In 1994, U.S. health care expenditures approached $1 trillion, of which private businesses paid a substantial portion. While the debates on regulatory measures to contain these costs have not been resolved, the market has responded decisively to pressures from increased health care spending, evidenced by the broad shift to managed care. This trend has moved decision-making authority from providers and recipients of health care goods and services. Increasingly, those who pay for health care, including insurers, managed care organizations, and employers, are asserting their influence over the way treatment decisions are ultimately made.1

If more effective management can curtail cost escalation in health care, an employer can channel some resources otherwise allocated to health care for other uses. To capitalize on this opportunity, some employers have attempted to intervene directly in the workplace to address health care issues, which requires a clear understanding of the costs they currently bear. Whereas employers routinely calculate and report out-of-pocket health care costs, they generally do not understand the opportunity costs associated with health care decisions. Our purpose is to illustrate the importance of the “indirect” costs of illness in the workplace, especially the effects of impaired functioning on the job as well as absenteeism. Without paying proper attention to these additional costs, employers are unlikely to make the best health care expenditure decisions, from a systemwide perspective.2

Costs of Illness in the Workplace

Cost-of-illness analysis requires a specific vantage point. For example, we can examine the costs of a particular illness — direct and indirect — with respect to an individual sufferer. Alternatively, we can explore society’s stake in the illness using a slight variant of the same methodology. While direct costs include all the out-of-pocket expenditures incurred for treatment, indirect costs include the lost output that results from illness. This latter category includes, but may not be limited to, lost resources due to premature death.

Although this type of analysis usually takes the perspective of the individual or of society, we can focus on the vantage point of other agents or groups of interested parties as well. Our discussion takes the perspective of employers who want to ascertain the cumulative resource depletion from illness in the workplace. A specific firm can easily develop a simple extension of the same analysis, based on that company’s unique features and the demographics of its workforce. Alternatively, a company can apply a similar approach to a range of illnesses that each may affect workplace costs in different ways.

Our discussion repeatedly refers to two types of employers, one with a very narrow view (employer A) and the other with a much broader appreciation for the impact of illness on workplace costs (employer B). In practice, of course, employers do not fall neatly into these stylized groupings, but are arrayed across a broad spectrum of concerns about the economic consequences of workplace illness.

Employer A: Focus Only on Out-of-Pocket Costs

If asked to define the costs of illness in the workplace, employers frequently respond based on out-of-pocket health costs alone. Figure 1, part A shows the historical growth in private business expenditures for health services and supplies from 1965 to 1991. Included in this aggregate level of spending are such categories as employer contributions to private health insurance premiums, employer contributions to the Medicare hospital insurance trust fund, workers’ compensation and temporary disability insurance, and industrial in-plant health services. By 1991, these expenditures totaled more than $205 billion, or almost 30 percent of all health care spending in the United States.

Private business expenditures for health services and supplies increased at a compound annual average rate of 8.3 percent between 1965 and 1991, after adjusting for inflation. With private business employment growing by only 1.9 percent during this time, real health spending per worker grew by 6.3 percent over this period (see Figure 1, part B).3 As a percent of wages, private business expenditures for health services and supplies grew steadily from 2 percent in 1965 to 9 percent in 1991. Thus, in 1991, for every dollar private businesses paid in wages, they paid an additional nine cents in out-of-pocket costs for health care benefits to employees. Figure 1, part C illustrates this trend in private business health expenditures as a percent of wages and shows that out-of-pocket health care costs have clearly become an increasingly important concern for private business.

Just as policy reform proposals have focused on preventing further escalation of health care spending by containing current costs and adding no new costs, employers may also be tempted to adopt a strict downsizing approach to manage the cost of illness in the work-place. Companies may be inclined to reduce the scale or scope of coverage, thereby hoping to manage better what for many firms have been wildly escalating costs. This approach may be flawed, however, because it fails to recognize the relationship between investments in employees’ health and their workplace productivity:

Many firms provide their employees with more health coverage than is required by law, essentially giving an extra $500 worth of medical insurance instead of an extra $500 in wages. At first this seems mysterious: Why not give employees the cash and let them spend it as they want? A partial answer — and perhaps the entire answer — is that employees prefer nontaxable benefits to taxable wages. But another possible answer is that good health care enhances productivity. If productivity were easily observed and rewarded, there would be no issue here, because employees would have ample incentives on their own to acquire good health care. But in a world of imperfect information, employee benefit packages can be the best way to enforce good behavior.4

The link between health and workplace performance makes it readily apparent that each measure in Figure 1 significantly understates the extent to which health care concerns impose costs on private businesses.

