PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Arthritis Rheum. Author manuscript; available in PMC 2010 August 23.
Published in final edited form as:
PMCID: PMC2925686
NIHMSID: NIHMS197560

Medical Care Expenditures and Earnings Losses Among Persons With Arthritis and Other Rheumatic Conditions in 2003, and Comparisons With 1997

Abstract

Objective

To obtain estimates of medical care expenditures and earnings losses associated with arthritis and other rheumatic conditions and the increment in such costs attributable to arthritis and other rheumatic conditions in the US in 2003, and to compare these estimates with those from 1997.

Methods

Estimates for 2003 were derived from the Medical Expenditures Panel Survey (MEPS), a national probability sample of households. We tabulated medical care expenditures of adult MEPS respondents, stratified by arthritis and comorbidity status, and used regression techniques to estimate the increment of medical care expenditures attributable to arthritis and other rheumatic conditions. We also estimated the earnings losses sustained by working-age adults with arthritis and other rheumatic conditions. Estimates for 2003 were compared with those from 1997, inflated to 2003 terms.

Results

In 2003, there were 46.1 million adults with arthritis and other rheumatic conditions (versus 36.8 million in 1997). Adults with arthritis and other rheumatic conditions incurred mean medical care expenditures of $6,978 in 2003 (versus $6,346 in 1997), of which $1,635 was for prescriptions ($899 in 1997). Expenditures for adults with arthritis and other rheumatic conditions totaled $321.8 billion in 2003 ($233.5 billion in 1997). In 2003, the mean increment in medical care expenditures attributable to arthritis and other rheumatic conditions was $1,752 ($1,762 in 1997), for a total of $80.8 billion ($64.8 billion in 1997). Persons with arthritis and other rheumatic conditions ages 18–64 years earned $3,613 less than other persons (versus $4,551 in 1997), for a total of $108.0 billion (versus $99.0 billion). Of this amount, $1,590 was attributable to arthritis and other rheumatic conditions (versus $1,946 in 1997), for a total of $47.0 billion ($43.3 billion in 1997).

Conclusion

Our findings indicate that the increase in medical care expenditures and earnings losses between 1997 and 2003 is due more to an increase in the number of persons with arthritis and other rheumatic conditions than to costs per case.

Concern about the individual and population economic burden of arthritis and other rheumatic conditions is growing in developed and developing countries alike, crystallized by the activities of the Bone and Joint Decade (14). In developed countries, this burden is in part the result of the aging of the population (4) and of the increased medical care costs associated with the expansion of joint replacement surgery, the rise (and then fall) of the coxibs, and, for inflammatory conditions, the development and increasing use of biologic agents (5). This concern is reflected in systematic reviews of evidence on the economic impact of select rheumatic conditions (summarizing studies on clinical samples) (613) and in a series of studies on the total economic burden associated with arthritis, many using the arthritis and other rheumatic conditions rubric and/or population-based data (1424).

The present report provides an update on medical care expenditures and earnings losses among persons with arthritis and other rheumatic conditions. The specific goals of the study were 1) to provide estimates of all medical care expenditures on behalf of all adults (ages 18 years and older) with any form of arthritis and other rheumatic conditions in the US in 2003 as well as the increment in expenditures specifically attributable to arthritis and other rheumatic conditions, 2) to provide estimates of lost earnings among working-age adults with arthritis and other rheumatic conditions (ages 18–64 years) and the increment in lost earnings also attributable to arthritis and other rheumatic conditions among persons in this group, and 3) to compare 1997 and 2003 estimates derived by the same methodology.

METHODS

Overview

The present study used the same data source and methods of analysis as 2 prior studies, on direct costs in 1996 (22) and direct and indirect costs in 1997 (24). The data source and methods of analysis are described briefly below. Specifically, we used the same methods to provide estimates for 2003 and then inflated estimates from the 1997 analysis to 2003 dollars, to compare expenditures and earnings losses in the 2 years.

Data source

The data source used was the Medical Expenditures Panel Study (MEPS). The MEPS is designed to provide data on health care use, medical care expenditures, sources of payment, and insurance coverage for a representative sample of the civilian noninstitutionalized US population of all ages. It also tracks the employment status and earnings of members of that sample.

The full MEPS data include survey responses from a sample of households (MEPS-H), information from providers for selected participants about their medical conditions, and information from their health plans about these plans (25). The MEPS-H sample is derived from the prior year’s National Health Interview Survey respondents, who are, in turn, derived from a clustered, random sample of the civilian noninstitutionalized population, with oversampling of African Americans and Hispanics (26). In the present report, we use data from the 2003 MEPS-H, which included 11,929 households with 34,215 persons, of whom 23,352 were 18 years of age or older. Of the latter, 4,801 met the present study’s definition of arthritis and other rheumatic conditions (see below). In the 1997 MEPS-H, there were 22,435 persons 18 years or older, of whom 4,449 met the definition of arthritis and other rheumatic conditions.

