TRENDS IN PERMANENT WORK DISABILITY ASSOCIATED WITH RHEUMATOID ARTHRITIS IN THE UNITED STATES, 1999 – 2015

Abstract

Background.

Advances in treatment over the past 20 years has resulted in improved control of rheumatoid arthritis (RA), but whether there has been a decrease in permanent work disability associated with RA in the U.S. has not been examined.

Methods.

Medicare data from 1999 to 2015 were used to identify beneficiaries age 20 to 59 with RA who became eligible for Medicare coverage under Social Security Disability Insurance. Diagnosis of RA was based on physician claims in the first year of enrollment. Annual rates of enrollment were sex- and age-standardized to the 2000 U.S. population.

Results.

The study included 97,787 beneficiaries with RA and Social Security Disability Insurance across all years. Medicare enrollment was 26.0 per million in 1999 and 26.0 per million in 2015. Rates increased following the Great Recession of 2008–09 before returning to pre-recession levels. There was no linear trend over time after adjusting for the annual national unemployment rate (relative risk 0.99 per year; 95% confidence interval 0.99, 1.00; p = 0.69). Risks of work disability were much higher among workers over age 50.

Conclusion.

Based on Medicare enrollment by recipients of Social Security Disability Insurance, there was no decrease in permanent work disability among young and middle-age workers with RA in the U.S. between 1999 and 2015.

The treatment of rheumatoid arthritis (RA) has changed dramatically over the past twenty years with the emphasis on early and consistent use of disease-modifying medications and the introduction of biologics. With better treatments, the health outcomes of patients with RA have improved over time, with many studies reporting decreases in disease activity, less functional difficulty, less joint damage, and reduced need for joint surgery.[1,2] Clinical remission is now an achievable goal.

Although joint inflammation and damage have decreased, it is not clear if other long-term outcomes such as permanent work disability have improved. In studies of patients observed in the late 1980s and 1990s, 30% – 50% of employed patients developed permanent work disability after 10 years of RA.[3,4] Many studies suggest that tumor necrosis factor inhibitors are efficacious in preventing short-term work loss, which holds the prospect of lower incidences of permanent work loss.[5,6] However, permanent work disability in RA depends not only on the severity of joint symptoms and impairments, but also on the type of work, workplace accommodations, psychological response to illness, and social and financial supports.[4] Risks of permanent work loss are higher among older persons with RA and those with less formal education and physically-demanding jobs.[4]

Population-based studies from Sweden and Finland reported decreases of more than 50% in rates of permanent work disability among persons with RA between 1990 and 2010.[7,8] In Norway, rates of permanent disability pensioning for RA were stable between 1968 and 1997, but lagged rising incidences in the general population between 1983 and 1997.[9] Similar national studies have not been reported in the United States. Social Security Disability Insurance (SSDI) provides permanent benefits to eligible American workers who are certified as having a medical condition that makes them unable to work for at least one year and is unlikely to improve. SSDI recipients are eligible for Medicare insurance after two years. The aim of this study was to examine rates of Medicare enrollment from 1999 to 2015 among SSDI recipients age 20 to 59 with RA.

METHODS

The data source was 100% fee-for-service Medicare inpatient and outpatient administrative claims files from 1999 to 2015. In each year, I identified newly-enrolled beneficiaries age 20 to 59 who entered Medicare via SSDI eligibility. I excluded those who entered because of end-stage renal disease. Among these beneficiaries, I identified those with RA based on presence of at least two claims on separate dates in the first year of Medicare enrollment that had a principal diagnosis of RA based on an International Classification of Diseases, Ninth Revision code of 714.[10] This criterion was based on the consideration that medical care would be sought specifically for the condition that permitted Medicare eligibility. The study protocol was approved by the National Institute of Diabetes and Digestive and Kidney Diseases Institutional Review Board, which waived the requirement for informed consent.