Employer B: Focus on Systemwide Costs

Resource utilization in the health sector often focuses on out-of-pocket expenditures alone. Direct treatment costs can incorporate a variety of categories, including hospitalization, nursing home care, outpatient services such as doctor visits, pharmaceutical therapy, and diagnostic procedures. However, these categories of health care costs do not capture the full effects of illness in the workplace. We must also consider additional important costs not included in the simple accounting, as well as their implications.

Omitted from the standard employer analysis are those categories described in traditional cost-of-illness research as the indirect costs associated with health decisions. This broad approach recognizes that health problems can impose significant additional costs beyond those paid out of pocket. For example, many illnesses impair functioning and well-being, which reduces overall quality of life. (In extreme cases, poor employee health may also lead to premature death, resulting in significant turnover costs to employers from the search for new workers and subsequent training.) Although its measurement has become widespread in clinical investigation and patient care, quality of life is highly subjective and generally difficult to measure clearly.5 In contrast, impaired quality of life may be related to reduced employee productivity, from either incapacitation at work or increased absenteeism. In this context, quality-of-life issues are more straightforward to measure and contrast for different illnesses since they can be measured in dollar terms.

Skeptical employers might question the value of incorporating yet another set of cost categories in the complex discussion of health care. After all, health care spending already is widely viewed as excessive and a source of competitive disadvantage in international trade. However, by ignoring the indirect costs of illness, firms risk making poor resource allocation decisions on health care issues.

Depression in the Workplace: A Case Study

Numerous illnesses, including asthma, arthritis, anxiety disorders, substance abuse, atherosclerosis, back pain, diabetes, and migraine headaches, most likely result in substantial workplace costs to an employer. Each of these and other conditions affect many individuals in the workforce, resulting in substantial productivity losses.

One way to understand the implications of an overly narrow focus on direct costs alone is by considering the most common clinical problem in primary care, which, interestingly, is not a physical illness but a mental illness that cuts across all demographic lines — depression.6 Despite its high prevalence, however, primary care physicians, who tend to focus their diagnosis and treatment on accompanying physical symptoms, frequently do not recognize this mental illness.7 As a result of both its widespread prevalence and the evident difficulty in diagnosis and treatment, depression imposes an enormous burden on society.

Each year, depression affects at least 11 million Americans — about 7.8 million women and about 3.2 million men. We estimate that this widespread illness costs society almost $44 billion annually. However, the vast majority of these costs do not accrue as a result of the direct treatment of this widespread, debilitating illness; almost three-quarters of the costs are indirect and result from a combination of reduced productivity of depressed workers and lost earnings due to depression-related suicides.

Employer A: Direct Cost Focus

Employer A focuses solely on the direct costs of depression. Included among these out-of-pocket expenditures are the costs of inpatient care at short-stay and long-term care facilities, outpatient care and partial care programs, visits to office-based psychiatrists, as well as pharmaceutical prescriptions. We have estimated that, in 1990, direct treatment costs for depression exceeded $12 billion, 28 percent of the annual cost of this illness.8

From this limited vantage point, minimizing the out-of-pocket costs associated with depression is the overriding objective. An employer can take various steps to reduce health expenditures, given such a narrow focus. It can limit spending on an illness category over a particular time frame, for example, specifying an annual ceiling on total mental health benefits among its workers. Alternatively, an employer can impose annual coverage restrictions on the total days a worker can spend in an in-patient or outpatient facility, perhaps a much lower ceiling than for physical illness. In another approach to cost containment, employers can share out-of-pocket health care costs with their employees on a differential basis for certain medical conditions. For example, the direct costs of physical illnesses could be split 85/15 percent between the employer and its employees, with expenditures on mental health disorders perhaps receiving a less generous 50/50 percent division.9

Employer A’s view clearly results in a heavy-handed response to the cost escalation of mental health benefits. When employers “carve out” a particular portion of the larger health care expenditure budget and specify a maximum per capita budget or “capitation” rate for such a category, they may narrowly control the costs of illness. Furthermore, they can replicate this model of disease cost management for different illnesses, treating each one as a stand-alone cost center that must be minimized. Unfortunately, from a systemwide perspective, these narrowly focused strategies may not be optimal, even in terms of direct costs alone. To see why, it is important to understand some salient characteristics of depression.