MEPS-H data are collected through 6 rounds of interviews over a 2.5-year period. The data used in the analysis for this report, for both 1997 and 2003, were obtained from the first 3 MEPS-H interviews, covering expenditures, employment, and earnings over an entire year.

Analyses

Data partitions

We first partitioned the 2003 MEPS-H data into 1 of the following 5 chronic condition groups on the basis of International Classification of Diseases, Ninth Revision (ICD-9) codes: 1) persons with only arthritis and other rheumatic conditions (ICD-9 codes 274, 354, 390, 391, 443, 446, 710–716, 719–721, and 725–729), 2) persons who had both arthritis and other rheumatic conditions and chronic conditions other than arthritis and other rheumatic chronic conditions (non–arthritis and other rheumatic conditions), 3) persons with 1 non–arthritis and other rheumatic condition, 4) persons with ≥2 non–arthritis and other rheumatic chronic conditions, and 5) persons with no chronic conditions. The definition of arthritis and other rheumatic conditions was based on a standard developed by the National Arthritis Data Workgroup but abridged to 3-digit ICD-9, Clinical Modification codes to be compatible with the MEPS-H public release files (24).

General considerations

Because the MEPS-H is based on a 2-stage cluster sample rather than a simple random sample of the noninstitutionalized population, it is necessary to weight the data to make inferences regarding the US population. In the MEPS-H, the sampling weights also take into account nonresponse in the households targeted for inclusion and attrition among respondents after completion of the first interview (27). We used SAS (version 9.1) Survey procedures (Surveymeans, Surveyreg, and Surveylogistic), which employ the Taylor expansion method to account for the cluster sampling design, in all analyses requiring the calculation of the standard error of estimates (28).

In the analyses, we estimated mean medical care expenditures by category (e.g., inpatient, outpatient, prescription medication) and earnings losses for persons in each data partition as well as total expenditures and earnings losses aggregated across all such persons. We also estimated the increment in mean expenditures and earnings losses attributable to arthritis and other rheumatic conditions and then provided a total figure for these increments aggregated across all persons with arthritis and other rheumatic conditions.

Description of expenditures

We began by enumerating medical care expenditures on behalf of persons of all ages by condition group and category of medical care expenditures, and also showed the distribution of medical care expenditures. In the foregoing analysis, we tabulated all expenditures among persons in the condition groups, without regard to whether the condition in question accounted for the expenditures. In the table for these analyses, we indicated those estimates having low statistical reliability, that is a relative standard error of >30%.

Increment in health care expenditures

To assess the incremental contribution of arthritis and other rheumatic conditions to medical care expenditures, we estimated a series of regressions separately for adults with and those without arthritis and other rheumatic conditions. We then simulated the expected level of expenditures for the arthritis and other rheumatic conditions group as if they did not have the arthritis and other rheumatic conditions, by applying the parameter estimates derived from the group without the condition to the data from the group with the condition. The mean increment was then calculated as the (weighted) mean difference between the expected expenditures and the “actual” expenditures for individuals in the arthritis and other rheumatic conditions group (29). In these estimations, “actual” expenditures were in fact adjusted means estimated using the regression models and parameter estimates for the group (adjustment variables are listed below). The total increment was calculated by multiplying the mean increment by the estimated number of adults with arthritis. This results in a slightly more accurate estimate of total US expenditures attributable to arthritis and other rheumatic conditions than the method used previously (24), in which only the weighted number of arthritis cases available for regression was used as the multiplier.

To make these calculations with respect to ambulatory care, inpatient, prescription drug, and residual expenditures (comprising home health care, visual aids, dental visits, and medical devices), we applied the 2-stage method outlined by Duan et al (30) to account for the skewed distribution of medical care expenditures, in which many persons have low medical care expenditures or none, while a small proportion have high expenditures. In this method, logistic regression is first used to estimate the probability that an individual has any expenditures, and then ordinary least squares regression on the logarithm of the expenditure is used to estimate the level of expenditures strictly among those with positive expenditures.

