Assessment of the Economic Burden of COPD in the U.S.: A Review and Synthesis of the Literature

Abstract

The costs of chronic obstructive pulmonary disease (COPD) pose a major economic burden to the United States. Studies evaluating COPD costs have generated widely variable estimates; we summarized and critically compared recent estimates of the annual national and per-patient costs of COPD in the U.S. Thirteen articles reporting comprehensive estimates of the direct costs of COPD (costs related to the provision of medical goods and services) were identified from searches of relevant primary literature published since 1995. Few papers reported indirect costs of COPD (lost work and productivity). The National Heart, Lung, and Blood Institute (NHLBI) provides the single current estimate of the total (direct plus indirect) annual cost of COPD to the U.S., $38.8 billion in 2005 dollars. More than half of this cost ($21.8 billion) was direct, aligning with the $20–26 billion range reported by two other recent analyses of large national datasets. For per-patient direct costs (in $US 2005), studies using recent data yield attributable cost estimates (costs deemed to be related to COPD) in the range of $2,700–$5,900 annually, and excess cost estimates (total costs incurred by COPD patients minus total costs incurred by non-COPD patients) in the range of $6,100–$6,600 annually. Studies of both national and per-patient costs that use data approximately 8–10 years old or older have produced estimates that tend to deviate from these ranges. Cost-of-illness studies using recent data underscore the substantial current cost burden of COPD in the U.S.

Key words: :

• Chronic obstructive pulmonary disease
• Economics
• Costs
• Cost analysis

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) poses a major economic burden to industrialized societies and their health-care systems. In the United States, COPD was the fourth leading cause of death in 2002, and in 2003 approximately 10.7 million adults were reported to have diagnosed COPD ([1], [2]). Its prevalence is expected to increase in coming years as the U.S. population ages ([3]). As a result, the costs of COPD are of increasing concern to public and private health-care payers in the U.S., with payers also noting the emergence of new benefits for seniors (Medicare Part D), as well as increasing enrollment in the Medicare program.

Numerous studies have reported estimates of COPD costs. However, cost-of-illness studies such as these tend to generate widely variable estimates even for a single disease in one country. This variability results from different sources of data (e.g., working-age COPD patients versus elderly Medicare beneficiaries), data from disparate points in time (as demographics and practice undergo change), inclusion of diverse components of costs (e.g., the direct costs of health-care services, the indirect costs of lost productivity), and incongruous methodologies (e.g., costs directly attributable to COPD, or all excess costs of health care for COPD patients versus non-COPD patients). Any cost-of-illness estimate must be interpreted with these factors in mind, especially when multiple estimates are to be compared.

The purpose of this paper is to summarize and critically compare recent cost-of-illness estimates of annual national and per-patient costs of COPD in the U.S. We sought to determine the reliability of current estimates and identify gaps in the recent literature.

METHODS

We retrieved candidate articles for a comprehensive, non-systematic review from a MEDLINE search of the primary literature pertaining to human subjects, written in English, providing abstracts, and published since 1995. We required the following keywords in the title or abstract: COPD or “chronic bronchitis” or emphysema, or obstructive and either lung or pulmonary; and either cost* or economic* (both terms truncated to retrieve variant endings). This strategy identified 442 articles, which we manually searched to find studies reporting direct (related to the provision of medical goods and services) or indirect (lost work and productivity) costs associated with COPD, exclusively in the U.S. The manual search found 48 candidate articles that met our criteria. Four additional articles were identified, three too recently published to have been indexed by MEDLINE at the time of our search ([4], [5], [6]) and one published in a journal not indexed by MEDLINE ([7]), for a total of 52 candidate articles.

From this group of articles, we selected papers reporting comprehensive cost-of-illness estimates of COPD in the U.S. Articles were excluded if they reported on the results from an economic model (to focus on real-world rather than simulated data), pertained to a disease definition larger or smaller than COPD (e.g., all respiratory disease, or acute exacerbations of chronic bronchitis), limited elements of care (e.g., only hospitalizations), or a population not intended to represent the general population of COPD patients (e.g., COPD patients treated with antibiotics). We excluded Tinkelman et al. because of incomplete delineation of methods and results ([5]). After these exclusions, 13 articles remained relevant to our review. Only two of these studies covered indirect costs ([8], [9]); therefore, we focused on summary and comparison of direct costs.

