock, may or may not benefit the employer. If it increases retention of valued workers, it is positive for the employer. On the other hand, if an employer is unable to attract a desired worker from another firm because of the preferred set of health insurance benefits offered by the current employer, the recruiting employer may feel disadvantaged, and offering benefits can be viewed as providing an advantage to the current employer. Health insurance may enhance worker effort and productivity because workers may feel that insurance is part of the package that makes a job a desirable one. Employees may work harder if they believe it would be difficult to find an equivalent or better paying job that included health insurance. Health insurance, then, may be considered an employer investment in employees along with education and training, services offered to employees, and general workplace environment for the purpose of maintaining morale and retaining workers. As Buchmueller notes, however, a particular employer will not necessarily reap the benefits of her investment in the better health consequent to offering health insurance to employees; to the extent the worker has the ability to move between employers, another employer may benefit from the worker’s enhanced productivity. If long-term health gains can be realized from having health insurance, the benefits would only accrue to the employer if employees were retained for a number of years. Employers that experience relatively high turnover in their workforce would likely continue to do so even if they offered health insurance coverage (Buchmueller, 2000). Employer Surveys Employer surveys and business management literature reveal that employers believe that health insurance contributes positively to firm performance. Three-quarters of small employers that offer health benefits reported in one national survey that these benefits had a positive effect on recruitment, on employee retention, or on employee attitudes and performance. About two-thirds believed that health benefits contributed to better employee health, and more than half reported that the benefits helped reduce absenteeism (Fronstin and Helman, 2000). The Employee Benefit Research Institute and the Consumer Health Education Council recently conducted a Web-based, nonrepresentative survey of 800 firms of all sizes (representing 3 million full- and part-time workers) to determine employer attitudes and policies regarding workplace health benefits (Christensen et al., 2002). Virtually all respondents (97 percent) represented firms that offered health insurance to full-time employees and a third of them to part-time employees. Eighty percent of the respondents identified health benefits as “extremely” or “very important” in recruiting and retaining workers. Forty percent thought that

OCR for page 62
such benefits were extremely or very important for improving worker productivity. Health Benefits and Workplace Productivity Studies Although employers who offer health benefits view them positively, the evidence for improved employee health and productivity as a result of health insurance is limited and mixed. Studies of absenteeism rarely include health insurance as an explanatory variable and provide little support for the notion that absenteeism is decreased by health insurance. One simulation study indicates that physician visits and work absences substitute for each other. That is, ill workers have fewer days absent if they receive medical attention, which requires taking time off from work, and they are more likely to receive medical attention if they have health insurance. Health insurance coverage and sick leave benefits together, however, increase both the number of workers taking sick days and the number of sick days taken (Gilleskie, 1998). Thus, having health insurance cannot be shown to reduce absenteeism. Studies have demonstrated that impaired health is related to absenteeism and reduced productivity (Chirikos and Nestel, 1985; Greenberg et al., 1995; Berndt et al., 1997; Bound et al., 1999; Druss et al., 2001; Fronstin and Holtmann, 2000; Blau and Gilleskie, 2001; Kessler et al., 2001a,b; Ramsey et al., 2002). Particular health care interventions have been demonstrated to make a difference for individual labor market outcomes, including labor force participation, hours worked, and earnings. Studies have shown that people in poor health or with specific illnesses (e.g., arthritis, depression or other psychological disorders, asthma, or chronic backache) work less and earn less than people in good health (Bartel and Taubman, 1986; Mitchell and Butler, 1986; Mitchell and Burkhauser, 1990; Berndt et al., 1997; Ettner et al., 1997; Greenberg et al., 1999; Kessler et al., 1999; Berndt et al., 2000; Birnbaum et al., 2002). Other studies have found that workers in poor health are more likely to quit work or retire early than are workers in better health (Diamond and Hausman, 1984; NAAS, 2000). Still, the measurement of workplace productivity in relation to health and as a function of particular health interventions is a relatively young field. Greenberg and colleagues (1995) suggest that employers can take either a narrow view of illness-related costs in the workplace or a much broader view. In the former case, employers focus only on their out-of-pocket costs related to illness and health care, including employer health insurance premium contributions, employer contributions to the Medicare trust fund, workers’ compensation and temporary disability insurance, and in-house health services. In the latter case, adopting a broader view, employers take into account indirect costs related to their decisions about health care. These indirect costs include productivity effects, in terms of both performance on the job and health-related absences. These authors propose a model to estimate the value of employee productivity lost to illness rates, based on both the prevalence of the illness in the workforce (the

