Objective To estimate the sick leave and disability pension trajectory in patients diagnosed with early rheumatoid arthritis (RA) 1999–2007, and in prevalent patients in 2007.
Methods Individuals aged 19–59 years diagnosed with early RA were identified in the Swedish Rheumatology Quality Register (1999–2007; n=3029; 47 years; 73% women). Additionally, prevalent patients in 2007 were identified in the National Patient Register (n=25 922; 52 years; 73% women). For each patient, five age-, sex-, education- and county-matched general population comparators were sampled. Sick leave and disability pension days were retrieved from national registers.
Results Sick leave and disability pension increased from a mean 43 to 77 days/year from 2 to 1 years before RA diagnosis. A further increase to 147 days/year was observed the next year, followed by a rebound to 116 days/year 4 years after diagnosis. During the 4 years following diagnosis, sick leave decreased from a mean 118 to 35 and disability pension increased from 29 to 81 days/year. In the prevalent RA population, patients had a mean 158 annual days of sick leave and disability pension compared to 71 in comparators. Large variations existed across age, sex and education level, but RA patients had consistently higher levels. In 2007, the costs associated with sick leave and disability pension were €16 000 per patient with €9 000 attributable to RA.
Conclusion Despite better drugs and improved treatment strategies, data from contemporary patients with early and established RA continue to indicate large unmet needs.
Statistics from Altmetric.com
Data from the 1990s show that work ability dramatically deteriorates from the time of rheumatoid arthritis (RA) diagnosis, but that it does not continue to decline beyond that expected in the general population thereafter.1,–,4 Since then, marked changes in the therapeutic approaches to early RA have occurred. The increasing treatment intensity and number of treatment options introduced the last decade may have led to improvements in work ability after diagnosis. Furthermore, nationwide assessments of the societal burden due to sick leave and disability pension in the prevalent RA population are lacking.
Regarding established RA, studies show substantial productivity losses.5 6 In a regional Swedish survey from 2002, productivity losses were reported to constitute >50% of the total cost of RA.7 Although estimates are generally regarded to be country-specific due to welfare system differences, a recent review found similar rates of work disability in US and northern European studies of RA.5
The aims of this study were: 1. to estimate the work ability trajectory, as measured by annual sick leave and disability pension days, and its temporal relation to diagnosis in patients diagnosed with early RA 1999–2007, and 2. to estimate the nationwide annual productivity losses in the prevalent RA population in 2007. Days on sick leave and disability pension, and costs, attributable to RA were assessed through comparison to matched general population comparators.
Patients aged 19–59 years diagnosed with RA between 1999 and 2007 were identified at diagnosis in the Swedish Rheumatology Quality Register (SRQ; n=3029; mean 47 years; 73% women; denoted early RA cohort) where diagnosis is set by rheumatologists.8 Additionally, prevalent RA patients aged 19–64 years in 2007 were identified by also using the National Patient Register, as described elsewhere (n=25 922; mean 52 years; 73% women; denoted prevalent RA cohort).9
Based on each resident's unique personal identification number, the early RA and prevalent RA cohorts were merged with the Social Insurance Office database including sick leave and disability pension data (1996–2010), and the Longitudinal Integration Database for Health Insurance and Labour Market Studies (LISA; 1996–2007). LISA includes variables such as education level, annual days of sick leave and disability pension (derived from the Social Insurance Office day-level data) and days with unemployment benefits. Ethical approval was granted by the Regional Ethics Committee, Karolinska Institutet, Sweden.
In 2007, Sweden had a population of 9.2 million with 5 517 813 (60%) being 19–64 years of age. The Swedish healthcare system is tax-funded and offers universal access, and prescription drugs are provided for free above a threshold of 1800 SEK cumulative expenses annually (≈€200). Patients with RA are typically diagnosed and treated by rheumatologists rather than general practitioners, and their care is provided in outpatient and inpatient facilities with the majority of rheumatologists working at hospitals rather than as private practitioners. The penetration of biologics in RA has been estimated to be 15–18% overall, and 22–27% in working-age patients.9
The Swedish welfare system provides compensation for sick leave, disability pension and unemployment, all of which may be complete (100%) or partial (eg, 25%, 33%, 50%, 67%, 75%). Data are recorded by the Social Insurance Office. The retirement age is 65 years, but one can continue working until 67 years or go into old-age retirement at 61 years. In this study, working age was defined as 19–64 years.
Data were retrieved from the Population and Housing Census database at Statistics Sweden to ascertain place of residence, vital and emigration status. From the same register, five general population comparators were sampled for each early and prevalent RA patient, using age (±1 year), sex, county and education level (≤9, 10–12 and >12 years of schooling) as matching factors.
The primary outcome was net annual days of sick leave and disability pension (maximum 365). As a secondary outcome in the prevalent RA cohort, days on unemployment benefits were evaluated.
During the study period, the first sick leave day of an episode was unpaid (‘waiting period’). Compensation for days 2–14 was paid by the employer (‘sick pay period’; 28 days between 1997 and 31 March 1998), and only episodes >14 days were recorded by the Social Insurance Office. After day 7 a doctor's note confirming the medical problem was required.