We adjusted all costs to 2005 U.S. dollars using average annual index values for the Medical Care component of the U.S. Consumer Price Index for All Urban Consumers (CPI-U) ([10]). We classified studies by whether they calculated national costs or costs per COPD patient. For per-patient studies, we also determined whether they measured excess costs (total costs incurred by COPD patients minus total costs incurred by non-COPD patients) or attributable costs (costs directly attributable to COPD). For the analysis by Hilleman et al., which stratified per-patient costs by COPD illness severity, we calculated an average cost per patient weighted by stage of disease ([11]).

RESULTS

Of the 13 recent studies included in this analysis, all but one article ([12]) were published in 2000 or later, indicating a recent surge of interest in estimating the cost burden of COPD in the U.S. ([13]). Four of the 13 articles calculated national costs, while 9 calculated costs per patient (Table 1). Most used an attributable cost approach, or reported both excess and attributable costs; only two studies of per-patient costs exclusively used an excess cost approach.

Table 1 Methodologies and analytical approaches in literature reviewed

Reference	Stated perspective	Costing approach	Base year for costs
Annual national costs
Ward et al. 2000 ([17])	Public-payer	Attributable	1994
Wilson et al. 2000 ([14])	Societal	Attributable	1996
Leigh et al. 2002 ([15])	Societal	Attributable	1996
NHLBI 2004 ([8])	Societal	Attributable	2004
Annual per-patient costs
Grasso et al. 1998 ([12])	Medicare	Excess	1992
Hilleman et al. 2000 ([11])	Single university health system	Attributable	1999
Mapel et al. 2000 ([22])	Managed care	Excess	1997
Wilson et al. 2000 ([14])	Societal	Attributable	1996
Strassels et al. 2001 ([21])	Societal	Attributable	1997
Halpern et al. 2003 ([9])	Societal	Attributable	1999
Stuart et al. 2004 ([18])	Medicare	Attributable	2000
Marton et al. 2006 ([6])	Medicaid	Attributable and excess	2001
Miller et al. 2005 ([7])	All-payer	Attributable and excess	2000

Total Costs of Illness

The National Heart, Lung, and Blood Institute (NHLBI), a part of the National Institutes of Health (NIH), estimated in 2004 that COPD costs the U.S. $37.2 billion annually, or $38.8 billion in 2005 dollars. This figure represents the total sum of national costs of COPD, both direct and indirect ([8]). In the last 10 years, the NHLBI has furnished the only such total (direct plus indirect) cost-of-illness estimate for COPD in the U.S. The $38.8 billion cost of COPD ranks higher than the NHLBI's estimate for congestive heart failure ($26.9 billion, 2005) but below the estimated costs of stroke ($55.9 billion, 2005) and hypertension ($57.8 billion, 2005) ([8]).

To generate cost-of-illness estimates for all of these diseases, the NHLBI uses data from large national datasets available through various U.S. government sources. They include: vital statistics and census data for the U.S. population; the annual National Health Interview Survey (NHIS); the National Health and Nutrition Examination Survey (NHANES), including NHANES 1999–2000; the annual National Hospital Discharge Survey (NHDS); and the annual National Ambulatory Medical Care Survey (NAMCS) ([8]).

The cost-of-illness estimate for COPD by the NHLBI indicates that by a small margin (56%), most of the national costs of COPD are direct, specifically resulting from the use of health-care resources and medications ([8]).

National Estimates of Direct Costs

Three recent studies estimate the annual direct cost of COPD to the U.S. on a national scale to be approximately $20–26 billion ($US 2005) ([8], [14], [15]) (Table 2). All three studies used multiple large national datasets as a source of prevalence and cost data. However, caution must be applied in any comparisons between cost-of-illness studies, as different methodologies, disease definitions, components of care, and choice and application of data sources will affect results. It should be noted that while all three of these studies state that they used a societal perspective, none of them truly did, as this would have required estimation of direct nonmedical costs and indirect costs in addition to direct medical costs ([16]). Only the NHLBI study assessed indirect costs, and even here, direct nonmedical costs are excluded, making the purity of the “societal perspective” somewhat questionable ([8]).