We estimated the incremental contribution of arthritis and other rheumatic conditions to total expenditures using a 4-stage model, also described by Duan et al (30). The first stage uses a logistic regression to predict the probability of any medical expenditures. The second stage uses a separate logistic regression to predict the probability of any hospital expenditures among individuals with medical expenditures. The third stage uses an ordinary least squares regression to predict the logarithm of the level of costs among persons without hospitalizations, and the fourth stage uses an ordinary least squares regression to predict the logarithm of costs among persons with hospitalizations. For both the 2-stage and the 4-stage models, the log transformation is used in the ordinary least squares regressions in order to account for the skewed distribution of expenditures. The resulting estimate, when transformed back to the original units by exponentiation, is biased downward (that is, the expected value of the estimate is lower than the expected value of the population mean). Duan (31) has shown that multiplying the estimate by an adjustment called the smearing coefficient can be used to adjust for this bias. The smearing coefficient is the (weighted) mean of the pooled exponentiated residuals from the regressions on the log-transformed expenditures for both groups. This coefficient is then multiplied by the exponentiated predicted log costs to obtain more accurate estimates of the predicted expenditures for the arthritis and other rheumatic conditions group.

All of the regression models included independent variables for the presence or absence of 9 high-cost chronic conditions: hypertension, other forms of heart disease, pulmonary disease, stroke, other neurologic conditions, diabetes, cancer, mental illness, and non–arthritis and other rheumatic musculoskeletal conditions.

In addition to the indicator variables for the 9 chronic conditions, we included variables for female versus male sex, age categories (45–64 years, ≥65 years, with 18–44 as the referent), white versus nonwhite race, Hispanic versus non-Hispanic ethnicity, level of formal education (high school graduate, some college, college graduate, graduate school, with less than high school as the referent), marital status (single, widowed, separated, divorced, with currently married as the referent), and health insurance status (presence of only public insurance, presence of no insurance, with presence of any private insurance as the referent).

Description of employment status and earnings losses

Earning losses were estimated among persons ages 18–64 years in the MEPS with a reported history of ever having been employed. Estimation of earnings losses was limited to this age group both for the substantive reason that labor force participation rates decline precipitously at age 65 as individuals reach the earliest age of eligibility for full Social Security retirement benefits, and for the methodologic reason that estimates among persons ages 65 and older, based on a relatively small sample size, would not be reliable.

To estimate earnings losses, we first tabulated the actual employment status of persons with and those without arthritis and other rheumatic conditions who were between the ages of 18 and 64 years and, among those who were employed, the average weekly hours of work. The absolute level of lost wages was then calculated as the sum of 2 estimates. The first estimate was the cost of lost wages among those individuals with arthritis and other rheumatic conditions who were not working at all. This estimate is equal to the difference in the employment rates of persons with and those without arthritis and other rheumatic conditions times the number of persons with arthritis and other rheumatic conditions times the mean yearly earnings among all employed persons with arthritis and other rheumatic conditions. The second estimate was the cost of lost wages among individuals with arthritis and other rheumatic conditions who continued to work. This estimate is the product of the mean reduction in yearly wages per employed individual with arthritis and other rheumatic conditions times the number of employed individuals with arthritis and other rheumatic conditions. The mean reduction in yearly wages is calculated as the difference in the mean hours of work per week of employed persons with and those without arthritis and other rheumatic conditions divided by the mean hours of work per week of all employed persons (yielding the percentage reduction in hours) times the mean yearly earnings among all employed persons with arthritis and other rheumatic conditions.

Increment in earnings losses

We then estimated the increment of earnings losses attributable to arthritis and other rheumatic conditions, in a manner analogous to the estimate of the expenditures increment. Specifically, we began by estimating the probability of employment in separate logistic regressions among persons with and those without arthritis and other rheumatic conditions, and then estimated, among the members of the 2 groups with earnings, the level of annual earnings, using separate ordinary least squares regressions. The regressions included the following independent variables: sex, categories of age (35–44, 45–54, and 55–64, with 18–34 as a referent), white versus nonwhite race, Hispanic ethnicity, education level, marital status, and indicator variables for the same comorbid chronic conditions as described above for the direct cost analyses.

After the regressions were completed, we applied the parameter estimates derived from regressions on persons without arthritis and other rheumatic conditions to the data on those with such conditions, to simulate the expected level of earnings among such persons if they did not have arthritis and other rheumatic conditions. The mean increment in earnings losses was then calculated as the difference between the arthritis and other rheumatic conditions and simulated non–arthritis and other rheumatic conditions values.

Translating 1997 expenditures and earnings losses to 2003 terms

We used the Medical Care Component of the Consumer Price Index (32) and the wage/salary component of the Employment Cost Index (33) to translate 1997 expenditures and earnings, respectively, into 2003 dollars. Because the components of medical care costs may not have increased at the same rates (with a greater rate of inflation in costs for prescription drugs than for other services or products within the overall Medical Care Component), this may have understated the overall increase in medical care expenditures.