I examined rates by year of enrollment relative to the U.S. population, based on census data. Rates were sex- and age-standardized (in 5-year age groups) to the U.S. population in 2000. Applications for SSDI benefits and SSDI enrollment increase during economic downturns and job loss.[11] National unemployment rates varied between 4.0% and 6.0% from 1999 to 2008, and increased to 9.3% and 9.6% in 2009 and 2010, before decreasing gradually to 5.3% in 2015.[12] To account for changes associated with the 2008–2009 Great Recession, I used Poisson regression models to determine if there was a linear trend in SSDI enrollment while adjusting for changes in the national unemployment rate over these years. The independent variables in the model were calendar year, the two-year lagged national unemployment rate (to account for the time between SSDI claim and Medicare enrollment), sex, and indicator variables for each 5-year age group. This analysis provided the relative risk of the average yearly change in rate of SSDI enrollment, while partitioning out the contribution of the unemployment rate and variations in the sex and age composition of the sample over time.

In sensitivity analyses, I examined beneficiaries who had three or more claims with RA as the principal diagnosis in the first year in Medicare, and those age 20 to 39 on enrollment in Medicare, who may have had greater potential to benefit from recent treatment advances.

RESULTS

Across all years, 97,787 beneficiaries (75.5% women; mean (standard deviation) age 50.1 (8.3) years) enrolled in Medicare under SSDI with RA claims in their first year. Sixty-nine percent had their first visit for RA within 90 days of Medicare enrollment. The rate of Medicare enrollment under SSDI was 26.0 per million in 1999 (Table 1 and Figure 1). Rates were comparable or higher in each subsequent year, and peaked at 40.0 per million in 2011, two years after the Great Recession.

Table 1.

Rates of Medicare enrollment via Social Security Disability Insurance by beneficiaries with rheumatoid arthritis.

Year	Number	Sex-age standardized rate, per million population (95% confidence interval)
1999	4057	26.0 (25.2, 26.9)
2000	4181	26.8 (26.0, 27.7)
2001	4685	29.3 (28.4, 30.2)
2002	5291	32.1 (31.2, 33.0)
2003	5757	34.2 (33.4, 35.2)
2004	5781	33.5 (32.7, 34.5)
2005	5721	32.3 (31.4, 33.2)
2006	5741	31.7 (30.8, 32.5)
2007	5361	29.4 (28.6, 30.2)
2008	5380	29.3 (28.5, 30.1)
2009	6027	32.5 (31.7, 33.4)
2010	6574	35.2 (34.3, 36.1)
2011	7530	40.0 (39.1, 41.0)
2012	7319	38.3 (37.4, 39.3)
2013	7013	36.5 (35.6, 37.4)
2014	6257	32.4 (31.5, 33.2)
2015	5112	26.0 (25.3, 26.8)

Figure 1.

Age and sex-standardized rates of Medicare enrollment under Social Security Disability Insurance by year for beneficiaries with rheumatoid arthritis. Error bars are 95% confidence limits.

There was no linear trend in Medicare enrollment among beneficiaries with RA over time (relative risk (RR) 0.99 per year; 95% confidence interval (CI) 0.99, 1.00; p = 0.69) after adjustment for the rise in unemployment rates in 2009–2012. Risks increased progressively with age. Compared to those age 20–24, the RR for enrollment among those age 25–29 was 1.75 (95% CI 1.63, 1.88), while the RR among those age 50–54 was 20.85 (95% CI 19.68, 21.98) and for those age 55–59 was 31.50 (95% CI 29.66, 33.11). Risks were lower among men compared to women (RR 0.33; 95% CI 0.33, 0.34).

The subgroup with three claims for RA in the first year included 75,930 beneficiaries (76.0% women; mean age 50.2 (8.3) years), while the subgroup age 20 to 39 included 12,564 beneficiaries (79.2% women; mean age 33.2 (5.2) years). While the absolute rates were lower in both subgroups, the incidences in these subgroups paralleled the incidence in the overall group, with no decrease over time (Figure 1). These results indicate that use of a more stringent requirement for inclusion did not affect the conclusion, and that the results of the large proportion of middle-aged beneficiaries were not obscuring a decrease in SSDI enrollment over time among young adults.