Two distinguishing features of depression are its treatability and lack of widespread recognition. The National Institute of Mental Health has noted that between 80 percent and 90 percent of individuals suffering from a major depressive disorder can be treated successfully, but that only one in three with the illness ever seeks treatment. Furthermore, because depression is largely unrecognized and often inappropriately treated, an even smaller percentage of those suffering receives proper care.10 A social stigma is likely associated with mental illnesses of all sorts, including depression, which may make people reluctant to seek appropriate care. Increased recognition and awareness of the illness could result in successful treatment of more sufferers of depression.

If we examine clinical practice, we can see the economic implications of broader, rather than more restrictive, treatment for depression. In a common scenario, an individual suffering from the illness visits a primary care doctor frequently to address a succession of physical ailments. The physician treats each somatic complaint in turn on the basis of physical symptoms. Eventually, the physician may diagnose the mental illness underlying these physical complaints; the patient is really suffering from depression. However, because of the improper diagnoses by primary care physicians who are not necessarily attuned to mental health concerns, the process may take several months —or even several years — to complete. Thus it is not surprising that people suffering from depression are currently among the highest users of the health care system.

This scenario highlights a very important fact about a disease like depression: its costs are often very closely related to other health care expenditures. In fact, immediate, effective treatment of depression may reduce other direct health care costs. While it is not clear whether direct medical cost offsets in the case of depression are significant enough to cover the added expenditure for proper treatment, these interactions underscore the importance of a broad view, even when the objective is defined strictly in terms of direct cost management alone.

Employer B: Systemwide Perspective

Employer B considers the possible interdependence of illnesses such as depression with a range of other physical and mental health conditions. As a result, its more efficient health care investments may lead to medical cost offsets of avoidable expenditures, as well as reductions in spending on related conditions that often occur together. In the case of individuals suffering from depression, for example, high costs may result from substance abuse disorders triggered by depression. Comorbidity costs can include those related to other physical or mental health illnesses that are more likely to occur following onset of depression or are exacerbated by its symptoms.

In addition to recognizing the economic interaction among illnesses, perhaps more important, employer B also recognizes that investing in a worker’s health has far-reaching implications in the workplace itself. Such investments can improve employees’ work performance, reduce absenteeism, result in fewer accidents, and free colleagues to concentrate on their own responsibilities rather than covering for an ill coworker.11 In this sense, employer B explicitly recognizes the inherent links between the direct and indirect costs of illness.12

It is important to note that 71 percent of those who suffer from depression in a given year are female, and that 70 percent are between eighteen and forty-five. The gender distribution suggests that, with an increasing number of women in the labor force, the workplace costs of this illness have likely become more significant over time. In addition, the age distribution underscores that this illness strikes people in their prime working years. Furthermore, depression symptoms — including fatigue, inability to concentrate, indecisiveness, and low motivation to accomplish even routine tasks — result in immediate, often significant declines in workplace productivity. Thus many of the illness’s inherent characteristics imply that workplace costs are likely to be quite high due to the effects of depression.

We have estimated that depression accounts for approximately $12.1 billion in costs as a result of employee incapacitation while at work, and about $11.7 billion in costs from employee absenteeism during episodes of the illness. Together, these productivity losses represent approximately 55 percent of depression’s total costs. The final cost component that we have estimated for depression amounts to $7.5 billion per year, or 17 percent of the total, in lost resources due to depression-related suicides.13 Of course, because these sizable cost categories are indirect, they are usually omitted from traditional compilations of health care costs.14

A major illness can impose large costs on society in a number of ways. For example, because the majority of AIDS victims are working-age men, this fatal disease results in large lifetime earnings loss each year. However, the relatively small number of people with this illness tends to limit turnover costs. In the case of cancer or coronary heart disease, there are substantial direct investments for intensive treatment. But because those who suffer from these illnesses are usually closer to retirement age, the majority of its victims are either no longer in the workforce or are nearing the end of their working years. Thus it is unlikely that these illnesses result in relatively large workplace costs. In contrast, because it primarily affects the working age population, depression imposes large costs that are much more gradual and far less dramatic. Its costs are skewed to relatively invisible categories, especially those that result from reduced productivity and increased absenteeism of depressed workers during episodes of the illness.

