Work Disability and Its Economic Effect on
55–64-Year-Old Adults With Rheumatoid Arthritis
SARALYNN ALLAIRE,1FREDERICK WOLFE,2JINGBO NIU,1MICHAEL LAVALLEY,1AND
Objective. To examine the extent and financial impact of work disability among older workers with rheumatoid arthritis
Methods. Year 2002 data from 5,419 subjects with RA <65 years of age in the National Data Bank for Rheumatic Diseases
were used, along with US population data. Measures of work disability were employment status, part-time work, sick day
use, and limitation in work demands; the latter was assessed by the Work Limitations Questionnaire (WLQ). Measures
of financial status were median household income and poverty level income. Statistical procedures included logistic and
linear regression, Wilcoxon’s rank sum test, and chi-square test.
Results. Despite being better educated, subjects with RA ages 55–64 years had lower employment rates than individuals
of the same age in the US (women 40% versus 53% and men 54% versus 66%). These older subjects with RA had stopped
working more often than younger subjects with RA, and more worked part time (40% versus 34%; P < 0.01). However,
the older subjects used sick time less often than younger subjects (35% versus 41%; P < 0.01) and were similarly limited
in job demands, e.g., physical demands (mean WLQ subscale score 27.0 versus 26.6; P ? 0.65). Median household income
of older employed subjects was $20,000 greater than that of retired subjects; 56% of retired subjects had incomes lower
than US median income versus 32% of employed subjects, and 11% had income below the poverty level.
Conclusion. Premature work cessation in persons with RA ages 55–64 years is a serious problem that needs to be
KEY WORDS. Work disability; Rheumatoid arthritis; Older adults.
Employment and work disability issues among persons
ages 55 to 64 years will become increasingly compelling as
the US workforce ages along with the baby boomer gener-
ation. Shortages of labor and skills are predicted as baby
boomers retire (1), and individuals will need enough in-
come to last through their retirement period, which is
lengthening as the population ages. These issues are espe-
cially pertinent for persons with rheumatoid arthritis (RA)
because the incidence (2) and prevalence of RA are high in
this age bracket, and the number of persons with RA in this
age bracket will increase as the baby boomer generation
Work disability among older workers has previously not
been regarded as a serious problem by society. It is gener-
ally believed that the work capacity of older workers is
limited (1,3), and that the prospect for return to work
among older workers with health problems is minimal
(4,5); plus, the workers’ reliance on disability pension
programs lasts only for a short while before they are trans-
ferred to retirement programs (6).
For many individuals, however, work disability has sub-
stantial negative ramifications. Studies suggest that those
who stop working prematurely due to health have reduced
physical and psychosocial wellbeing (7–9), and they com-
monly experience considerable loss of income. Long-term
and immediate income can be affected because work dis-
ability in persons of older age prevents many from devel-
oping adequate retirement savings or may necessitate them
prematurely withdrawing their retirement savings. As new
RA treatments improve mortality and expected lifespan
Supported by the NIH (grant P60-AR-47785 from the Na-
tional Institute of Arthritis and Musculoskeletal and Skin
Diseases), Bethesda, MD.
1Saralynn Allaire, ScD, Jingbo Niu, MS, Michael LaVal-
ley, PhD: Boston University, Boston, Massachusetts;2Fred-
erick Wolfe, MD: National Data Bank for Rheumatic Dis-
eases, Wichita, Kansas;
Bank for Rheumatic Diseases, Wichita, Kansas, and Center
for Primary Care and Outcomes Research, Stanford Univer-
sity, Stanford, California.
Address correspondence to Saralynn Allaire, ScD, Clini-
cal Epidemiology Research and Training Unit, A203, 715
Albany Street, Boston, MA 02118. E-mail: email@example.com.
Submitted for publication December 6, 2004; accepted in
revised form March 6, 2005.
3Kaleb Michaud: National Data
Arthritis & Rheumatism (Arthritis Care & Research)
Vol. 53, No. 4, August 15, 2005, pp 603–608
© 2005, American College of Rheumatology
lengthens, work disability among persons with RA ages
55–64 years could sentence many of them to live the
remaining years of their life in poverty.