Table 2 Annual national direct medical costs of COPD

Reference	Cost components	Base year for costs	Base year $ (millions)	2005 $ (millions)
Ward et al. 2000 ([17])	Inpatient	1994	$2,836	$4,344
	Outpatient		$3,277	$5,020
	Pharmacy		$462	$708
	Total		$6,575	$10,071
Wilson et al. 2000 ([14])	Inpatient	1996	$8,384	$11,874
	Outpatient		$1,044	$1,479
	Pharmacy		$5,064	$7,172
	Total		$14,492	$20,525
Leigh et al. 2002 ([15])
	Total	1996	$18,634	$26,391
NHLBI 2004 (8)	Inpatient	2004	$11,400	$11,882
	Outpatient		$4,500	$4,690
	Pharmacy		$5,000	$5,211
	Total		$20,900	$21,783

In the analysis by Leigh et al., the human capital method was used to estimate direct costs. The principal focus of this study was occupational COPD, although the authors also calculated direct costs for non-occupational COPD. A “top down” estimation approach assumed that all direct costs were proportional to hospital days for COPD, adjusted for the estimated inpatient/outpatient mixture of costs. This method produced an estimate of $18.6 billion for the national direct costs of all COPD (occupational and non-occupational) in 1996, or $26.4 billion in 2005 dollars ([15]).

Wilson et al. used a narrower range of ICD-9-CM codes to define COPD than that used by Leigh et al., and modeled aggregate and individual resource utilization by American Thoracic Society (ATS) stages of COPD severity ([14]). Wilson et al. calculated the national direct cost of COPD to be $14.5 billion in 1996, or $20.5 billion in 2005 dollars, slightly lower than the $26.4 billion estimate by Leigh et al. for the same base year ([14]). Falling between these two figures, and using a wider disease definition similar to that of Leigh et al., the NHLBI estimate for the direct costs of COPD totals $21.8 billion in 2005 dollars ([8]).

One older analysis of the direct national cost of COPD yields a substantially lower estimate than the three more recent studies. Ward et al. evaluated 10 components of COPD care using national survey databases and assigned costs based on rates of reimbursement by Medicare and Medicaid. With an estimated 15 million people in the U.S. affected by COPD (an estimate Ward et al. made after projecting older epidemiological data to the present day), Ward et al. calculated the total direct cost to a public payer in 1994 to be $6.6 billion, or $10.1 billion in 2005 dollars (with a finding that long-term oxygen therapy represents one-third of these costs) ([17]). The difference between this $10.1 billion estimate and the more recent figures of $20–26 billion demonstrates the disparity between costs of illness calculated at widely separated points in time.

Table 2 shows estimates of the annual national direct costs of COPD with their components. Across all studies, inpatient care received at a hospital or nursing home represents a major component of these costs. The NHLBI reports that in 2004, hospital and nursing home care accounted for approximately 55% of direct costs. Another 24% of direct costs were for prescription drugs, and outpatient physician services represented an additional 18% of costs ([8]).

Patient-Level Estimates of Direct Costs

We found nine studies that reported the direct per-patient costs of treating COPD in the U.S. (Table 3). Five of the nine studies measured only attributable costs, which only include costs directly attributable based on diagnosis codes for COPD. Two of the nine studies used only an excess cost approach, which measures the difference between the medical expenditures of patients with and without diagnosed COPD. Two of the nine studies used both attributable and excess cost approaches ([6], [7]).