Since illnesses have different underlying cost distributions, considering out-of-pocket expenditures alone shifts the focus of economic concern toward those health conditions with relatively high direct costs. But in light of the importance of indirect costs, an employer ought to analyze cost effectiveness of treatment on a systemwide basis, including the opportunity costs related to health care treatment decisions.

Although our results show that $24 billion of depression’s total costs result from the diminished workplace performance of those who suffer, employers most likely do not completely understand the costs they currently bear. Only the direct costs of health coverage are regularly incorporated in company financial statements. Because the indirect costs of depression in the workplace are not reported to employers as a line item among all the specific health care expenditures, employers do not generally focus explicitly on these “hidden” costs.

A Simple Model of Indirect Workplace Costs of Illness

Workplace costs of illness clearly extend beyond the direct costs of employee health insurance coverage. To understand the magnitude of health-related workplace costs in a specific firm or for a particular illness, we suggest a simple model of the factors that contribute to resource depletion from health care decisions. The model has broad applicability to a wide range of chronic and acute illnesses, including, for example, migraine headache, arthritis, impairment from viruses or other infectious diseases, various respiratory ailments, and numerous other medical conditions. We can also extend the model to analyze the implications of investments in occupational safety, wellness plans, and even education programs to promote awareness and recognize specific health issues in the workplace.

Diminished workplace performance due to illness has three main components, each answering a different question about the costs of illness:

  • How widespread is the problem among employees? Estimating the percentage of available work time adversely affected by illness involves several steps. First, an employer must ascertain the prevalence of illness in the workplace, given the particular sociodemographic distribution of its workforce. For most illnesses, there are reliable data in medical journal articles, government sources, or trade association publications. Second, to determine how much work time is affected by illness, an employer must estimate the number of workdays per period during which the illness adversely affects performance.15 Prevalence rates are multiplied by expected episode days to complete the calculation. The results can be examined at the disease level, to gain insight into the potential workplace costs associated with a particular illness, at a division or firmwide level, for all illnesses that can adversely affect productivity in a specific company, or for the entire economy, to measure the overall impact of illness in the workplace.
  • How significant is the resulting impairment of workers? When a particular job lends itself to objective measures of productivity declines, the employee impairment rate that results from illness is easy to measure. For example, when an employment agreement requires compensation on a piece-rate or commission basis, it is easy to measure the impairment rate due to illness compared with performance during periods of good health. In many occupations, however, individual productivity may be either difficult to disentangle from a team’s performance or inherently difficult to assess for other reasons. In such instances, an employer can calculate a range of possible productivity cost estimates for alternative impairment rates. When combined with estimates of available work time adversely affected by illness, the employer can develop a range of productivity loss estimates resulting from illness.
  • How costly can the problem be in the workplace? An employer can obtain wage information for specific workers. Alternatively, the employer can compute average wages for each age and gender category described in the prevalence data, based on the salary distribution throughout the firm.

With sound information on each of these issues, an employer can calculate the costs associated with impairment while at work as well as absenteeism from work during illness. If estimates do not seem reliable, alternative methods of analysis may be needed. This may require application of more formal estimation procedures, including the possibility of designing prospective workplace studies to measure the indirect costs of illness and the effects of alternative health interactions on productivity.

Applying the Framework

A range of costs associated with illness in the workplace for the economy as a whole in 1990 are shown in Table 1, which compares annual percentages of available work time affected by illness with alternative impairment rates due to illness to estimate the costs of lost productivity. For example, suppose epidemiological data suggest that illness affects 10 percent of all available work time, with an average impairment rate of 10 percent. In addition, suppose the wage bill for the economy was $2.7 trillion (as in 1990). Under these conditions, the economy bears an annual indirect cost of approximately $27 billion, or 1 percent of the $2.7 trillion wage bill.16

· Depression.

The basic framework underlies the calculation of both components of workplace costs in the depression results we described earlier. Our estimates suggest that approximately 2.25 percent of available work time in the labor force is adversely affected by depression symptoms each year.17 An average reduction in performance during the work year of 20 percent results in costs totaling $12.1 billion due to employee impairment at work. This estimate is obtained by interpolating the data in Table 1.