OCR for page 62
impairment rate) and the annual percentage of work time affected by the illness. Obviously, both of these factors vary by the illness in question. Particular employers also must consider the wage profiles of the employees affected in order to calculate their specific costs. A number of management tools for the measurement of workplace productivity are currently under development (Lynch and Riedel, 2001). Experts in the field of health and productivity management stress the importance of focusing analyses relatively narrowly on particular health conditions and job performance requirements and criteria if these tools are to be useful to managers. Although this field of inquiry is promising, it has little to contribute at this time to informing an employer’s decision to offer workers health insurance benefits. The benefits of having healthier workers may include reductions in other labor costs, especially long-term and short-term disability insurance rates and workers compensation costs. Studies of workers compensation and health insurance fail to show significant reductions in these related costs, however (Card and McCall, 1996; Buchmueller, 2000). Experience rating of workers compensation premiums for large firms might show small reductions, but virtually all those firms offer health insurance to at least some of their employees anyway. Small firms that do not offer health insurance are also not paying workers compensation premiums that vary with their own employees’ claims experience. The Small Group Market The practices and policies of insurers that sell in the small group market also figure in the decisions of some firms to offer health insurance. These insurers require firms perceived as having higher than average medical risk to pay higher premiums, relative to the value of the coverage purchased. Nearly every state has enacted legislation to curb the more extreme of these practices. These reforms have neither resulted in the increase in coverage of workers hoped for by the reformers, nor in the decrease feared by the critics of those reforms. Overall, the impacts of reform legislation have not been large. One reason for this may be that small employers are not aware of the reforms and may continue to have a distorted impression of the barriers to their offering insurance (Fronstin and Helman, 2000; Mulkey and Yegian, 2001). Education of small employers might help to expand coverage, although its potential impact is limited. Many small employers are also employers of relatively low-waged workers. The cost of health insurance benefits represents a proportionately larger share of low-waged workers’ total compensation package than it does for higher-waged workers. Both the employers of low-waged workers and low-waged workers themselves may be reluctant to trade off take-home pay for health benefits (Hadley and Reschovsky, 2002).

OCR for page 62
Summary Evidence shows that, other things being equal, having health insurance im-proves the health of working-age Americans (Hadley, 2002; IOM, 2002a). Nonetheless, the information available does not lead to the conclusion that any productivity benefits accruing to individual employers of insured workers are sufficient to induce them to offer health insurance when offering it increases their payroll costs. Worker demand for health insurance is the primary determinant of whether or not a firm offers it (Buchmueller, 2000; Christensen et al., 2002). Those employers that provide this workplace benefit have demonstrated the value that they ascribe to it through their action. The “business case” probably cannot be made, however, for the group of employers that now do not offer it to any workers, predominantly smaller and lower-wage firms. It is unlikely that additional small employers can be induced to offer health insurance to their workers without additional public subsidies. HEALTH SYSTEMS IMPACTS Finding: Not only those who lack coverage but others in theircommunities may experience reduced access to and availability of primary care and hospital services resulting from relatively high rates of uninsurance that imperil the financial viability of health care providers and institutions. In addition, population health resources and programs, including disease surveillance, communicable disease control, emergency preparedness, and community immunization levels, have been undermined by the competing demands for public dollars for personal health care services for those without coverage. Uninsurance throughout the United States at its present level (16.5 percent of the population below the age of 65) has deleterious effects on the financial stability of health care providers and institutions and may affect the availability and quality of health care services not only to those who lack coverage but also to others who share common health care facilities and community resources (IOM, 2003a). Although extensive insurance coverage has led to excess capacity in the health care system over the past three decades, this trend has recently been countered by its converse. In this section, the Committee summarizes its findings, presented in A Shared Destiny: Community Effects of Uninsurance, regarding two aspects of health care services that are adversely affected by uninsurance: the availability and quality of personal health care services within communities and the ability of public health agencies to perform their core mission of protecting population health. Access to and Quality of Health Care As discussed in Chapter 3, health care practitioners and institutions provide substantial amounts of uncompensated care to uninsured patients. The cost of this