Table 3 Annual per-patient direct medical costs of COPD

Reference	Cost Components	Base Year for Costs	Base Year $ (millions)	2005 $ (millions)
Grasso et al. 1998 ([12])	Inpatient	1992	$3,408	$5,794
	Outpatient		$1,563	$2,657
	Pharmacy		n/a	n/a
	Total		$4,971	$8,451
Hilleman et al. 2000^1 ([11])	Inpatient	1999	$2,553	$3,293
	Outpatient		$1,420	$1,831
	Pharmacy		$625	$806
	Total		$4,598	$5,930
Mapel et al. 2000 ([22])	Inpatient	1997	$3,067	$4,225
	Outpatient		$1,996	$2,750
	Pharmacy		$806	$1,110
	Total		$5,869	$8,086
Wilson et al. 2000 ([14])	Inpatient	1996	$708	$1,002
	Outpatient		$60	$85
	Pharmacy		$320	$453
	Total		$1,088	$1,540
Strassels et al. 2001 ([21])	Inpatient	1987	$1232	$3061
	Outpatient		$257	$638
	Pharmacy		$176	$437
	Total^2		$1665	$4136
Halpern et al. 2003 ([9])	Inpatient	1999	$2891	$3729
	Outpatient		$1025	$1322
	Pharmacy		$202	$261
	Total		$4118	$5311
Stuart et al. 2004^3 ([18])	Inpatient	2000	$663	$821
	Outpatient		$67	$83
	Pharmacy		$256	$317
	Total		$985	$1221
Marton et al. 2006 (excess cost)^4 ([6])	Inpatient	2001	$2,912	$3,450
	Outpatient		$1,115	$1,321
	Pharmacy		$1,535	$1,818
	Total		$5,561	$6,589
Marton et al. 2006 (attributable cost) ([6])	Total	2001	$2,289	$2,712
Miller et al. 2005 (excess cost) ([7])	Inpatient	2000	$3,298	$4,087
	Outpatient		$1,550	$1,921
	Pharmacy		$796	$986
	Total unadjusted		$5,645	$6,996
	Total adjusted^5		$4,932	$6,112
Miller et al. 2005 (attributable cost) ([7])	Inpatient	2000	$1,365	$1,692
	Outpatient		$767	$951
	Pharmacy		$375	$465
	Total		$2,507	$3,107

¹Hilleman et al. stratified costs by COPD disease severity. To estimate costs across all levels of severity, a weighted average was calculated (see the Methods section).

²Total differs from the sum of the components due to rounding.

³Stuart et al. stratified costs by prescription drug coverage. To estimate costs for both coverage groups, a weighted average was calculated (see the Methods section).

⁴Marton et al. reported costs for Florida and California patients separately. To estimate costs for all patients, a weighted averaged was calculated (see the Methods section).

⁵Miller et al. used multivariate linear least-squares regression to adjust for potentially confounding factors.

Inpatient hospitalization accounts for most of the annual direct per-patient cost of COPD, just as for national costs. In the five most recent studies of direct per-patient costs, inpatient hospitalization represented 52–70% of all direct costs ([6], [7], [9], [11], [18]).

Attributable Costs

Dividing the current NHLBI estimate of $38.8 billion (direct and indirect) as an annual attributable cost of illness for COPD by the 10.7 million diagnosed adults in the U.S. yields an annual per-patient cost of $3,626 in 2005 dollars ([1], [8], [19]). Corresponding direct costs would total $2,037 per year, which is at the low end of the range of recent estimates (Figure 1). Using data collected between 1999 and 2001, four of the most recent studies of the attributable cost of COPD have produced a range of approximately $2,700–$5,900 in 2005 dollars ([6], [7], [9], [11]) (Table 3) (Figure 1).

Figure 1 Estimated annual per-patient direct medical costs of COPD ($US 2005) by first author of study and base year of estimate. Excess costs = difference between medical expenditure of patients with and without COPD. Attributable costs = costs directly attributable based on diagnosis codes for COPD.

Using data from the 2000 Medical Expenditure Panel Survey (MEPS), Miller et al. estimated the direct attributable cost of COPD to be $3,107 per patient after multivariate adjustment for selected demographic factors and for smoking ($US 2005). The authors hypothesized that the estimate would have been higher if it had included long-term oxygen therapy and the cost of COPD in institutionalized adults, neither of which is collected in the MEPS ([7]). However, similar estimates were calculated by Marton et al. using administrative claims data from Medicaid programs, which do include institutionalized adults (but excluding those residing in skilled nursing facilities) and long-term oxygen therapy. Costs attributable to COPD were $2,559 per patient in the California Medicaid program and $2,769 in Florida ($US 2005) ([6]).