The other component of workplace costs that we analyzed in detail concerns all the time lost from work due to illness. In the case of depression, we estimated that approximately .44 percent of all available work time each year was lost due to increased absenteeism during episodes of depression.18 By definition, this component of indirect workplace costs results in a 100 percent impairment rate. In the approach shown in Table 1, additional losses in productivity total about $11.7 billion. Although the extent of absenteeism in the case of depression appears to be modest, it nonetheless implies substantial workplace costs.

· Cigarette Smoking Breaks.

The same framework has broad application to various illnesses and health conditions. For example, suppose cigarette smoking breaks result in the loss of fifteen minutes of productive time each workday, which occurs if smokers do not extend their work time to make up for brief, periodic absences. Assuming a standard eight-hour workday, this is equivalent to a 100 percent impairment rate during 3.125 percent of all available work time. If 20 percent of the employees in the workplace smoke, this implies that the value of productivity losses attributable to smoking breaks is approximately $16.875 billion.19 Thus the loss of even a small share of labor output — in this case, 20 percent x 3.125 percent, or .625 percent — results in substantial workplace costs.

· Arthritis.

In another application of the framework, we examine the economics of rheumatoid arthritis (RA) and osteoarthritis (OA) in the workplace. The top part of Table 2 gives several key estimates from the medical literature about the scope of these crippling illnesses in the working population. With only a few types of data concerning arthritis — prevalence, annual workdays lost, and how many people in the labor force suffer from the disease — we can estimate the associated workplace costs.

The bottom part of Table 2 shows our calculations based on data from the literature and a conservative assumption of a $20,000 average annual salary. Using this information, we estimate that the annual absenteeism costs to an employer are $804 per worker with RA and $570 per worker with OA.20

Table 3 shows the costs of diminished productivity due to arthritis under alternative average impairment scenarios. Suppose, for example, that arthritis results in only a 5 percent average impairment rate annually. Although this may appear unduly low, it is an average rate of diminished performance during an entire year. Furthermore, it reflects the contributions of those who remain in the workplace despite symptoms; presumably, workers with the most severe arthritis leave the labor force. Thus self-selection implies that the average impairment of workers with arthritis may not be as significant as might otherwise be expected. Although a 5 percent impairment rate is rather conservative, it nonetheless results in lost productivity estimates per sufferer of almost $1,000 with either form of arthritis. As shown in Table 4, this implies indirect workplace costs of $1,764 per RA sufferer and $1,541 per OA sufferer.

When an employer has difficulty detecting particular types of illness in its workforce, it can calculate the unconditional expected loss from any given worker. For instance, as a direct result of underlying RA costs, there would be an expected $.80 cost of absenteeism and a $.96 loss in workplace productivity, whether or not the particular worker suffered from arthritis. As shown in Table 4, in the case of a worker with OA, an employer can expect to forgo $24.80 in absenteeism costs and an additional $42.26 in lost productivity. The large difference in expected costs per worker by type of illness is due to the much larger prevalence of OA in the labor force compared with RA. In this example, our estimate is based on medical literature reports that more than 6.5 million workers suffer from OA, while only 150,000 workers suffer from RA.

These data are also helpful for estimating costs to the overall economy from increased absenteeism and lost productivity due to an illness such as arthritis. If we assume a $20,000 average annual salary for each worker with one of these medical conditions, RA accounts for approximately $120 million in absenteeism costs and $144 million in impairment costs, while the annual costs of OA in the workplace are estimated at $3.7 billion for absenteeism and $6.3 billion for diminished performance while at work.

Break-even Analysis of Smoking Cessation

Although our approach may not generate precise indirect cost estimates if the impairment rate is unknown, it nonetheless allows us to perform an insightful break-even analysis. One application of the framework is to determine the required improvement in productivity that justifies further investments in employee health. Such investments may include education and awareness programs, screening procedures, or even more proactive measures such as medical interventions. If the critical, break-even productivity improvement seems achievable in practice for a specific illness, such an investment would be cost justified.

Employers who want to recover forgone productive capacity may wonder how much productivity needs to increase, after a health intervention in the workplace, to justify a particular investment. The answer depends on the ratio of the intervention costs — for example, drug therapy, counseling, wellness programs, or even exercise equipment — to the average salary of employees targeted for health improvement.