RESULTS

Between 1997 and 2003, the prevalence of arthritis and other rheumatic conditions as defined in the MEPS-H increased from 36.799 million adults (18.7% of the US civilian noninstitutionalized population) to 46.114 million (21.5%) (Table 1). The increase is due, approximately equally, to an increase in population (predominantly in the “baby boomer” cohort) and an increase in the age-specific rates of arthritis and other rheumatic conditions. The increase in rates of arthritis and other rheumatic conditions from 1997 to 2003 occurred for most ages, but was especially pronounced for those ages ≥50 years (40 per thousand higher for age ≥50 compared with younger ages).

Table 1
Number and percent of the US noninstitutionalized population ages ≥18 years, by AORC status, 1997 and 2003*

In 2003, 2.5% of the population reported having arthritis and other rheumatic conditions alone and 19.0% reported having both arthritis and other rheumatic conditions and non–arthritis and other rheumatic conditions; the analogous percentages in 1997 were 2.6% and 16.1%, respectively. Thus, all of the overall increase in the prevalence of arthritis and other rheumatic conditions occurred among persons who also reported having non–arthritis and other rheumatic conditions. The proportion of the population with non–arthritis and other rheumatic conditions but without arthritis and other rheumatic conditions remained the same between 1997 and 2003. Because of the increase among those with arthritis and other rheumatic conditions, the percentage with no chronic conditions decreased, from 31.5% to 28.8%.

Table 2 provides detailed information on medical care expenditures of the noninstitutionalized population of the US in 2003. Mean expenditures are shown for persons with arthritis and other rheumatic conditions only, for those with arthritis and other rheumatic conditions as well as non–arthritis and other rheumatic conditions, and, among those with only non–arthritis and other rheumatic conditions, for those with 1 versus ≥2 such conditions. Mean expenditure data for persons with arthritis and other rheumatic conditions and those with non–arthritis and other rheumatic conditions in 1997 and 2003 are presented in Figure 1.

Figure 1
Comparison of mean medical care expenditures in the arthritis and other rheumatic conditions (AORC) and non-AORC (i.e., chronic conditions other than AORCs) subgroups by expenditure category, 2003 versus 1997 (1997 expenditures inflated to 2003 dollars). ...
Table 2
Mean individual and total medical care expenditures for the US noninstitutionalized population ages ≥18 years, by detailed AORC status and expenditure category, 2003*

Between 1997 and 2003, mean expenditures among persons with arthritis and other rheumatic conditions increased across all categories by almost 10% in relative terms, from $6,346 to $6,978. In addition, there was a change in the composition of expenditures. Out-patient costs for arthritis and other rheumatic conditions increased slightly, from 29% of the total in 1997 to 32% in 2003. Inpatient expenditures for arthritis and other rheumatic conditions decreased in both absolute terms (from $2,504 to $2,217) and relative terms (from 39% to 32% of the total). In contrast, expenditures for prescription medications for arthritis and other rheumatic conditions almost doubled in absolute terms (from $899 to $1,635) and increased by >50% in relative terms (from 14% to 23% of the total expenditures in 1997 and 2003, respectively). The overall increase in expenditures for filled prescription medications between 1997 and 2003 was the result of an increase in the mean number of prescriptions, from 18.7 to 25.2 per person, and in the mean cost per prescription expressed in constant terms, from $48 to $65.

In 1997, the 36.799 million adults with arthritis and other rheumatic conditions incurred $233.5 billion in total medical care expenditures (37% of all expenditures for that year), or ~2.3% of the Gross Domestic Product (GDP) for the US in that year. By 2003, the 46.114 million persons with arthritis and other rheumatic conditions incurred $321.8 billion in expenditures (40% of all expenditures) (Table 2), or ~3.0% of the GDP. Total expenditures for prescription medications among persons with arthritis and other rheumatic conditions increased from ~$33 billion in 1997 to ~$75 billion in 2003.

In 2003, expenditures on behalf of persons with arthritis and other rheumatic conditions were much higher if non–arthritis and other rheumatic conditions were also present (mean expenditure $1,891 per year among persons with arthritis and other rheumatic conditions alone versus $7,643 among those with arthritis and other rheumatic conditions and non–arthritis and other rheumatic conditions) (Table 2). This is similar to the relationship of expenditures between the arthritis and other rheumatic conditions only and arthritis and other rheumatic conditions plus non–arthritis and other rheumatic conditions groups observed in 1997 ($1,381 versus $7,135, respectively).