Two studies using data contemporary with those used by Miller et al. and Marton et al. did produce higher estimates of direct costs attributable to COPD. Halpern et al. used U.S. responses to a large, international survey of middle-aged and elderly adults; direct costs attributable to COPD totaled $5,311 ($US 2005) ([9]). A study by Hilleman et al. reviewed the treatment at an academic medical center of 413 patients aged 35 to 80 years, and calculated direct costs of $5,930 ($US 2005) ([11]). Both of these estimates are more than $2,000 higher than the $3,107 estimate of direct costs attributable to COPD reported by Miller et al. ($US 2005) ([7]). Nearly all of this difference is due to higher inpatient costs in the Halpern et al. and Hilleman et al. analyses, and high costs of home oxygen therapy in the Hilleman et al. study. As in the Miller et al. analysis, neither Halpern et al. nor Hilleman et al. included costs associated with nursing home or other long-term care ([7], [9], [11]).

Hilleman et al. also calculated the costs of COPD by ATS stages of COPD severity ([20]). This stratification shows that annual direct costs more than double as patients progress between stages, with costs attributable to COPD totaling $2,168 for patients in stage I, $6,496 in stage II, and $13,944 in stage III ($US 2005). As patients progress from stage I to III, they spend increasing proportions on inpatient care (40% of costs in stage I to 63% in stage III) and decreasing proportions on pharmacy (30% of costs in stage I to 7% in stage III) (11).

While these four studies of 1999–2001 data all estimate that annual attributable costs were between $2,700 and $5,900, one other study using data from this recent time period produced a much lower estimate (Figure 1). Stuart et al. assessed costs of COPD treatment using data from 462 patients responding to the 1999 and 2000 U.S. Medicare Current Beneficiary Survey. The attributable cost of treating these patients totaled only $1,221 ($US 2005). One possible reason for this low estimate is that it covered hospitalizations, pharmacy, and outpatient physician visits, but no other components of care, such as nursing home stays, oxygen therapy, or laboratory tests. In addition, the study excluded ICD-9-CM code 496.x (chronic airway obstruction, not elsewhere classified), as used in most other studies. Perhaps most importantly, the authors note that only three percent of expenditures on outpatient physician visits had a COPD diagnosis, suggesting probable under-coding ([18]). Therefore, we have excluded this cost estimate from our range based on more recent data sources of $2,700–$5,900 annual attributable costs per patient.

Two studies using data prior to 1999 yielded somewhat lower cost estimates (Figure 1). Using 1996 data from large national databases, Wilson et al. calculated the direct costs of COPD to be $1,540 in 2005 dollars ([14]). Strassels et al. published in 2001 an analysis of data from the 1987 National Medical Expenditure Survey (NMES), the precursor to the MEPS, and calculated the direct attributable cost of COPD to be $4,136 ($US 2005). While this estimate appears comparable to the range of estimates generated with recent data, its reliability is undermined by changes in COPD practice patterns and drug therapies across the nearly 20 years since the underlying data were collected ([21]).

Excess Costs

Both the MEPS study and the Medicaid study by Marton et al. calculated the direct per-patient cost of COPD using both an attributable and an excess cost approach. The excess cost estimate by Miller et al., $6,112 per patient after adjustment for demographic differences and other potential confounders, closely approximates corresponding figures from the Marton et al. study, in which excess costs total $6,589 ($2005). Comorbidities linked to smoking were estimated to account for about a third of excess costs in this Medicaid population ([6]).