OCR for page 62
care is ultimately borne by providers themselves as bad debt and donated services, local, state, and federal taxpayers; and, to a much lesser extent, organized philanthropy. Where support and subsidies for care provided to uninsured patients is not adequate, physicians, clinics and hospitals may cut back or withdraw services from areas with large uninsured populations, affecting access to and quality of care to local residents more broadly. A Shared Destiny assesses available evidence about what happens, and lays out the Committee’s hypotheses about what reasonably can be expected to happen within communities when one factor, the local rate of uninsurance, is relatively high or rising. This section reviews the findings from that report and suggests how health insurance coverage for the whole population could improve the availability and quality of health care within communities. The effects of uninsurance at the community level are components for assembling the national picture of spillover costs. These effects are often easier to detect locally than nationally, where the aggregation of information averages out marked local variation in the organization, financing, and delivery of health care. One key causal pathway by which uninsurance affects communities is lower provider revenues resulting from the combination of less use of services by the uninsured compared with that of insured persons and the costs of uncompensated care that providers incur when uninsured patients receive services for which they cannot pay. Uninsurance may affect the availability of health services within communities. In an effort to avoid the burden of uncompensated care or to minimize its impact on the financial bottom line, health care providers may cut back on services, reduce staffing, relocate, or close. Already overcrowded hospital emergency departments may be further strained as they increasingly serve as the provider of first and last resort for uninsured patients. Physicians’ offices or even hospitals may relocate away from areas of towns or entire communities that have concentrations of uninsured persons. Especially for institutions that serve a high proportion of uninsured patients such as center-city community hospitals or academic medical centers, a large or growing number of uninsured persons seeking health care may “tip” a hospital’s or clinic’s financial margin from positive to negative. The quality of care for both uninsured and insured persons may be adversely affected by uninsurance within the community. The IOM Committee on the Quality of Health Care in America describes the goals for health care in the United States as a systematic approach to care that is safe, effective, patient centered, timely, efficient, and equitable for all Americans, irrespective of insurance status (IOM, 2001b). While the Committee’s second and third reports, Care Without Coverage and Health Insurance Is a Family Matter, have documented the lesser effectiveness of health care received by the uninsured, its fourth report considers how high uninsured rates undermine the capacity of health care institutions to provide high-quality care more generally. A Shared Destiny documents reduced availability within the community to clinic-based primary care, specialty services, and hospital-based care, particularly emergency medical services and

OCR for page 62
trauma care, in areas with relatively large uninsured populations (Gaskin and Needleman, 2003; Needleman and Gaskin, 2003). Reduced access to primary care increases the demand for services by both insured and uninsured persons in already overcrowded hospital emergency departments (EDs) (Derlet, 1992; Grumbach et al., 1993; Baker et al., 1994; Billings et al., 2000). In 1986, the Emergency Medical Treatment and Labor Act (EMTALA), which was conceived to counter the practice of hospitals turning away or inappropriately transferring patients who did not have the means to pay for their care, established a right to medical screening and stabilization and hospitalization, if necessary, for all patients presenting to EDs for treatment regardless of ability to pay (Bitterman, 2002). Thus hospital EDs may be one of the few health care providers to whom uninsured patients can turn when they seek routine or urgent, as well as emergency care. Large metropolitan areas and multicounty rural areas depend on highly specialized and resource-intensive care provided by trauma centers affiliated with EDs. The lack of adequate financing for the emergent care of uninsured ED and trauma patients risks diminished access for all residents of a region. In many urban and rural areas, hospital emergency departments are often filled beyond capacity, affecting all who rely on them (Richards et al., 2000; Derlet et al., 2001; Lewin Group, 2002). Relatively high uninsured rates are associated with the lessened availability of on-call specialty services to hospital emergency departments and the decreased ability of primary care providers to obtain specialty referrals for patients who are members of medically underserved groups (Asplin and Knopp, 2001; Bitterman, 2002). One strategic response of some hospitals to such cost pressures has been to eliminate specialty services with relatively high levels of uncompensated care, such as burn units, trauma care, pediatric and neonatal intensive care, emergency psychiatric inpatient services, and HIV/AIDS care (Gaskin, 1999; Commonwealth Fund Task Force on Academic Health Centers, 2001). In rural areas, all residents may experience lessened access to specialty care (as with primary care) if providers leave the community because of financially unviable practice conditions (Ormond et al., 2000). The health sector is a critical component of many local economies, particularly in rural areas, where hospitals serve as social, historic, and civic anchors, as well as economic engines. A high uninsured rate and the corresponding burden of uncompensated care on the local health care system may reduce the economic base of the community. The economic role of local hospitals is particularly influential in rural areas (Cordes et al., 1999; Colgan, 2002). Rural hospital closings result in the loss of community physicians and the jobs and tax revenues that private practices generate (Hartley and Lapping, 2000; Doeksen et al., 1997). Public financing, particularly through Medicare and Medicaid, is also especially important for the rural health services infrastructure (Cordes, 1998). Universal health insurance coverage could be expected to reinforce the health services