Age is known to increase risk for RA work disability
(5,10); however, the extent of work disability and its im-
pact have not been examined in older workers with RA.
This may be because work disability among older workers
has not been considered a serious problem and because
sample sizes in previous clinical studies have often been
This study examined the extent and financial impact of
work disability among persons in the older worker age
bracket using 3 research questions. Because most persons
in the US retire prior to age 65 (1,11), our first question
was, “Does the employment rate of persons with RA ages
55 to 64 years differ from that of their US peers?” Next,
although it is generally believed that the work capacity of
older workers is limited compared with that of younger
workers (1,3), and as a consequence vocational interven-
tion or employer efforts for older workers are less worth-
while (5,12), no studies have compared the work capacity
of the 2 age groups. Therefore we asked, “Are older work-
ers with RA more limited in their capacity for work than
younger workers with RA?” Finally, until very recently
there has been a trend in the US toward earlier retirement
(1,11), which is presumably based on many individuals
having a comfortable retirement income. The actual finan-
cial impact of early withdrawal from the workforce on
persons with RA ages 55–64 years is unknown, although it
is anticipated to be negative. Consequently, we asked,
“Among older persons with RA, what is the financial im-
pact of early withdrawal from the workforce?”
SUBJECTS AND METHODS
Subjects. The definition of older workers used by the
US Census Bureau is the age range of 55–64 years (11).
Although the US eligibility age for full retirement benefits
has currently increased to 67 years, it was 65 years in 2002
when data used in this study were collected; therefore we
used 64 as the upper age limitation for older workers.
Study subjects were participants in the National Data
Bank for Rheumatic Diseases (NDB) longitudinal study of
RA outcomes. The NDB-RA is a dynamic cohort of
?10,000 individuals; participants are added continuously,
and ?8% of the participants in the NDB-RA withdraw
from the Data Bank each year. Participants are recruited
from 2 sources: nonselected patients from the practices of
US rheumatologists, and individuals enrolled as part of
pharmaceutical company-sponsored registries. At the time
of our study enrollment, rheumatology practice patients
were referred from 429 rheumatologists located through-
out the US; more than 90% of the rheumatologists were in
private practice and were not full-time university physi-
Data are collected from NDB participants biannually.
For our study, we used data collected in either January or
June, 2002. Data from the first time period only were used
for each subject; the questionnaire time reference periods
were either July to December 2001, or January to June
The majority of NDB participants fill out detailed, com-
prehensive questionnaires, but ?12% fill out short ques-
tionnaires that contain no questions about employment. In
addition, we had defined the upper age limit of older
workers as 64 years. Therefore our study’s sample was
limited to NDB participants who supplied employment
information and were younger than 65 years of age. Be-
cause of the different requirements of the 3 research ques-
tions, different sample subsets were used to address each
of the questions.
Sample subsets, measures, and analyses for research
questions. Does the employment rate of persons with RA
ages 55–64 years differ from that of their US peers? The
RA sample subset used for this question included subjects
ages 55–64 years. Sex-specific data on employment rates
are available for this age group in the 2002 US Current
Population Survey (13), therefore we examined employ-
ment rates separately for men and women. We also exam-
ined the educational and racial characteristics of US citi-
zens in this age group using 2000 Census Bureau data (14)
because these characteristics, in addition to sex, strongly
influence employment rates. NDB-RA participants as a
whole have increased educational attainment and are less
likely to be members of minority groups than the U.S.
population, and differences in these characteristics might
obscure differences in employment rates.
Because statistical comparison between our sample and
the US Current Population Survey sample was not possi-
ble, we observed the proportions of men and women in
both the RA and US samples in the 55–64-year age group
who were employed. US population employment data for
this age group are available only by sex, educational at-
tainment, and race characteristics alone, so we examined
the educational attainment and racial characteristics of the
RA and US samples separately.
Are older workers with RA more limited in their capacity
for work than younger workers with RA? The first mea-
sure of work capacity was work cessation, so the appro-
priate sample subset for this analysis was subjects who
were employed at disease onset. Within this subset, we
compared the proportion of subjects ages 55–64 years who
were no longer employed with that of younger subjects,
i.e., those ages 18–54 years.