Suppose, for example, that a particular smoking cessation program costs an annual $500 per employee, and covers nicotine patches, counseling, and other support services. Furthermore, suppose those targeted for behavior modification earn an average annual salary of $20,000. Under these conditions, the required improvement in productivity that justifies this investment is 2.5 percent. On the basis of narrowly defined workplace performance issues alone, a program that succeeds in recapturing 2.5 percent of the productive capacity of the smokers in a particular workforce is economically viable. In such a program, most of the benefits may result from reduced smoking breaks and a decrease in implicit absenteeism, which we estimated earlier at just over 3 percent of work time among smokers. Alternatively, in the case of drug intervention to address depression or arthritis in the workplace, the benefits likely result from a combination of improved performance while at work and reduced absenteeism.

Table 5 and Figure 2 show how to use this simple model of the indirect workplace cost of illness to assess the required improvement in targeted employees’ productivity. The higher treatment costs of a particular intervention require larger projected improvements in productivity to be cost effective. Alternatively, a targeted employee group’s higher average incomes imply that, even with lower resulting productivity improvements, the benefits of interventions may outweigh their costs for those particular employees.

Areas for Future Research

The chairman of Chrysler Corporation, Robert Eaton, once stated that there are “$1,100 of health care costs in each one of our cars.” Eaton’s estimate is probably extremely conservative, since he undoubtedly did not account for the indirect workplace costs of illness. In fact, failing to incorporate indirect costs into a health care discussion can have particularly adverse consequences.

On the basis of direct costs alone, the most cost-efficient health care intervention strategy in general is to do nothing. Not treating a medical condition is always less expensive than treating it, at least in the short term. Only when indirect costs, including quality-of-life considerations, are included in the analysis is it clear that doing nothing to address widespread, persistent health concerns is a particularly inefficient “solution” to a medical problem. An unduly narrow focus on health insurance expenditures alone often ignores the very substantial costs of workers’ absenteeism and diminished performance due to ill health. Our insight from the example of depression in the work-place shows that, for some categories of illness, indirect costs are a multiple of the direct costs employers bear.

Although our discussion has focused on the relationship between health and employees’ performance at work, a broader interpretation of this relationship warrants attention. When workers are unhealthy, their diminished performance can dramatically affect coworkers. For example, if ill health results in more accidents or increased errors, all who explicitly or even implicitly interact with unhealthy employees can become less productive.

One way to incorporate this cost is to permit impairment rates to exceed 100 percent of salary. In various scenarios, the impairment rate due to illness in the work-place is far above 100 percent of a single worker’s usual productive contribution. As a result, such an employee’s absence from work may be preferable to his or her continued attempts to work while ill. So an employer may, at times, be inclined to avoid high costs due to work-place impairment by encouraging the worker to be absent during periods of ill health.

Robert Fogel has stated:

The average efficiency of the human engine in Britain increased by about 53 percent between 1790 and 1980. The combined effort of the increase in dietary energy available for work, and of the increased human efficiency in transforming dietary energy into work output, appears to account for about 50 percent of the British economic growth since 1790.21

With the benefits of hindsight, it now seems self-evident — and is clearly documented in Fogel’s research — that workers’ productivity closely depends on their nutritional intake. Although the link between individual job performance and underlying health status is inescapable, its precise magnitude has not been well established and documented for different workplace environments in occupational studies.

Thus it would be useful to extend our discussion of the relationship between health status and worker performance by investigating several additional areas. For example, employers likely want to know how much they will save in indirect costs for every dollar they invest in direct costs. It is not yet clear how sensitive indirect costs are to changes in the nature and extent of health coverage. For some illnesses, and for some types of occupations, low incremental out-of-pocket investments may quickly result in substantial productivity improvements. In other scenarios, however, it may be necessary to make larger investments to realize substantial improvements in performance.

It would also be useful to assess which illnesses most likely result in substantial workplace costs and to examine their common features. For example, a high prevalence of illness among a working age population generally is conducive to high workplace costs. Additional investigation might include analyzing the occupations most likely to be adversely affected by health disorders, as well as their common characteristics. A given illness may impair worker performance very differently in jobs that require substantial manual dexterity, for instance, compared with those requiring mostly interpersonal interactive capabilities or cognitive skills.

