Abstract
Chronic obstructive pulmonary disease (COPD) is a rapidly growing public health problem in the United States and elsewhere. Although direct costs of COPD are well documented, the impact of COPD and its severity on labor force participation is not well known. Using population-based data from the Third National Health and Nutrition Examination Survey (NHANES III), we determined the adjusted relationship between COPD (and its severity) and labor force participation in the U.S. We used data from 12,436 participants involved in NHANES III; 1,073 of these participants (8.6% of the total) reported COPD. These participants were 3.9% (95% confidence interval, 1.3% to 6.4%) less likely to be in the labor force than those without COPD. Increasing severity of COPD was associated with decreased probability of being in the labor force (p for linear trend = 0.001). Mild, moderate, and severe COPD was associated with a 3.4%, 3.9%, and 14.4% reduction in the labor force participation rate relative to those without COPD. These data suggest that COPD has a considerable adverse impact on work force participation. Based on these data, we estimate that, in 1994, COPD was responsible for work loss of approximately $9.9 billion in the U.S.
Keywords: chronic obstructive pulmonary disease; indirect costs; employment; severity; NHANES III
Chronic obstructive pulmonary disease (COPD) affects over 17 million people and is responsible for 2.2 million disability-adjusted life years and one–half million potential years of life lost in the United States (U.S.) (1). Direct costs for COPD are estimated to be over $18 billion annually (2).
Although COPD has been traditionally considered as a disease of the elderly, COPD can also afflict the working age population. According to the 1995 National Ambulatory Medical Care Survey, ∼ 70% of COPD patients were under the age of 65 and they consumed 67% of total COPD office visits and 43% of all hospitalizations (3, 4).
Due to its progressive and debilitating nature, COPD has the potential to interfere with a person's ability to work, leading to productivity losses and lost wages for workers and their employers. A recent study (5) revealed a striking association between COPD and work loss, suggesting that COPD is a problem of considerable magnitude for people in the labor force. However, this study did not control for the potentially confounding effects of age, sex, race, and, most importantly, disease severity. Numerous studies have indicated that disease severity, as measured by spirometry, is a key element influencing a person's capacity to engage in strenuous activities (6-8). Hence, disease severity is likely to be a critical determinant of labor force participation in those with established COPD.
We therefore conducted a population-based study using data from the Third National Health and Nutrition Examination Survey (NHANES III) to determine the relationship between COPD and its severity and labor force participation in the community. We hypothesized that patients with COPD would have a substantially lower labor force participation rate than the rest of the population and that those with the greatest COPD severity would be the most affected, exhibiting the lowest rates of labor force participation among all COPD patients.
We gathered data from people aged 18 to 64 years who participated in NHANES III (9). The NHANES III databases contain comprehensive information on patient demographics as well as socioeconomic, occupational, medical, and dietary histories. Data pertaining to lung function also exist for some individuals. These data were collected using spirometry protocols consistent with the guidelines established by the American Thoracic Society (10). Each participant performed at least five forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) maneuvers on a calibrated, dry-rolling seal spirometer. Only the largest absolute spirometric test values from acceptable maneuvers were recorded. To adjust for height, age, sex, and race we used published prediction equations for FEV1 and FVC, which were derived from studies of well defined, healthy populations (11, 12).
Individuals were classified according to the presence (or absence) of COPD based on the questions: “Has a doctor ever told you that you had chronic bronchitis?” and “Has a doctor ever told you that you had emphysema?” An affirmative answer to either of these questions indicated COPD. As there might be considerable misclassification of COPD on the basis of self-report, we used FEV1 to FVC ratios of less than 0.70 to define significant airflow obstruction (13). Severity of COPD was determined using the Global Initiative for Chronic Obstructive Lung Disease criteria. Mild, moderate, and severe COPD was defined as FEV1 of > 80% of predicted, 30% to 80% of predicted, and < 30% of predicted, respectively, in the presence of significant airflow obstruction (13).
A labor force participation model was used to assess the impact of COPD on the probability of participating in the work force (14). This model is premised on the assumption that individuals spend their time engaged in either leisure activities or at work, and that common forces, including demographic, economic, and health-related factors, influence people's incentive and ability to work, which, in turn, affects the total labor force participation rate in the community. In this model, work can be defined to include or exclude at-home production. We decided a priori to include only paid jobs in the case definition of work for this study. Specifically, we used the question: “During the past two weeks, did you work at any time at a job or business, not counting work around the house?” to create a response variable indicating an individual's working status. A value of one was assigned to an affirmative response and a value of zero was given to all other responses.