Expenditures among persons with arthritis and other rheumatic conditions increased with age. In 2003, expenditures among persons with arthritis and other rheumatic conditions ages 18–44 years averaged $3,688; among such persons ages 45–54, 55–64, and ≥65, expenditures averaged $5,741, $8,341, and $9,520, respectively. Of note, of the 9.315 million increase in the number of individuals with arthritis and other rheumatic conditions between 1997 and 2003, ~66% of the increase occurred among persons ages 45–54 or 55–64. Also, expenditures increased by 19% and 25% between 1997 and 2003 in these 2 age groups, respectively, while increasing by only 2% among persons ages 18–44 or ≥65. Thus, the overall increase in expenditures for arthritis and other rheumatic conditions was associated both with population growth and with expenditure increases among persons ages 45–54 and 55–64.

Expenditures among persons with arthritis and other rheumatic conditions are only partially attributable to the arthritis and other rheumatic conditions; a significant fraction is due to other chronic conditions in these persons, as well as to their acute care, well care, and demographic characteristics. Table 3 shows the results of the analysis of the various kinds of expenditures attributable to arthritis and other rheumatic conditions among persons 18 years or older in 1997 and 2003. Overall, we estimated that in 1997 adults with arthritis and other rheumatic conditions had a mean increment in expenditures of $1,762 per person beyond what would be expected in similar persons without such conditions. By 2003 that estimate was virtually unchanged, at $1,752 per person. However, as with all expenditures incurred by persons with arthritis and other rheumatic conditions, there was a shift in the composition of incremental expenditures. The estimated increment in outpatient expenditures attributable to arthritis and other rheumatic conditions increased from $758 to $914 per person between 1997 and 2003 and the increment in expenditures for prescription medications increased from $141 to $338 per person, while the increment in inpatient expenditures declined from $508 to $352 per person.

Table 3
Increment in medical expenditures for AORC for the US noninstitutionalized population ages ≥18 years, 1997 versus 2003, by expenditure category*

Estimated aggregate incremental expenditures attributable to arthritis and other rheumatic conditions increased from $64.8 billion in 1997 to $80.8 billion in 2003; this was due wholly to the increase in the number of persons 18 and older with arthritis and other rheumatic conditions, since there was almost no change in per-person incremental expenditures. The aggregate incremental expenditures attributable to arthritis and other rheumatic conditions in 2003 represent ~0.7% of US GDP for that year.

We cross-validated the estimates of incremental costs presented in Table 3 by randomly dividing the data into 10 groups and calculating the estimated means, omitting each of the groups in turn. Virtually all of the 10-way cross-validation estimates (including total expenditures) were within 25% of the point estimates, with the exception of inpatient expenditures in 2003, for which estimates ranged from 46% below to 45% above the point estimate.

In 2003, as a result of the lower employment rates among persons with arthritis and other rheumatic conditions relative to those without such conditions (79.8% versus 90.6% [difference of 10.8%]), ~3.2 million persons with arthritis and other rheumatic conditions not currently employed would be working if that group had the same employment rate as persons without such conditions (Table 4). However, after controlling for differences between the 2 groups in demographic characteristics and comorbidity, the incremental employment gap was shown to be 2.3%, or ~0.7 million persons.

Table 4
Individual and aggregate employment and earnings losses among US persons with AORC ages 18–64 years, raw differences, and increments attributable to the AORC, 1997 and 2003*

The raw wage loss cost associated with the forgone earnings of the ~3.2 million persons with arthritis and other rheumatic conditions who would be working but were not was $113.0 billion ($3,826 per working-age person). We estimated that currently employed persons with arthritis and other rheumatic conditions actually work slightly more hours than those without arthritis and other rheumatic conditions and thus, have slightly higher earnings ($5 billion, or $213 per worker). The $113.0 billion lost because of lower employment rates and the $5 billion higher earnings among those employed netted a total earnings gap of $108.0 billion, or $3,613 per capita. After controlling for demographic characteristics and comorbidity, the increment in the net mean earnings gap attributable to arthritis and other rheumatic conditions averaged $1,590, for a total of $47.0 billion.

In 1997, the total earnings gap among persons with arthritis and other rheumatic conditions, updated to 2003 terms, was $99.0 billion, while the total incremental earnings gap attributable to arthritis and other rheumatic conditions was $43.3 billion. Thus, the total and incremental earnings gaps between persons with and those without arthritis and other rheumatic conditions did not change much between 1997 and 2003.

Because almost all of the earnings losses associated with arthritis and other rheumatic conditions occurred among those who were not working at all, stopping work loss is essential to containing the earnings losses associated with this set of conditions. Table 5 shows, for both 1997 and 2003, the demographic factors associated with employment status among persons ages 18–64 with arthritis and other rheumatic conditions who had worked. As indicated by the magnitude of the odds ratios for the demographic factors, there was little change in the factors affecting work loss between 1997 and 2003. In both years, the characteristics associated with a significantly reduced odds of employment included female sex (for 1997, odds ratio 0.6 [95% confidence interval 0.4–0.8]; for 2003, odds ratio 0.8 [95% confidence interval 0.7–1.0]) and increments of age, while increasing levels of education and white race were associated with increased odds of employment.