Even if an employer adopts a systemwide perspective for assessing illness costs in the workplace, it is not immediately clear which strategies are best for addressing these far-reaching concerns. For example, it would be useful to investigate specific health interventions likely to be successful in offsetting indirect workplace costs. In this context, the side effects of various drugs, such as sedation, may have dramatic consequences on the differential, systemwide intervention costs.

While health service research is, in many respects, still in early stages, several lines of investigation have been undertaken or are currently underway. For instance, to examine the implications of some workplace cost issues more formally, we have been assessing the economic burden of illness using data from prospective clinical research. One objective is to quantify meaningfully the influence of particular medical factors on various aspects of workplace performance.

Computer-based simulation models hold promise as tools for enhancing managers’ understanding of the salient features of these enormous cost categories. These models describe the interaction of direct costs, workplace costs, and medical parameters in a simulated workplace environment, thereby permitting a broad range of “what if” scenarios to be investigated and analyzed. They also allow rank ordering of interventions so a company can make limited investments that maximize realized return. Some companies may invest in a broad screening program to assess health status, while others may target investments by illness or occupation within the company.

Another area of research involves ongoing retrospective economic investigation to document the actual relationship between individual employee productivity, where measurable, and health status, particularly the effects of different illnesses and alternative interventions on employee performance. These and other avenues will eventually yield insight into the specific relationships among direct and indirect costs across illnesses and occupations.

In the past several years, both businesses and government have committed to “outcomes research” to measure, among other things, returns on investment in health care products and services. For many useful medical interventions, including drugs and clinical procedures, it has been difficult to document improvements in death rates or disease occurrence. But there are claims that these advances in medical treatment or care improve function or quality of life.

Measurement of functional or quality-of-life improvement as a result of medical treatment has relied on patients’ reports of benefits to their physical, psychological, and social well-being. Questionnaires on these issues have been extensively validated in various settings by health services researchers. One such measurement scheme, for example, has successfully differentiated two otherwise similar drugs for hypertension.

But academic researchers and business executives would likely prefer to address the value of such health benefits in familiar measures like dollars, the usual currency of economic comparison. Measurements of the effects of health status on workplace performance, including both absenteeism and lost productivity, will be very valuable in advancing these important lines of investigation. Documenting the value of particular interventions in changing employees’ health status should facilitate such comparisons and ultimately improve employers’ economic decision making.


1. P. Danzon, L. Boothman, and P. Greenberg, “Consolidation and Restructuring in the Health Care Sector: A Precursor to Health Care Reform” (Cambridge, Massachusetts: Analysis Group, Inc., Discussion Paper 94–5, October 1994).

2. A straightforward analogy can be made between the effects of employee health status on accidents in the workplace, including attention to out-of-pocket costs as well as forgone output.

3. More recent surveys have estimated health expenditures per employee in the range of $3,500 to $4,000. Those results can be reconciled with the data presented in Figure 1, part C with attention to the fact that the data in the figure are based on all employees in the labor force. In contrast, the more recent surveys are usually based on the total number of employees covered by employer-sponsored health plans, which is a lower number.

4. S.F. Lansburg, The Armchair Economist: Economics & Everyday Life (New York: Free Press, 1993).

5. T.M. Gill et al., “A Critical Appraisal of the Quality-of-Life Measurements,” Journal of the American Medical Association, 24/31 August 1994, pp. 619–626.

6. See P.E. Greenberg, L.E. Stiglin, S.N. Finkelstein, and E.R. Berndt, “The Economic Burden of Depression in 1990,” Journal of Clinical Psychiatry 54 (1993): 405–418; and

P.E. Greenberg, L.E. Stiglin, S.N. Finkelstein, and E.R. Berndt, “Depression: A Neglected Major Illness,” Journal of Clinical Psychiatry 54 (1993): 419–426.

7. L. Eisenberg, “Treating Depression and Anxiety in Primary Care: Closing the Gap between Knowledge and Practice,” New England Journal of Medicine 326 (1992): 1080–1084.

8. Not all of these direct costs are necessarily paid by employers, however. As we described earlier, private business paid less than 30 percent of all health care expenditures. Households paid approximately 35 percent in private insurance and out-of-pocket expenditures, while federal, state, and local governments paid an additional 35 percent.