The remaining measures of work capacity assessed work
productivity, so the appropriate sample subset for these
comparisons was subjects who were currently employed.
The work productivity measures were part-time versus
full-time work, use of sick days, and limitation in ability to
work. Part-time work was defined as ?35 hours per week
or ?40 weeks per year. Sick day use was defined as any
sick days taken in the prior 6 months. Limitation in ability
to work was measured using the Work Limitations Ques-
tionnaire (WLQ), which was included in the 2002 NDB
survey questionnaires. The WLQ assesses the on-the-job
impact of chronic health problems and health-related pro-
ductivity loss (15). Items assess limitations in performing 4
dimensions of job demands: physical, time, mental-inter-
personal, and output demands; each dimension is mea-
604Allaire et al
sured in a separate subscale. These items are common
attributes of job roles, are sensitive to the effects of phys-
ical and/or mental health problems, and are relevant to
productivity. The WLQ has established reliability and va-
lidity in arthritis samples (16).
We used logistic regression to assess the impact of older
age versus younger age on current employment status, i.e.,
employed versus not employed. This analysis was ad-
justed for sex, race, education, functional limitation, and
disease duration. Functional limitation was measured by
the Health Assessment Questionnaire (HAQ) (17). Similar
logistic regression analyses assessed the impact of older
age on part-time versus full-time work and any use of sick
time; in addition to sex, education, race, functional limi-
tation, and disease duration, these analyses were adjusted
for the physical demand level of subjects’ jobs and, in the
case of sick time use, part-time versus full-time work. Job
physical demand was assessed using a scale from the 1978
Survey of Disability and Work (18,19). Linear regression
was used to examine differences in limitation in the 4
types of job demands, i.e., physical, time management,
mental/interpersonal, and output demands, in the older
and younger worker age groups. These analyses were ad-
justed for sex, job physical demand, functional limitation,
and disease duration.
Among older persons with RA, what is the financial
impact of early withdrawal from the workforce? We com-
pared the financial status of 2 different sample subsets of
subjects ages 55–64 years. The first subset included sub-
jects who were currently employed, and the second in-
cluded subjects who were employed at disease onset, but
were not currently working. Measures of financial status
were median household income, the proportion with less
than median income, and the proportion below the US
poverty income amount. The US median and poverty level
income amounts were based on 2000–2001 US Census
Bureau data; median income was $42,000 and poverty
level income was $11,000 (20). In the NDB questionnaire
the response categories for the household income question
cover increments of approximately $10,000 of income,
e.g., $10,000–$19,999, beginning with ?$10,000 and end-
ing with ?$100,000.
The household income frequency distributions of each
of the currently and formerly employed subject groups
were examined. The NDB income response categories clos-
est to the US median and poverty level incomes were
chosen, i.e., ?$40,000 and ?$10,000, and the proportions
of subjects in both groups below these levels were ob-
served. We then used the Wilcoxon rank sum test to com-
pare the total household incomes of the employed and
retired subject groups and the chi-square test to compare
the proportions of each group that were below the US
median income or below the poverty level income.
More than 11,000 NDB participants with RA provided data
in 2002, and 9,725 participants filled out comprehensive
questionnaires. Of these, 422 participants gave no infor-
mation about their employment status. A total of 5,419 of
the remaining 9,303 participants were younger than 65
years of age, and these participants formed the full sample
of subjects for our study. Of these subjects, 2,499 were ages
55–64 years (mean ? SD age 60.9 ? 2.9 years, 77%
women, 91% white, 75% married) (Table 1). Mean ? SD
disease duration was 14.7 ? 10.2 years, and mean func-
tional limitation (HAQ score) was 1.1 ? 0.7 on a 0–3 scale,
where 0 equals no limitation. Fifty-four percent of the
subjects had an education level beyond high school, the
level considered to confer an employment advantage; 43%
were currently employed, whereas 84% had been em-
ployed at disease onset.
Employment rate of persons with RA ages 55–64 years
versus US population. A larger portion of the 2,499 sub-
jects with RA ages 55–64 years had attained education
beyond high school than their US peers (54% versus 46%),
and subjects with RA were more often of white race (91%
versus 78%). However, despite these employment advan-
tages, they were employed at lower rates (women with RA
40% versus 53%, and men with RA 54% versus 66%).