Disability pension (in 2003 replaced by ‘sickness compensation’ (‘sjukersättning’) for 30–64-year-olds, and ‘activity compensation’ (‘aktivitetsersättning’) for 19–29-year-olds) was defined as either sickness compensation or activity compensation (2003–2010), or disability pension (<2003). Both sickness and activity compensation may be time-limited or permanent, and require at least 25% reduction in work ability expected to remain for ≥1 year.
Individuals on full disability pension were not entitled to unemployment benefits, whereas those on partial disability pension generally were. Furthermore, individuals on unemployment benefits may receive sick leave. Therefore, days on unemployment are presented separately from sick leave and disability pension to prevent double counting.
Statistical analyses were performed using SAS statistical software (version 9.2; SAS Institute, Cary, North Carolina, USA). Due to the non-normal distribution of days, the percentages of patients in categories of days of sick leave and disability pension were calculated. Despite the skewed distribution, the mean was used as summary measure, as the mean multiplied by the number of patients represents the overall burden to society. The early RA cohort and their matched comparators were followed from 4 years before date of diagnosis to 4 years after regarding annual days of sick leave and disability pension. Also, a closer analysis by month from 12 months before to 12 months after diagnosis was conducted.
For the prevalent RA cohort and their matched comparators, total annual days of sick leave and disability pension, as well as unemployment, were calculated for 2007. The associated cost of sick leave and disability pension was calculated using the human capital method. Each day was valued using the mean sex-specific wage (women 23 500, men 28 000 SEK/month; http://www.scb.se) and the applicable social fees in 2007 (32.42%; http://www.skatteverket.se). All costs are presented in 2007 SEK and € (converted using purchasing power parities for 2007; http://www.oecd.org).
A p value of <0.05 was regarded as statistically significant.
In the early RA cohort, the clinical and demographic characteristics of newly diagnosed patients were similar over time, although the percentage initiating biologics within 1 year of diagnosis increased until 2003–2004 (table 1).
Monthly sick leave and disability pension in relation to RA diagnosis
One year before diagnosis, the monthly number of days on sick leave and disability pension was similar in RA patients and matched general population comparators (figure 1). However, in RA patients the mean monthly days increased steadily during the 12 months prior to diagnosis and peaked in the month after diagnosis at 15 days (median 15). Thereafter it decreased and stabilised at around 11 days per month. Six months after diagnosis the median monthly work loss was back at 0 days.
Annual sick leave and disability pension in relation to RA diagnosis
In the early RA cohort patients with 4 years of follow-up after diagnosis (diagnosis 1999–2006), the trajectory of mean annual days of sick leave and disability pension was similar to the general population's until 1 year prior to diagnosis, with small annual increases (figure 2). The mean annual number of days on sick leave and disability pension increased from 43 to 77 from 2 to 1 year prior to RA diagnosis and peaked at 147 days the year after diagnosis, before rebounding to a level about twice as high as in general population comparators.
The percentage of patients with RA with no sick leave episodes >14 days or any disability pension decreased from 74% 2 years before to 36% 1 year before to 30% 1 year after RA diagnosis, but thereafter increased again (figure 3, upper panel). In parallel, the percentage of patients with 365 days on sick leave or disability pension increased from 6% 2 years before to 19% 1 year after diagnosis.
After the peak 1 year after RA diagnosis, there was a decline in sick leave days, which was almost fully offset by an increase in disability pension (figure 3, lower panel). Disability pension days accounted for 31% of total days 1 year prior to diagnosis, but 70% 4 years after.
Calendar period trends in sick leave and disability pension in relation to RA diagnosis
The level of sick leave and disability pension days 1 year prior to diagnosis was similar between patients diagnosed in 1999, 2003 and 2007, whereas the level 1 year after diagnosis decreased over time (figure 4). The increase from 1 year prior to diagnosis to the year after diagnosis was greater in 1999 (89 days, 95% CI 73 to 105) and 2003 (75 days, 63 to 87) compared to 2007 (42 days, 32 to 51). However, the trajectory in the general population comparators matched to RA patients diagnosed in 2007 was also different: whereas comparators matched to patients diagnosed in 1999 and 2003 displayed a steady increase in sick leave and disability pension over time, the comparators for the 2007 cohort remained at the same level during follow-up.
Sick leave, disability pension and unemployment in prevalent RA patients
In the prevalent RA cohort, 46% were on disability pension in 2007 compared to 19% of the general population comparators (25% vs 12% full-time). The corresponding percentages for sick leave were 26% versus 15%, and for any unemployment benefits 9% versus 11%.
In prevalent RA patients and general population comparators, marked heterogeneity across demographic segments was seen, with more days of sick leave and disability pension per year in women compared to men, at older compared to younger age and in those with a lower education level (table 2). However, in each segment the mean was higher in RA patients, ranging from a difference of +30 days in the youngest to +116 days per year in the oldest age group. There was substantial heterogeneity of these differences across levels of education, with greater differences within the RA population than between the RA and the general population comparator.