As observed in studies of attributable costs, excess cost studies using older data have tended to yield estimates that vary widely from the range derived from recent data, $6,100–$6,600 (Table 3) (Figure 1). Using administrative claims and some medical records of 1,522 members of a New Mexico health system in the mid-1990s, Mapel et al. estimated the direct excess cost of COPD to be $8,086 ($US 2005) ([22]). Using 1992 data from the Health Care Financing Administration (now the Centers for Medicare and Medicaid Services), Grasso et al. estimated the direct excess cost of COPD among elderly Medicare beneficiaries to be $8,541 ($US 2005). It should be noted that the estimate by Grasso et al. excluded ICD-9-CM code 496.x, and the $8,541 estimate likely would have been higher had this study used the broader definition of COPD subscribed to by most other studies included in this review ([12]).

Indirect Costs

We identified two recent cost-of-illness studies reporting indirect costs of COPD ([8], [9]). No other recent studies have calculated the indirect costs of COPD, which are more difficult to quantify than direct costs ([23]).

The total indirect national costs of COPD have been calculated by the NHLBI to be $16.3 billion in 2004, or $17.0 billion in 2005 dollars. In estimating this cost, the NHLBI considered costs resulting from morbidity and early mortality of patients with COPD, and thus the loss of potential work productivity by patients with the disease ([8]).

One estimate of the indirect per-patient costs of COPD has been published in the last 10 years, by Halpern et al., using U.S. responses to the “Confronting COPD in North America and Europe” survey. Working-age patients responding to this survey reported losing time from work equivalent to 18.7 days per patient annually, amounting to $1,970 per patient each year ($US 2005). Seven percent of survey respondents reporting that they were caregivers for a person with COPD also described losing time from work—an average of 1.7 days per year. While time lost from work by caregivers of COPD patients does present an economic burden on society, Halpern et al. did not include this component in their estimate of indirect costs per COPD patient ([9]).

DISCUSSION

The NHLBI provides the single current estimate of the total (direct plus indirect) annual cost of COPD to the U.S., $37.2 billion in 2004, or $38.8 billion in 2005 dollars. Because it is the single current estimate of the national cost of COPD, it is widely cited, and must be updated on a regular basis to remain current with changes in demographics, treatment patterns, and the structure of the healthcare system.

Most of the total annual cost estimated by the NHLBI was direct. While cost-of-illness comparisons are limited by major differences in methods and data sources, it is notable that the NHLBI's $21.8 billion estimate of direct costs aligns with the $20–25 billion range reported from two other recent analyses of large national datasets ([14], [15]). For per-patient direct costs attributable to COPD, studies using recent cost data yield estimates in the range of $2,700–$5,900 annually, and $6,100–$6,600 for excess costs ($US 2005).

A strength of our study is that we report results from two alternative costing approaches, each with advantages and disadvantages. Because the attributable cost approach only counts costs specific to a diagnosis of COPD, patients with comorbidities may have had their diagnosis linked to those conditions, rather than correctly coded to COPD. As a consequence, true COPD-attributable costs may be underestimated. This problem is particularly likely when data are derived from claims databases that allow for only one diagnosis code. In contrast, the excess cost method measures the overall difference in expenditures, regardless of specific diagnoses, between COPD patients and a comparison group without COPD. However, in this approach, a portion of excess costs may reflect other, related respiratory conditions. For example, Marton and colleagues found that comorbidities were estimated to account for 35–38% of the excess cost of COPD, and that asthma as a concomitant diagnosis with COPD was found to contribute an additional 8–9% to excess costs ([6]).

Studies of direct costs using data approximately 8–10 years old or older have produced estimates that tend to deviate from the ranges stated above. Ward et al., using national survey data from 1994, calculated the direct national cost of COPD to be $10.1 billion in 2005 dollars, much lower than the $20–26 billion range generated by analyses of data from the last 10 years ([17]). Grasso et al., Wilson et al., and Mapel et al., using data approximately 10–15 years old, generated estimates varying more or less by $1,100–$2,500 from the recent range of $2,700–5,900 for attributable and $6,100–$6,600 for excess costs ($US 2005) ([12], [14], [22]). An exception is Strassels et al., who used data nearly 20 years old but produced an estimate that appears similar to recent figures ([21]). The disparities between most estimates derived from older versus newer data illustrate both the large differences in study design inherent in comparing cost-of-illness studies and the short shelf life of health economic data. In the time since data for older studies were collected, dramatic changes have occurred in the organization, financing, and delivery of health-care in the U.S. There have also been important changes in the sociodemographics of the U.S. and in COPD treatment patterns, which are influenced by factors such as guideline revisions, newly available medications, efforts to cut costs, and the results of prominent clinical trials. As a result, earlier estimates of these costs may not accurately reflect current costs of this illness ([24]).