Work capacity of older workers with RA versus
younger workers with RA. A total of 4,470 subjects in our
study’s full sample were employed at disease onset, 2,044
ages 55–64 years and 2,426 ages 18–54 years. A substan-
tially greater portion of the older subjects had stopped
working compared with younger subjects (53% versus
31%; P ? 0.0001) (Table 2).
A total of 3,009 subjects in the full sample were cur-
rently employed, 1,073 ages 55–64 years and 1,936 ages
18–54 years. The older workers were more likely to work
part time than younger workers (40% versus 34%; P ?
0.0005) (Table 2). However, the older workers’ work pro-
ductivity by the other measures was equal to or better than
that of younger workers. Thirty-five percent of older work-
ers had used sick days in the prior 6 months compared
with 41% of younger workers (P ? 0.0006), and the older
workers were not more limited in meeting the physical,
time management, or output demands of their jobs and
were less limited in meeting the job mental/interpersonal
demands (mean WLQ subscale score 14.1 versus 15.9; P ?
Table 1. Demographic, disease, and employment
characteristics of 2,499 subjects ages 55–64 years*
Age, mean ? SD years
Education ? high school
Disease duration, mean ? SD years
Functional limitation, mean ? SD score‡
Employed at disease onset
60.0 ? 2.9
14.7 ? 10.2
1.1 ? 0.7
* Unless otherwise indicated, values are the number (%).
† Denominator values are slightly different for each characteristic
because of missing data.
‡ Health Assessment Questionnaire (range 0–3, where 0 ? no
Economic Effect of Work Disability on Adults 605
0.02). Their functional limitation was greater than that of
the younger workers (mean ? SD HAQ score 0.83 ? 0.6
versus 0.75 ? 0.6; P ? 0.0005), and they had longer disease
duration (mean 13.4 ? 10.0 years versus 11.8 ? 8.4 years).
Older workers also were more likely to have a comorbidity
than younger workers (63% versus 49%; P ? 0.0001),
although they were not more likely to have a major comor-
bidity (43% versus 39%; P ? 0.1).
Financial impact of early withdrawal from the work-
force among older persons with RA. A total of 1,000 of
1,073 currently employed subjects ages 55–64 years and
952 of 1,081 subjects of the same age employed at disease
onset but not currently working provided financial infor-
mation. On average, subjects who were no longer working
were considerably less financially secure than those who
were currently employed. As depicted in Table 3, the
median household income of the subjects no longer work-
ing was $30,000 compared with the $50,000 median in-
come of employed subjects (P ? 0.0001), and those no
longer working were much more likely to have a house-
hold income that was below the US median (56% versus
32% of employed subjects; P ? 0.0001). Although these
subjects were educationally and racially advantaged com-
pared with US individuals of the same age, 11% of those
who had stopped working had household incomes below
the US poverty level, compared with 2% of employed
subjects (P ? 0.0001).
Not unexpectedly, we found evidence of substantial work
disability defined as premature work cessation among per-
sons with RA ages 55–64 years. The ?2,500 subjects with
RA in our sample of this age were considerably less likely
to be employed than their US age mates, despite having
educational and racial employment advantages. Also,
compared with younger subjects with RA, a larger percent-
age of older subjects had stopped working, and they were
less likely to work full time. However, the work produc-
tivity of older, employed subjects with RA, as measured by
use of sick days and on-the-job limitation in job de-
mands, was at least equivalent to that of their younger
peers. Because older subjects had more extensive func-
tional limitation, longer disease duration, and were
Table 2. Work disability in older versus younger subjects with rheumatoid arthritis*
Measure of work disability
18–54 years Adjusted P
No. subjects employed at disease onset
No longer working, no (%)
No. currently employed subjects
Part time work, no (%)
Use of any sick days within prior
6 months, no (%)
Limitation in job demands, mean (95% CI)¶
Time management demands
* 95% CI ? 95% confidence interval.
† Adjusted for sex, race, education, functional limitation, disease duration.