To estimate the likelihood of work force participation, a Probit equation was used. The model included the following independent variables: sex (M/F); age (18–24, 25–44, 45–54, and 55–65); education (completed high school or less, or completed more than high school); race (white/black/other); marital status (married, single, separated, divorced, or widowed); family size; area of residence (urban/rural); current smoker (yes/no); and self-reported COPD. We applied sampling weights provided by NHANES to all statistical analyses to obtain 1994 population-based estimates.
The final study sample contained data from 12,436 participants. Participants were, on average, 37.9 (± 13.2 SD) years of age and 5,777 (46.5%) were males. There were 8,000 (64.3%) participants who were white, 3,981 (32.0%) who were black, and 455 (3.7%) who were of other race; 8,845 (71.1%) were nonsmokers. A self-report of COPD was present in 1,073 (8.6%) participants.
The demographic, socioeconomic, and health-related characteristics of the study sample stratified by self-reports of COPD are presented in Table 1. Increasing age, male sex, white race, and increasing smoking exposure were all significant risk factors for self-reports of COPD. In contrast, urban residence was associated with decreased prevalence of COPD.
| Has Self-Reported COPD | ||||||
|---|---|---|---|---|---|---|
| Yes % | No % | p Value* | ||||
| Sample size | 1,073 | 11,363 | ||||
| Gender | ||||||
| Male | 58.47 | 47.43 | 0.001 | |||
| Female | 41.53 | 52.57 | ||||
| Age, yr | ||||||
| 18–24 | 4.62 | 18.40 | 0.001 | |||
| 25–44 | 32.98 | 55.91 | ||||
| 45–54 | 26.47 | 15.12 | ||||
| 55–65 | 35.94 | 10.57 | ||||
| Race | ||||||
| White | 87.16 | 83.05 | 0.001 | |||
| Black | 8.80 | 12.51 | ||||
| Other | 4.04 | 4.44 | ||||
| Marital status | ||||||
| Married | 72.65 | 64.96 | 0.001 | |||
| Single | 9.02 | 22.62 | ||||
| Separated/divorced/widowed | 18.33 | 12.42 | ||||
| Smoking status (packs per year) | 10.64 | 2.84 | 0.001 | |||
| Education | ||||||
| Primary | 4.18 | 3.82 | 0.565 | |||
| Post primary | 95.82 | 96.18 | ||||
| Labor force participation | ||||||
| In the labor force | 69.21 | 77.24 | 0.001 | |||
| Not in the labor force | 30.79 | 22.76 | ||||
| Urban residence | 44.79 | 49.79 | 0.001 | |||
| Family size, mean | 2.92 | 3.24 | 0.001 | |||
| FEV1 (I), mean | 2.67 | 3.46 | 0.001 | |||
| FVC (I), mean | 4.19 | 4.25 | 0.150 | |||
The effects of specific demographic, socioeconomic, and health-related independent variables on the probability of being in the work force are shown in Table 2. Results from the estimating equation indicated that men had an 18.2% higher work force participation than women, whereas blacks had 9.6% lower work force participation than whites.
| Variable | Probability of Employment in the Labor Force (%) | Significance | ||
|---|---|---|---|---|
| Sample size | 12,436 | |||
| COPD (self-report) | Reference | |||
| −3.9 (−1.3 to −6.4) | 0.032 | |||
| Gender | ||||
| Females | Reference | |||
| Males | 18.2 (16.7 to 19.7) | 0.001 | ||
| Age | ||||
| < 25 yr | Reference | |||
| 25–44 | −7.2 (−4.9 to −9.5) | 0.001 | ||
| 45–54 | −2.7 (−0.5 to −4.9) | 0.016 | ||
| ⩾ 55 yr | −23.2 (−21.0 to −25.5) | 0.001 | ||
| Education | ||||
| Less than high school | Reference | |||
| High school | 11.4 (9.5 to 13.3) | 0.001 | ||
| Post-high school | 16.7 (14.8 to 18.7) | 0.001 | ||
| Race | ||||
| Whites | Reference | |||
| Blacks | −9.6 (−7.4 to −11.8) | 0.001 | ||
| Other racial groups | −12.8 (−9.4 to −16.2) | 0.001 | ||
| Marital status | ||||
| Married | Reference | |||
| Single | −2.7 (−0.1 to −5.0) | 0.017 | ||
| Separated/ divorced/widowed | −2.2 (0.0 to −4.4) | 0.065 | ||
| Urban residence | ||||
| No | Reference | |||
| Yes | 2.5 (1.0 to 4.0) | 0.020 | ||
| Smoker | ||||
| No | Reference | |||
| Yes | 0.1 (0.0 to 0.1) | 0.075 | ||
| Census region | ||||
| South | Reference | |||
| North-East | −2.9 (−0.1 to 5.1) | 0.006 | ||
| Mid-West | −0.9 (−2.9 to 1.1) | 0.383 | ||
| West | −4.6 (−2.5 to −6.7) | 0.001 |
Compared with those without COPD, a self-report of COPD was associated with a 3.9% reduction in the adjusted probability of being in the work force (p = 0.032; Table 2). Severity of airflow obstruction, as measured by spirometry, was inversely associated with work force participation, such that those with the most severe airflow obstruction had the lowest probability of being in the work force, whereas those with the least amount of airflow obstruction had the highest probability (Figure 1). Compared with those without any significant airflow obstruction, mild, moderate, and severe COPD was associated with a 3.4%, 3.9%, and 14.4% reduction in the work force participation rate, respectively (p for linear trend = 0.001).