Table 5
Factors associated with employment of US persons with AORC ages 18–64 years, 1997 and 2003*

DISCUSSION

Concern about the economic impact of arthritis and other rheumatic conditions is usually associated with more general concerns about the economic impact of population aging. The results presented above demonstrate some of the initial impacts of these phenomena. Between 1997 and 2003, the number of persons with arthritis and other rheumatic conditions increased by more than a quarter, from just under 37 million to over 46 million, with almost all of the growth occurring among persons with both arthritis and other rheumatic conditions and other conditions. By the latter year, >21% of the population reported having arthritis and other rheumatic conditions. Furthermore, mean expenditures per arthritis and other rheumatic conditions case increased by ~10%. Of note, the increase in expenditures per case occurred disproportionately among persons ages 45–54 and 55–64 years, indicating that a large number of persons with arthritis and other rheumatic conditions will have these conditions for many years to come.

Thus, on the surface, there is some evidence consistent with the pessimistic scenario of increased numbers of persons with arthritis and other rheumatic conditions, increased expenditures per case, long durations of disease, and thus, rising costs associated with arthritis and other rheumatic conditions. Indeed, expenditures among persons with arthritis and other rheumatic conditions now account for the equivalent of ~3% of the US GDP.

On the other hand, almost all of the growth in the number of persons with arthritis and other rheumatic conditions occurred among those who also had other conditions: the number of persons with arthritis and other rheumatic conditions alone increased only from ~5.0 million to 5.3 million, whereas the number with arthritis and other rheumatic conditions and other conditions increased from ~31.8 to 40.8 million. Furthermore, much of the increase in the expenditures associated with arthritis and other rheumatic conditions between 1997 and 2003 is probably due to the increase in other chronic conditions experienced concomitantly by persons with arthritis and other rheumatic conditions, as well as their acute and well care, since the increment in expenditures for arthritis and other rheumatic conditions was virtually unchanged (from a mean of $1,762 in 1997 to $1,752 in 2003). Aggregated across the increased number of persons with arthritis and other rheumatic conditions, the increment in expenditures for arthritis and other rheumatic conditions increased from ~$65 billion in 1997 to ~$81 billion in 2003.

Although there was relatively little change in mean expenditures per case for arthritis and other rheumatic conditions and no change in incremental expenditures per case for this group of conditions, there has been substantial change in the relative magnitude of the different kinds of expenditures. In particular, expenditures for inpatient care declined in absolute and relative terms, whereas there was almost a doubling in expenditures for prescription medications among persons with arthritis and other rheumatic conditions (from $899 in 1997 to $1,635 in 2003) and more than a doubling in the increment in such expenditures (from $141 to $338). The growth in expenditures for prescription medicines should not be surprising since these were the years when the coxibs were marketed aggressively and biologic response modifiers first became available. Between 1997 and 2003, the mean number of prescriptions per person with arthritis and other rheumatic conditions and mean expenditure for each such person both increased by more than a third.

During this interval, the difference in the employment rate of persons with and those without arthritis and other rheumatic conditions narrowed, as did the per capita earnings gap of the 2 groups. As a result, the increase in the aggregate earnings gap of persons with and those without arthritis and other rheumatic conditions—whether measured in raw or in incremental terms—was entirely the result of the growth in the number of persons ages 18–64 with arthritis and other rheumatic conditions between 1997 and 2003. Moreover, population aging need not result in increased odds of work loss. After several decades in which employment rates among persons ages 55–64 declined, an increased demand for labor in recent years has slowed, and perhaps reversed, this trend (31). However, since persons with arthritis and other rheumatic conditions, like those with other health problems, are prone to the “last-hired, first-fired” phenomenon (34,35), they may be especially vulnerable to cyclical downturns.

The effect of population aging, which puts a greater number of persons at risk for developing arthritis and other rheumatic conditions, is reflected in the growth in both aggregate medical expenditures and earnings losses between 1997 and 2003. The increase in the prevalence of risk factors for the onset and heightened severity of arthritis and other rheumatic conditions, such as obesity, may intensify the effect of population aging on the economic impact of arthritis and other rheumatic conditions in the years to come. However, the impact of population growth can be mitigated by efforts to reduce medical care expenditures and increase employment, and there is some evidence consistent with this view. Total medical care expenditures per arthritis and other rheumatic conditions case increased very slowly between 1997 and 2003, while incremental medical care expenditures were stable, and raw and incremental earnings losses per case actually declined during this time.