‡ Adjusted for sex, race, education, functional limitation, disease duration, job physical demand.
§ Adjusted for sex, race, education, functional limitation, disease duration, job physical demand, full-
time versus part-time work.
¶ Work Limitations Questionnaire (subscale scores 0–100, where 0 ? no limitation).
# Adjusted for sex, job physical demand, functional limitation, disease duration.
Table 3. Financial status of early retirees versus employed older subjects with
Measure of financial status
n ? 952
n ? 1,000
Median household income
No (%) household income ? median
No (%) household income ? poverty
103 (11) 17 (2)
* Subjects ages 55–64 years employed at disease onset, but not currently.
† Subjects ages 55–64 who are currently employed.
‡ US median level in 2000 is $42,000.
§ US median level in 2001 is $11,000.
606Allaire et al
more likely to have a comorbidity than younger subjects,
their comparable on-the-job function may be explained
by more frequent self-selection into jobs with less phys-
ical and time demands (21) and/or greater use of job
In previous studies, older age has been examined as a
risk factor for work disability and was consistently found
to increase risk for premature work cessation (5,10). We
believe this is the first study to examine the extent and
impact of work disability among persons with RA ages
55–64 years. The lack of attention to work disability in this
age group likely stems from the belief that it is not a
serious problem and that employment preservation efforts
for older workers are not worthwhile (4,5). However, we
found that the financial toll of early retirement on the
household incomes of subjects ages 55–64 years was sub-
stantial. Fifty-six percent of these subjects who had
stopped working had household incomes below the US
median income compared with 32% of employed subjects,
and the household income of 11% of those no longer
working was below the US poverty level.
The financial toll of premature retirement on persons
with RA in the US is likely to increase in the future
because of changes in Social Security and other retirement
pension plans. The age of eligibility to receive the full
Social Security retirement benefit amount is increasing
from 65 to 67 years, and the partial benefit for early retire-
ment is dropping from 80% to 70% of the full benefit.
Furthermore, the type of private retirement plan available
to most workers has changed from the traditional employ-
er-provided, defined-contribution plan to tax-deferred ac-
counts (e.g., 401[k]) or cash balance plans, and this change
is expected to reduce retirement income and penalize
early retirement (22,23). Work disability among older per-
sons will also become more of a societal problem because
of the large size of the baby boomer generation. This will
increase the number of applications to disability benefit
programs, and, because of the increasing age of eligibility
for retirement benefits, lengthen time on disability pro-
Because of the strongly negative financial consequences
of premature retirement on many persons with RA, we
believe work disability in persons ages ?55 years is a
serious problem that needs to be addressed. There is good
reason to believe that intervention can reduce RA work
disability. The results of 2 studies suggest that treatment
with disease-modifying antirheumatic drugs increases the
number of hours worked (25,26).
In addition, many older persons with RA would benefit
from vocational rehabilitation services, especially when
those services are delivered prior to job loss. Our recently
completed randomized trial proved that a vocational reha-
bilitation intervention delivered to employed subjects
with rheumatic diseases who are at risk for work disability
prevented job loss (27). Job accommodation was a major
component of the intervention, and data from the US
Health and Retirement Survey have shown that job accom-
modation extends the working life of employees with
health conditions, even those who are older (28).
Vocational rehabilitation services are available to per-
sons with health conditions and disabilities in the US
through the public vocational rehabilitation program. This
program is funded largely by the federal government, but
services are delivered in local communities of each state
(www.jan.wvu.edu/SBSES/VOCREHAB.htm). Services can
be provided to persons who are employed, but are at risk
for job loss and/or who are underemployed. The program
is extremely underutilized by persons with rheumatic dis-
eases, partly because most patients and health profession-
als are unaware of it (29).
Many older workers in our sample worked part time,
and the role part-time work plays in maintaining employ-
ment among persons who experience an unplanned event
such as chronic disease is unknown. It may be a way of
managing fatigue; a study in the Netherlands showed that
subjects with RA who worked ?32 hours per week were
more likely to remain employed than those working longer
hours (30). However, in a US study, subjects with RA who
worked ?30 hours per week were more likely to stop
working over a 9-year period than those who worked more
hours, and the authors perceived part-time work as part of
the premature work cessation process (31). Nevertheless,
increased employer flexibility in the number of hours
worked has been recommended as a policy that might
promote the continued employment of workers with
chronic diseases (32). More research is needed to deter-
mine if and when part-time work is beneficial.