DISCUSSION

These results do not indicate a decrease in rates of permanent work disability among American workers with RA between 1999 and 2015. Apart from temporary increases in Medicare enrollment under SSDI as a consequence of labor market changes following the Great Recession, rates of enrollment were stable over this period.

Older workers were most susceptible to work disability, as also shown in many prior studies.[4] Compared to younger workers, this group likely included a higher proportion of workers with more longstanding RA, whose joint damage may have been less amenable to improvement.[13] However, rates were also stable among those age 20 to 39, who might have been expected to gain more benefit from recent treatment advances.

These findings raise questions about access to newer treatments, particularly by low income, blue-collar workers who are most at risk for work disability related to RA.[4] We did not have data on medication use, and could not examine this question directly. However, lower income and less formal education have been consistently associated with less access to disease-modifying medications, including biologics, in patients with RA in the U.S.[14,15] The absence of a national decrease in work disability over time may therefore reflect lack of access to treatment advances by workers of lower socioeconomic status, who are at highest risk of work disability. Access to new treatments by highly-educated workers may not have had an impact on national rates of work disability because their baseline risk of health-related job loss was comparative low. It is also possible that despite improvements in physical health over time among workers with RA, social or psychological influences continued to foster work disability. Differences between these results and those from Nordic countries may relate to differences in the organization and financing of health care.[7,8]

An individual’s decision to pursue a work disability claim includes considerations of the severity of illness, coping resources and skills, prospects for future improvement in health, job requirements and accommodations, opportunities for retraining, age and the time horizon, social supports, including economic support from family, and personal wealth.[3,4] Our results also highlight the association of disability claims with the national economy, and demonstrate the major role that macro level factors have on what has often been considered a personal decision. In the U.S., SSDI claims increased by more than 100,000 per quarter following the Great Recession, and tracked more closely with changes in the national unemployment rate than the gross domestic product.[11] These data show a similar peak among persons with RA in 2011–12, which aligns with the lag between enrollment in SSDI and eligibility for Medicare. The smaller peak in 2003–04 may be a consequence of the 2001 recession.[11] If similar associations hold for the current economic downturn and unemployment due to the COVID-19 pandemic, we might expect sharp increases in disability claims for RA in the near future. This effect may be particularly pronounced given the disproportionate economic impact of the pandemic on female service workers.

The study is limited in that data after 2015 were not available. However, there was no suggestion of a decrease in work disability through 2015, many years after the introduction of new treatments. RA might have developed between enrollment in SSDI and enrollment in Medicare in some individuals, but this was likely rare. The analysis also assumed the prevalence of RA has been stable over these years. As in any study of administrative claims, there may be inaccuracies in coding, but differential inaccuracies over time that would be large enough to alter the trends are unlikely. The validity of coding is supported by the detection of the peak in rates following the Great Recession, as also found in all-cause disability claims.[11]

These findings suggest that recent treatment advances have not yet had a major impact on permanent work disability associated with RA at the population level in the U.S. Future research should investigate the association between access to treatment and rates of work disability across the population.

SIGNIFICANCE AND INNOVATION.

• This is the first population-based study of rates of permanent work disability associated with rheumatoid arthritis in the United States.

• Rates of Medicare enrollment under Social Security Disability Insurance by persons with rheumatoid arthritis were the same in 1999 and 2015, after peaking in 2011 following the Great Recession.

• This is the first study to demonstrate the impact of national economic downturns on work disability in rheumatoid arthritis.

• The absence of a decrease in rates of permanent work disability among U.S. workers with rheumatoid arthritis may indicate inadequate access to treatment among those at highest risk for work disability.