Nevertheless, population growth did result in an increasing aggregate economic burden associated with arthritis and other rheumatic conditions between 1997 and 2003: total medical care expenditures grew from $233.5 billion to $321.8 billion after accounting for inflation, while incremental expenditures grew from $64.8 billion to $80.8 billion. Similarly, raw earnings losses increased from $99.0 billion to $108.0 billion, while incremental earnings losses expanded from $43.3 billion to $47.0 billion.

Arthritis and other rheumatic conditions exact a large and growing economic toll on the nation as a result of the increase in numbers of persons affected, rather than an increase in mean expenditures and earnings losses. Because the number of persons with arthritis and other rheumatic conditions is projected to increase steadily to nearly 67 million by 2030 (36), the economic impact is likely to continue to grow. The ability to prevent the onset of various types of arthritis is quite limited, so blunting of this growing economic impact will require cost-effective efforts to decrease mean medical expenditures and progress in meeting the nation’s Healthy People 2010 arthritis objectives to reduce the unemployment rate among adults with physician-diagnosed arthritis and decrease the proportion who are limited in their ability to work for pay. The latter may be achieved with greater use of underutilized interventions that have been shown to reduce the disability associated with arthritis and other rheumatic conditions, such as self-management education (37,38) and programs to increase physical activity (39). Population aging need not necessarily result in a proportionate increase in the economic impact of arthritis and other rheumatic conditions.

Acknowledgments

Supported by the Arthritis Program, Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention.

Footnotes

AUTHOR CONTRIBUTIONS

Dr. Yelin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Yelin, Murphy, Cisternas, Pasta, Helmick.

Acquisition of data. Not applicable (data in public domain).

Analysis and interpretation of data. Yelin, Murphy, Cisternas, Foreman, Pasta, Helmick.

Manuscript preparation. Yelin, Murphy, Cisternas, Foreman, Pasta, Helmick.

Statistical analysis. Yelin (as collaborator), Cisternas, Foreman, Pasta.