Fig. 1. The relative probability of being in the labor force according to severity of airflow obstruction. Reference group: those without objective COPD. Error bars represent standard error. p For linear trend = 0.001.
[More] [Minimize]Because unpaid housework may be the primary occupation for some individuals, we performed a sensitivity analysis excluding those whose primary job was housework. This analysis produced similar results to the overall findings. A self-report of COPD in this subgroup was associated with a 2.9% (95% confidence interval [CI], 0.1% to 5.1%) reduction in the probability of being in the work force compared with those without COPD (p = 0.009). As with the overall analysis, severity of airflow obstruction was inversely associated with work force participation. Compared with those without any significant airflow obstruction, mild, moderate and severe airflow obstruction was associated with a 2.4%, 3.0%, and 13.4% reduction in the probability of work force participation, respectively, in this subgroup of participants (p for linear trend = 0.001).
Our study produced several interesting and novel findings concerning COPD and work force participation in the U.S. We found that a self-report of COPD was associated with a significant reduction in work force participation, independent of age, sex, and other important factors. Compared with those without COPD, participants with COPD were 3.9% less likely to be in the work force, suggesting an adverse impact of COPD on their ability to work. We also found that the severity of airflow obstruction was a very powerful determinant of work force participation in COPD. Those with severe COPD demonstrated the largest reduction in work force participation, while those with mild COPD showed the least reduction.
These findings are consistent with those of Strassels and coworkers (4), who used the 1987 National Medical Expenditure Survey (NMES) and showed that persons with COPD lost, on average, 3.6 workdays per year because of their disease. However, because NMES lacked critical data for lung function (15), they were unable to determine the effect of disease severity on the relationship between work and COPD. Moreover, they did not evaluate the potential adverse effect of COPD on unemployment, which constitutes by far the biggest potential loss of income for patients and society. By measuring the “employment effect,” and taking into account lung function, we extend the findings of Strassels and colleagues and show that COPD (and in particular moderate to severe COPD) is associated with a marked reduction in work force participation.
A reduction in employment of even a few percentage points can have important economic implications for society. Based on the NHANES III population weights, there were ∼ 9.4 million persons of working age in the U.S. who had COPD in 1994. Using the point estimate for COPD generated in our study (3.9% reduction in the work participation rate), we estimate an excess unemployment of ∼ 366,600 persons in 1994 because of their COPD. If we estimate that persons without COPD worked 52 weeks at 34.5 hours per week (16) and earned the average wage of $11.00 hourly plus fringe benefits ($27,193 annually) (16), then the total lost productivity due to COPD in 1994 was approximately $9.9 billion.
We should note several limitations to our study. First, by excluding at-home production, we used a narrow definition of the labor force participation model. However, when we performed a sub-analysis, excluding those whose primary responsibility was “housework,” results were not significantly different than those of the main analysis, suggesting that our case definition of work did not materially affect the relationship between COPD and work force participation. Second, because the survey was cross-sectional in nature, we cannot be certain that severe COPD is causally related to decreased work participation. There is, however, a strong biologic and physiologic rationale for why this should be so. Many studies have demonstrated that COPD is a progressive and debilitating disease which, in its severe form, leads to marked dyspnea and general malaise that are usually not responsive to therapy (17, 18). These symptoms limit exercise tolerance and impair patients' ability to work (2). Thus, although we cannot entirely rule out the possibility that loss of work could lead to more severe COPD, severe COPD (and its associated symptoms of dyspnea, cough, and impaired exercise tolerance) leading to unemployment seems more plausible. Third, we did not have data on partial work loss for COPD patients still working. Due to their illness, some patients with COPD might have been working part-time or were experiencing frequent missed workdays even though they were still considered “employed.” Thus, our findings are likely to be a conservative estimate of the impact of COPD on work force participation. Fourth, our model did not include several “supply-side” variables that might influence a person's incentive to work, including taxation, welfare subsidies and transfers, family income, and spousal earnings, all of which represent an important limitation to our present work. If any of these unmeasured factors have an appreciable effect on both the “exposure” (i.e., COPD status) and “outcome” (work force participation) variables, then confounding could have been introduced. However, it is unlikely that factors such as taxation loads and family earnings would have been greater in COPD than non-COPD participants making this possibility remote.