In contrast to the work that has been done to assess the direct costs of COPD, few attempts have been made to assess the indirect costs of the disease. The NHLBI's $17 billion estimate of indirect costs to the U.S. remains the only recent tally of national indirect costs of COPD to date; it includes the estimated costs of morbidity and early mortality resulting from the disease. One estimate of per-patient indirect costs has been published in the last 10 years, by Halpern et al., who calculated annual time lost from work due to COPD at $1,970 per patient ($US 2005) ([9]). Despite the dearth of indirect cost estimates in the literature, it is evident that COPD presents a substantial economic burden in lost labor time and home production. A model by Sin et al. used NHANES III data to calculate the effects of COPD on labor force participation and estimated that in 1994, COPD resulted in work loss totaling $9.9 billion in the U.S. ($15.2 billion in $US 2005) ([25]). Using NHIS data, Ward et al. found that approximately 20% of persons with chronic respiratory conditions are unable to work at all (although only 2–3% of these patients attributed their inability to work to their respiratory conditions) ([26]).

Given the large direct and indirect cost burden of COPD in the U.S., efforts have been (and continue to be) made to identify strategies to avert and control these costs. Because patterns of resource utilization ultimately determine costs, more work is needed to understand differences in resource utilization among health systems and populations. In our current analysis, we have limited our comparisons to costs, but a similar comparison of utilization data is needed to inform cost-control strategies.

Hospitalization costs are a particular target, as both national and per-patient cost analyses consistently show that inpatient care accounts for most of the direct cost of COPD ([7], [9], [11], [18]). Much of this cost stems from hospitalization for acute exacerbations of chronic bronchitis (AECBs) ([13], [27], [28], [29]). While AECBs frequently are managed on an outpatient basis, it is possible that even greater shifting of management from inpatient to outpatient settings could offer significant cost savings, as could better prevention of acute episodes through improved control of disease ([27], [30]). This could be accomplished through better adherence to treatment guidelines for early diagnosis and treatment, smoking cessation, recommended vaccinations, and more effective symptom management ([23], [31]). In particular, the use of long-acting bronchodilators and inhaled corticosteroids has been shown to reduce exacerbation rates by approximately 20–25% ([32]). Pulmonary rehabilitation and better patient training regarding the use of long-term oxygen therapy may also lead to decreased use of health-care resources ([23]). Guidelines incorporating these goals for effective treatment of patients within a disease management program are offered by the Global Initiative for Chronic Obstructive Lung Disease (GOLD), a partnership between the NHLBI and the World Health Organization (WHO) ([33]). More work is needed to determine whether treatment that adheres to these guidelines will reduce costs.

Finally, some recent work suggests that the most benefit may come from interventions targeting younger adults; Joyce et al., modeling lifetime costs of COPD, calculate that by the time patients reach age 65, remaining lifetime costs to Medicare are only $7,942 more for a patient with COPD than for a patient without COPD, because the latter will likely live longer ($US 2005) ([34]). Unfortunately, the U.S. failed by large margins to meet smoking prevention goals outlined in the government's Healthy People 2000 initiative ([35]). Better strategies are needed to reduce the number of young people entering the pool of long-term smokers at greatest risk for developing this expensive and incurable disease.

In conclusion, COPD poses a major burden to the health-care system of the U.S., with total annual national costs estimated by the NHLBI at $38.8 billion in 2005 dollars. Most of the total costs estimated by the NHLBI are direct, and most direct costs derive from inpatient care; there is a need for more studies of the indirect costs of COPD. Cost-of-illness studies using recent data underscore the substantial current economic burden to the U.S.

Funding for this study was provided by Pfizer, Inc., New York, New York.