References

1. Brooks P. Impact of osteoarthritis on individuals and society: how much disability? Social consequences and health economic implications. Curr Opin Rheumatol. 2002;14:573–7. [PubMed]
2. Brooks P. The burden of musculoskeletal disease—a global perspective. Clin Rheumatol. 2006;25:778–81. [PubMed]
3. Reginster J. The prevalence and burden of arthritis. Rheumatology (Oxford) 2002;41 (Suppl 1):3–6. [PubMed]
4. Woolfe A, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ. 2003;81:646–56. [PubMed]
5. Michaud K, Messer J, Choi HK, Wolfe F. Direct medical costs and their predictors in patients with rheumatoid arthritis: a three-year study of 7,527 patients. Arthritis Rheum. 2003;48:2750–62. [PubMed]
6. March L, Bachmeier C. Economics of osteoarthritis: a global perspective. Ballieres Clin Rheumatol. 1997;11:817–34. [PubMed]
7. Pugner K, Scott D, Holmes J, Hieke K. The costs of rheumatoid arthritis: an international long-term view. Semin Arthritis Rheum. 2000;29:305–20. [PubMed]
8. Cooper N. Economic burden of rheumatoid arthritis: a systematic review. Rheumatology (Oxford) 2000;39:28–33. [PubMed]
9. Fautrel B, Guillemin F. Cost of illness studies in rheumatic diseases. Curr Opin Rheumatol. 2002;14:121–6. [PubMed]
10. Rat A, Boissier M. Rheumatoid arthritis: direct and indirect costs. Joint Bone Spine. 2004;71:518–24. [PubMed]
11. Soderlin M, Kautiainen H, Jonsson D, Skogh T, Leirisalo-Repo M. The costs of early imflammatory joint disease: a population-based study in southern Sweden. Scand J Rheumatol. 2003;32:216–24. [PubMed]
12. Breedveld F. Osteoarthritis—the impact of a serious disease. Rheumatology (Oxford) 2004;43 (Suppl 1):i4–8. [PubMed]
13. Gupta S, Hawker G, Laporte A, Croxford R, Coyte P. The economic burden of disabling hip and knee osteoarthritis (OA) from the perspective of individuals living with this condition. Rheumatology (Oxford) 2005;44:1531–7. [PubMed]
14. Rice D. Estimating the cost of illness. Hyattsville (MD): National Center for Health Statistics; 1966. Health Economic Series no. 6.
15. Cooper B, Rice D. The economic cost of illness revisited: Health Economics Series no. 6. Soc Secur Bull. 1979;39:21–35. [PubMed]
16. Rice D, Hodgson T, Kopstein A. The economic costs of illness: a replication and update. Health Care Fin Rev. 1985;7:61–80. [PMC free article] [PubMed]
17. Rice D. The economic burden of musculoskeletal conditions, 1995. In: Praemer A, Furner S, Rice D, editors. Musculoskeletal conditions in the United States. Rosemont (IL): American Academy of Orthopaedic Surgeons; 1999. pp. 139–62.
18. Wigle D, Mao Y, Wong T, Lane R. Economic burden of illness in Canada, 1986. Chronic Diseases in Canada. 1991 May-June12
19. Access Economics. Cost of arthritis to the Australian community: The Arthritis Foundation Submission to the Industry Commission Inquiry into Charitable Organizations. Canberra: Arthritis Foundation of Australia; 1994. pp. 14–22.
20. Badley E. The economic burden of musculoskeletal disorders in Canada is similar to that for cancer, and may be higher. J Rheumatol. 1995;22:204–6. [PubMed]
21. Katz PP, Yelin EH. The development of depressive symptoms among women with rheumatoid arthritis: the role of function. Arthritis Rheum. 1995;38:49–56. [PubMed]
22. Yelin E, Herrndorf A, Trupin L, Sonneborn D. A national study of medical care expenditures for musculoskeletal conditions: the impact of health insurance and managed care. Arthritis Rheum. 2001;44:1160–9. [PubMed]
23. Dunlop DD, Manheim LM, Yelin EH, Song J, Chang RW. The costs of arthritis. Arthritis Rheum. 2003;49:101–13. [PubMed]
24. Yelin E, Cisternas MG, Pasta DJ, Trupin L, Murphy L, Helmick CG. Medical care expenditures and earnings losses of persons with arthritis and other rheumatic conditions in the United States in 1997: total and incremental estimates. Arthritis Rheum. 2004;50:2317–26. [PubMed]
25. Cohen S. Sample design of the 1996 Medical Expenditure Panel Survey Household Component. Rockville (MD): Agency for Health Care Policy and Research; 1997. AHCPR publication no. 97-0027.
26. Cohen S. Sample design of the 1997 Medical Expenditure Panel Survey Household Component. Rockville (MD): Agency for Healthcare Research and Quality; 2000. AHRQ publication no. 01-0001.
27. Cohen S, DiGaetano R, Goksel H. Estimation procedures in the 1996 Medical Expenditure Panel Survey Household Component. Rockville (MD): Agency for Health Care Policy and Research; 1999. AHCPR publication no. 99-0027.
28. SAS Institute. SAS v9.1. Cary (NC): SAS Institute; 2002–2003.
29. Harwood H, Fountain D, Livermore G. The economic costs of alcohol and drug abuse in the United States. Rockville (MD): National Institute on Drug Abuse; 1998. NIH publication no. 98-4327, Appendix B.
30. Duan N, Manning W, Morris C, Newhouse J. A comparison of alternative models for the demand for medical care. J Bus Econ Stat. 1983;1:115–26.
31. Duan N. Smearing estimate: a nonparametric retransformation method. J Am Stat Assoc. 1983;78:605–10.
32. US Bureau of the Census. Statistical abstract of the United States, 2006. Washington: USGPO; 2006. p. 482.
33. Carroll R. Changes affecting the Employment Cost Index: an overview. Monthly Labor Review. 2006;129:3–5.
34. Yelin EH, Katz PP. Labor force participation among persons with musculoskeletal conditions, 1970–1987: national estimates derived from a series of cross-sections. Arthritis Rheum. 1991;34:1361–70. [PubMed]
35. Trupin L, Yelin E. Final Report. Social Security Administration Disability Research Institute; 2006. Multiple jeopardies in the California labor market: the conjoint role of disability, race, gender, and age.
36. Hootman JM, Helmick CG. Projections of US prevalence of arthritis and associated activity limitations. Arthritis Rheum. 2006;54:226–9. [PubMed]
37. Redelmeier D, Lorig K. Assessing the clinical importance of symptomatic improvements: an illustration in rheumatology. Arch Intern Med. 1993;153:1337–42. [PubMed]
38. Kruger J, Helmick C, Callahan L, Haddix A. Cost-effectiveness of the Arthritis Self-Help Course. Arch Intern Med. 1998;158:1245–9. [PubMed]
39. Minor MA. Exercise and Physical Activity Conference, St. Louis, Missouri: Exercise and Arthritis “we know a little bit about a lot of things. .” [editorial] Arthritis Rheum. 2003;49:1–2. [PubMed]