COPD is a rapidly growing problem in the community, and its economic impact on society will escalate over the next three decades (19). Although much of the focus has been on direct COPD costs, indirect costs (associated with work loss) are also substantial. Using data from NHANES III, we show that COPD is associated with a 3.9% reduction in the work force participation rate in the community. These data suggest that COPD, particularly moderate to severe COPD, leads to significant unemployment in the U.S. These findings, therefore, indicate an urgent need to improve COPD care to decelerate the progression of this disease so that COPD patients can remain productive in the work force for as long as possible.
The authors wish to thank Casandra Higgs-Carey for her expert secretarial assistance in preparing this manuscript.
| 1. | Gross CP, Anderson GF, Powe NRThe relation between funding by the National Institutes of Health and the burden of disease. N Engl J Med340199918811887 |
| 2. | Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest 2000;117(2 Suppl):5S–9S. |
| 3. | Strassels SA, Smith DH, Sullivan SD, Mahajan PSThe costs of treating COPD in the United States. Chest1192001344352 |
| 4. | Strassels S, Sullivan S, Smith D. Characteristics of the costs of chronic obstructive pulmonary disease (COPD) [abstract]. Eur Respir J 1996; 9(Suppl 23):421S. |
| 5. | Stang P, Lydick E, Silberman C, Kempel A, Keating ET. The prevalence of COPD: using smoking rates to estimate disease frequency in the general population. Chest 2000;117(2 Suppl):354S–359S. |
| 6. | Carter R, Nicotra B, Huber GDiffering effects of airway obstruction on physical work capacity and ventilation in men and women with COPD. Chest106199417301739 |
| 7. | LoRusso TJ, Belman MJ, Elashoff JD, Koerner SKPrediction of maximal exercise capacity in obstructive and restrictive pulmonary disease. Chest104199317481754 |
| 8. | Wegner RE, Jorres RA, Kirsten DK, Magnussen HFactor analysis of exercise capacity, dyspnoea ratings and lung function in patients with severe COPD. Eur Respir J71994725729 |
| 9. | National Center for Health Statistics. Plan And Operation Of The Third National Health And Nutrition Examination Survey, 1988–94. Hyattsville, MD.: US Dept of Health and Human Services; 1994; Publication No. (PHS) 94–1308. |
| 10. | American Thoracic Society. Standardization of spirometry–1987 update. Am Rev Respir Dis 1987;136:1285–1298. |
| 11. | Crapo RO, Morris AH, Gardner RMReference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis1231981659664 |
| 12. | Knudson RJ, Slatin RC, Lebowitz MD, Burrows BThe maximal expiratory flow volume curve. Am Rev Respir Dis1131976587600 |
| 13. | Pauwels RA, Buist AS, Ma P, Jenkins CR, Hurd SSGlobal strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) Workshop summary. Am J Respir Crit Care Med163200112561276 |
| 14. | Ehrenberg RG, Smith RS. Modern labor economics: theory and public practice, 7th ed. Reading, MA: Addison-Wesley; 2000. |
| 15. | Cohen SB, DiGaetano R, Waksberg J. Sample design of the 1987 household survey, National Expenditure Survey Methods 3. Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, 1991; AHCPR Publication No. 91–0037. |
| 16. | Bureau of Labor Statistics. National employment, hours and earnings, series ID: EEU 00500005, average weekly hours of production workers. Available at: http://stats.bls.gov |
| 17. | Celli BR. The importance of spirometry in COPD and asthma: effect on approach to management. Chest 2000:117(2 Suppl):15S–19S. |
| 18. | Senior RM, Anthonisen NR. Chronic obstructive pulmonary disease (COPD). Am J Respir Crit Care Med 1998;157(Suppl):S139–S147. |
| 19. | Murray CJ, Lopez ADGlobal mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet349199714361442 |
Supported in part through an unrestricted research grant from the Alberta Lung Association. Don D. Sin is supported by a New Investigator Award from the Canadian Institute of Health Research