OCR for page 62
infrastructure similarly because it would provide a stable source of revenues for providers and one that corresponds to the size of the population served. Public Health System Capacity The sheer number of uninsured persons in an area adds to the community burden of disease and disability. Uninsured children and adults have diminished health status and greater likelihood of premature mortality, in part because they receive less timely and appropriate care. Geographic differences in self-reported health status across the states and across urban, suburban, and nonmetropolitan communities are negatively correlated with the local uninsured rate, suggesting unmet health needs (Cunningham and Ginsburg, 2001; Holahan, 2002). Areas with higher uninsured rates and correspondingly higher burdens of disease place relatively greater demands on the population-based public health services that health departments provide, because uninsured residents are more likely to rely on these services (e.g., immunization and well-baby clinics, treatment for sexually transmitted and other communicable diseases) than are residents with public or private health insurance (IOM, 2003b). Health departments in communities with higher uninsured rates, however, are also likely to face higher demands to provide personal health services to uninsured persons, in their role as providers of last resort. Budgets for population-based public health, including disease and immunization surveillance, emergency preparedness, environmental health, and restaurant inspections, which benefit all members of a community, can be squeezed by demands on health departments to provide or pay for safety-net services for uninsured persons, adversely affecting the capacity to deliver public health services to the community (IOM, 1988, 2003b). In many parts of the country, health department officials have expressed their perceptions of being caught between the increasing demand and need for care of growing numbers of uninsured persons and diminished budgets (Goldberg, 1998; Lewin and Altman, 2000). Some health departments have tried to address the unmet health needs of sizable or growing uninsured populations by shifting discretionary funds toward the delivery of health services at the expense of population-based public health programs (IOM, 2003a,b). One result of the cutting back in population-based public health activities is the risk of higher incidence and prevalence of vaccine-preventable and communicable diseases, especially in areas where health departments have been chronically short of funding. Communicable disease control is a core health department function that helps prevent the spread of disease through the screening, tracing and notification, and for some diseases, treatment, of persons with whom an infected individual has come into contact. For example, underimmunization increases the vulnerability of entire communities to outbreaks of diseases such as measles, pertussis (whooping cough), and rubella (IOM, 2000). Both categorical immunization efforts and public and private health insurance that provide these services to

OCR for page 62
their enrollees are needed to ensure adequate levels of immunization among children and adults (DeBuono, 2000; Fairbrother et al., 2000; Sisk, 2000). Uninsured children and adults are both the primary clientele of health department immunization programs and substantially tax the capacity of these public agencies to serve them. To give another example, tuberculosis (TB) is a preventable and treatable contagious disease whose spread may be exacerbated by community uninsurance. In past years linked with HIV/AIDS, substance abuse, and life in institutional settings (e.g., homeless shelters, prisons), TB more recently has been associated with low-income, foreign-born communities in urban areas, along the southwestern U.S. border, and in other medically underserved communities with high uninsured rates (Geiter, 2000; CDC, 2001; Kershaw, 2002). Timely diagnosis and appropriate treatment are integral to stemming the transmission of TB. Inadequate access to primary and follow-up care has contributed to the increase in drugresistant tuberculosis, which poses a public health threat (CDC, 1999). Screening for TB, contact tracing, notification, and oversight of the daily medical treatment that the disease requires are all part of the communicable disease control activities carried out by health departments. The diversion of resources from communicable disease control to the provision of personal health care services to uninsured and other medically indigent residents has been cited by health department officials as likely to result in increased numbers of persons with tuberculosis (Geiter, 2000). The diversion of resources from public and population health activities to personal health care for uninsured persons is particularly dangerous because this frequently occurs without explicit public knowledge or endorsement, but rather as the result of ad hoc resource management decisions within state and local health departments (Fairbrother et al., 2000; IOM, 2003b). Providing health insurance to those who now lack it will certainly require public resources in addition to those that health departments spend on personal health care for the same populations. Financing the personal health care of uninsured Americans through insurance mechanisms instead of as direct services would, however, make public resource allocations and programmatic tradeoffs more transparent than they are at present. SUMMARY In this chapter the Committee has examined a variety of internal (private) and external (societal) economic implications of the lack of insurance within the United States. Many of these cannot be quantified. The Committee has undertaken an innovative approach to gauging the probable magnitude of the opportunity costs associated with uninsurance within the American populace. By estimating the monetary value of the diminished health and longevity among uninsured children and adults, consistent with the practices of federal agencies that make policy choices about investments in health and safety, the Committee has exposed some of the hidden costs of uninsurance.

OCR for page 62
The following chapter presents estimates of the resources that would be needed to increase the health services utilization of uninsured children and adults to the levels used by their insured counterparts. These estimates of resource demands will be used in the concluding chapter, along with the estimates of the value of forgone healthy life years presented in this chapter, to consider the costeffectiveness of providing health insurance to those currently uninsured.