9. Some employers might at times pursue other strategies, including ones that may be unlawful. An employer could subtly attempt to minimize the extent of the problem at the most senior levels of the firm by changing its hiring and promotion practices to inhibit such “high risk” individuals. Alternatively, it may attempt to pay a depressed employee less than an asymptomatic counterpart.

10. D.A. Regier et al., “The NIMH Depression Awareness, Recognition, and Treatment Program: Structure, Aims, and Scientific Basis,” American Journal of Psychiatry 145 (1988): 1351–1357.

11. Recent research at the First National Bank of Chicago illustrates the substantial impact of depressive disorders on absenteeism in a corporate setting. It was found that depression resulted in longer disability periods and more frequent disability recurrences than a number of physical disorders with much greater visibility. See:

D.J. Conti and W.N. Burton, “The Economic Impact of Depression in a Workplace,” Journal of Occupational Medicine 36 (1994): 983–988.

12. In recent years, the focus has shifted from attention on cost considerations to greater concern with value for dollars spent. Employer B would shift this focus even further with explicit attention to employee health as opposed to employee health care.

13. From a workplace perspective, the implications of depression-related suicide can be expressed in terms of turnover costs. For example, suppose it generally requires an investment of one year’s salary to search for, identify, and fully train a new employee. The cost associated with a fatality in the workplace due to a specific illness would then be the number of employee deaths per year multiplied by the average annual salary of those who had died. In the case of depression, it is commonly estimated that 60 percent of all suicides are directly attributable to depressive symptoms. If we assume that half these individuals were in the labor force prior to their premature deaths, this implies that in 1990, approximately 7,200 working men and 1,600 working women committed suicide as a result of depression. This results in added turnover costs totaling approximately $200 million per year, or less than one percent of all the workplace costs associated with depression.

14. Depending on the nature of the disability plan, however, a portion of the absenteeism costs may not be borne by the employer. For example, if the direct costs of illness in the workplace already include disability insurance premiums that cover all short-term and long-term disability days for mental illness, this additional cost component should not be added to the costs of illness from an employer’s perspective. It is important to recognize, however, that extended absenteeism from work due to a mental health disorder may fall outside the scope of many disability plans.

15. For a chronic illness, the amount of work time missed can be subtracted from an entire work year to estimate the amount of time that employees perform at diminished productivity. In contrast, in the case of an acute or episodic illness, only a portion of the entire work year likely would be characterized by impaired work performance.

16. This approach is conservative in the sense that the costs of employing a particular worker typically exceed that employee’s salary (i.e., with attention to supplementary benefits).

17. Based on the annual prevalence of depression in the workplace, as well as the average length of time of impaired performance at work due to the disorder.

18. Based on the annual prevalence of depression in the workplace, as well as the average duration of absenteeism from work as a result of the illness.

19. Alternatively, if smokers uniformly extend their workdays to offset the time spent on smoking breaks, the workplace costs associated with this nonproductive activity at work could be zero. These estimates can be obtained by interpolating the data in Table 1.

20. In a specific employer context, this assumption can be replaced with actual salaries in the particular workforce.

21. R.W. Fogel, “Economic Growth, Population Theory, and Physiology: The Bearing of Long-Term Processes on the Making of Economic Policy,” The American Economic Review 84 (1994): 369–389.


The authors are grateful for the research assistance of Tara Nells and Tamar Sisitsky, both of Analysis Group, Inc. Any remaining errors are the sole responsibility of the authors. This research was partly supported by an educational grant from Eli Lilly and Company.

3 Comments On: Economic Consequences of Illness in the Workplace

  • Kirk McIntyre | June 10, 2010

    There are many illnesses which affect the economics of the workplace. Some of these cause time off work, but there is another set of illnesses which do not cause absences but cause functional impairment. Things like arthritis can cause decreased performance and efficiency.

  • mia obran | September 27, 2010

    Interesting point of view on illness at workplace

  • deomadov | June 12, 2011

    Indirect costs of illness in the workplace are SIGNIFICANT! In our office the subject of depression has been historically taboo, not any longer. In fact, recent company surveys have uncovered a very large percentage of our office workforce has admitted, at least on a printed form they deal with depression, day-to-day…

Add a comment