The advantage of our clinical sample is accurate diag-
nosis. Because the sample had the typical characteristics
of a survey sample, higher educational attainment and
primarily white race, it is not generalizable to the whole
population of persons with RA. Also, being a clinical
sample, it may not include many people with very mild
RA. NDB participants who filled out only short question-
naires were not included in our sample; examination of
their demographic characteristics showed that they were
the same age, but slightly more likely to be men and have
lower educational attainment. They supplied no func-
tional status data, but their degree of limitation is likely to
be similar to that of the other participants recruited
through pharmaceutical registries who did fill out compre-
hensive questionnaires, i.e., a little greater than that of
participants recruited through rheumatology practices
(mean HAQ score 1.2 versus 1.1). The 422 subjects who
provided no information about their employment status
were older (mean age 70 years), had greater functional
limitation, were more likely to be women, and had lower
educational attainment than subjects in our sample.
In conclusion, we found that ?50% of our national
sample of persons with RA ages 55–64 years and em-
ployed at disease onset had stopped working, and that,
despite having greater educational attainment, their rate of
employment was considerably lower than that of their US
age mates. The financial impact of premature work cessa-
tion on older workers was substantial; 56% of those who
had stopped working versus 32% of employed subjects
had household incomes below the US median income, and
11% had incomes below the poverty level. The work pro-
ductivity of employed older workers, as measured by use
of sick time and limitation in job demands, was at least
equivalent to that of younger workers. We urge that pre-
mature work cessation in persons with RA ages 55–64
Economic Effect of Work Disability on Adults607
years be considered a serious problem, and we recommend
referral for vocational rehabilitation, in addition to access
to the best medical care.
1. Rix SE. The labor market for older workers. Generations 2002;
2. Gabriel SE, Crowson CS, O’Fallon WM. The epidemiology of
rheumatoid arthritis in Rochester, Minnesota, 1955–1985. Ar-
thritis Rheum 1999;42:415–20.
3. Taylor P, Walker A. Employers and older workers: attitudes
and employment practices. Ageing Soc 1998;18:641–58.
4. Ycas M. Patterns of return to work in cohort of disabled-
worker beneficiaries. In: Mashaw JL, Reno V, Burkhauser RV,
Berkowitz M, editors. Disability, work and cash benefits.
Kalamazoo (MI): WE Upjohn Institute for Employment
Research; 1996. p. 169–88.
5. Verstappen SM, Bijlsma JW, Verkleij H, Buskens E, Blaauw
AA, ter Borg EJ, et al. Overview of work disability in rheuma-
toid arthritis patients as observed in cross-sectional and lon-
gitudinal surveys. Arthritis Rheum 2004;51:488–97.
6. Mont D. Comments on chapter 4. In: Rupp K, Stapleton DC,
editors. Growth in disability benefits: explanations and policy
implications. Kalamazoo (MI): WE Upjohn Institute for Em-
ployment Research; 1998. p. 175–6.
7. Mehnert T, Krauss HH, Nadler R, Boyd M. Correlates of life
satisfaction in those with disabling conditions. Rehabil Psy-
8. MacKinnon JR. Occupational profiles: individuals with rheu-
matoid arthritis and a matched comparison sample. Work
9. Fifield J, Reisine ST, Grady K. Work disability and the expe-
rience of pain and depression in rheumatoid arthritis. Soc Sci
10. Allaire SH. Update on work disability in rheumatic diseases.
Curr Opin Rheumatol 2001;13:93–8.
11. Hursh NC. “Pro-work” strategies for older workers with
disabilities: a disability management approach. In: The aging
American workforce: the impact on persons with disabilities.
Switzer Seminar Series. Alexandria (VA): National Rehabili-
tation Association; 2003. p. 22–30.
12. Berkowitz ED. Disabled policy: America’s programs for the
handicapped. New York: Cambridge University Press; 1987.
13. Bureau of Labor Statistics. Labor force statistics from the
Current Population Survey: 2002. URL: http://www.bls.gov/
14. Bureau of the Census. Educational attainment by selected
characteristics: 2000: demographic and economic profiles of
selected racial and Hispanic origin populations. In: Statistical
abstract of the United States: 2002. Tables 36 and 210. Wash-
ington (DC): US GPO; 2002.
15. Lerner D, Amick BC 3rd, Rogers WH, Malspeis S, Bungay K,
Cynn D. The Work Limitations Questionnaire. Med Care
16. Allaire SH. Measures of adult work disability: the Work Lim-
itations Questionnaire (WLQ) and the Rheumatoid Arthritis
Work Instability Scale (RA-WIS). Arthritis Rheum 2003;49
17. Fries JF, Spitz PW, Young DY. The dimensions of health
outcomes: the Health Assessment Questionnaire, disability
and pain scales. J Rheumatol 1982;9:789–93.
18. Bye B, Schecter E. 1978 survey of disability and work: tech-
nical introduction. Washington (DC): US Social Security
Administration; 1982. Publication no. 13-11745.
19. US Social Security Administration. User’s manual: 1978 sur-
vey of disability and work. Washington (DC): US Social Se-
curity Administration; 1981. Publication no. 13-11732.
20. Bureau of the Census. Money income of households: median
income and income level by household type: 2000. Weighted
average poverty thresholds by size of unit: 1980 to 2001.
Persons below poverty level by selected characteristics. In:
Statistical abstract of the United States: 2002. Tables 655, 670,
671. Washington (DC): US GPO; 2002.
21. Walker N, Michaud K, Wolfe F. Work limitations among
working persons with rheumatoid arthritis: results, reliabil-
ity, and validity of the Work Limitations Questionnaire in 836
patients. J Rheumatol 2005;32:1006–12.
22. Munnell AH, Sunden A. Coming up short: the challenge of
401 (k) plans. Washington (DC): Brookings Institution Press;
23. Byrnes N. Pensions that discriminate against older workers.
Bus Week 2003;3864:44–6.
24. Rupp K, Scott CG. Determinants of duration on the disability
rolls and program trends. In: Rupp K, Stapleton DC. Growth
in disability benefits: explanations and policy implications.
Kalamazoo (MI): WE Upjohn Institute for Employment
Research; 1998. p. 139–74.
25. Puolakka K, Kautiainen H, Mottonen T, Hannonen P, Korpela
M, Julkunen H, et al. Impact of initial aggressive drug treat-
ment with a combination of disease-modifying antirheumatic
drugs on the development of work disability in early rheuma-
toid arthritis: a five-year randomized followup trial. Arthritis
26. Yelin E, Trupin L, Katz P, Lubeck D, Rush S, Wanke L.
Association between etanercept use and employment out-
comes among patients with rheumatoid arthritis. Arthritis
27. Allaire SH, Li W, LaValley MP. Reduction of job loss in
rehabilitation: a randomized controlled trial. Arthritis Rheum
28. Chirikos TN. Employer accommodation of older workers with
disabilities: some empirical evidence and policy lesions. In:
Blanck PD, editor. Employment, disability, and the Ameri-
cans with Disability Act. Evanston (IL): Northwestern Univer-
sity Press; 2000. p. 228–57.
29. Allaire SH, Partridge AJ, Andrews HF, Liang MH. Manage-
ment of work disability: resources for vocational rehabilita-
tion. Arthritis Rheum 1993;36:1663–70.
30. Chorus AM, Miedema HS, Wevers CW, van der Linden S.
Work factors and behavioural coping in relation to with-
drawal from the labour force in patients with rheumatoid
arthritis. Ann Rheum Dis 2001;60:1025–32.
31. Reisine S, Fifield J, Walsh SJ, Feinn R. Factors associated with
continued employment among patients with rheumatoid
arthritis: a survival model. J Rheumatol 2001;28:2400–8.
32. Williamson JB, McNamara TK. The effect of unplanned
events on retirement. Presented at: The Fifth Annual Joint
Conference of the Retirement Consortium; May 15-16, 2003;
608Allaire et al