Overall, and in both sexes, days on disability pension constituted approximately 80% of the total annual days in patients with RA and in the general population comparators, although the proportion varied by age and education (table 2).
In contrast to sick leave and disability pension, mean days of unemployment were lower in patients with RA compared to the general population, although this was not consistent across subgroups and the differences were small (table 2). As expected, the age-gradient was opposite to that seen for sick leave and disability pension, with the highest unemployment in the youngest age group.
Economic costs associated with sick leave and disability pension
Excluding sick leave episodes ≤14 days, the estimated costs from sick leave and disability pension in the Swedish RA population were 4.2 billion SEK (€391 million; 4.0 million days) in 2007. This can be compared with an estimated 1.9 billion SEK (€177 million; 1.8 million days) in a matched general population cohort of the same size, resulting in an annual cost attributable to RA of 2.3 billion SEK (€215 million). For the average patient with RA, the cost estimate in 2007 was 168 000 SEK (€16 000), or a cost attributable to RA of 92 000 SEK (€9000) compared to a general population comparator.
In patients diagnosed with early RA 1999–2007, a dramatic and persistent increase in days of sick leave and disability pension was found during the year preceding RA diagnosis. On a national level in prevalent RA patients, this resulted in an average 87 more days on sick leave and disability pension per patient compared to the general population in 2007, corresponding to an annual cost attributable to RA of €9000 per working-age patient. Thus, despite earlier and more aggressive use of a growing number of antirheumatic therapeutic options, the rapid and often irreversible work loss around diagnosis remains, as does a large unmet need for patients with established RA.
Several small studies of recent onset RA from the 1990s have investigated the trajectory of work ability after,2 3 10 or before and after RA diagnosis.1 Some focused only on disability pension and reported an increasing prevalence with RA duration.3 10 The current study and others have found disability pension to increase sharply in parallel to a decrease in sick leave after RA diagnosis, indicating the importance to investigate both outcomes jointly.1 4 Björk et al analysed the outcomes jointly from 3 years before to 3 years after RA diagnosis in 120 patients with recent onset RA (≤1 year) identified in 1996–1998 (before the introduction of biologics). Largely congruent with the present findings, they observed no differences compared to the general population until 0.5 years prior to diagnosis, but large and persistent differences thereafter.1 The results from the Finnish Rheumatoid Arthritis Combination Therapy trial of 162 patients with recent onset RA (<2 years; median 0.5 years) recruited in 1993–1995 and receiving either mono or combination disease-modifying antirheumatic drug (DMARD) therapy, are also largely congruent with the present findings.4 Puolakka et al found sick leave and disability pension peaking during the first year of follow-up and thereafter remaining stable over the subsequent 4 years, with small annual increases. As in the current study, sick leave decreased sharply over time, but increases in disability pension almost fully reciprocated this decrease.4
The burden of sick leave and disability pension in the prevalent RA population in 2007 was also estimated. The finding that 46% (25% full-time) of RA patients were on disability pension is within the 19–52% range reported in a systematic review.5 Regarding sick leave days, the same review reported a mean duration of 39 days per year (range 7–84).5 The sick leave estimate of 31 days in the current study agrees well with this finding, whereas the addition of the mean annual 126 days on disability pension results in a substantially higher estimate of total days.
The relation of days on disability pension versus sick leave in established RA in the current study (80% vs 20%) is similar to a Swedish community-based survey from 2002 of 613 patients (74% vs 26%),7 and highlights the importance of jointly analysing the outcomes.11 In that study, the cost per patient associated with sick leave and disability pension was 59 000 SEK versus 168 000 SEK from the current study. This may be explained by the authors of the community-based survey using all patients irrespective of age (mean age 66 years) in the denominator, as well as using a questionnaire asking respondents for productivity losses attributed specifically to RA (L Jacobsson, personal communication). To attribute losses to RA, the current study used matched general population comparators instead, and collected information on workforce participation from an external and independent source. This resulted in an estimated annual productivity loss attributable to RA of 92 000 SEK (€9 000) per working-age patient.
The large differences in sick leave and disability pension between the RA and general populations reveal a substantial unmet need already present at the time of RA diagnosis. This indicates also that effective treatment has the potential to prevent major economic losses for society. Improvements in productivity appear to hold the greatest promise to counter some of the rapidly increasing cost burden from biologics, which in 2009 constituted 5% of the total drug expenditure in Sweden.12 13 In this regard, it is disappointing that the current study did not find any clear difference for the RA population as a whole regarding the work ability trajectory in contemporary patients compared to estimates from the 1990s.1,–,4 Similar to US data on Health Assessment Questionnaire and quality of life, the greatest deterioration was found occurring around RA diagnosis.14 The current study did, however, find less work loss after (but not before) diagnosis in patients diagnosed in 2007 compared to 1999 and 2003, along with an increase from 5% in 1999 to 15% in 2007 of patients on biologics 1 year after diagnosis. The interpretation of the finding for the 2007 cohort is complicated by policy changes regarding sick leave and disability pension around this time and the parallel finding that matched comparators to the 2007 cohort did not experience any increases with age, as seen for the other comparator cohorts. Further investigation of the work ability trajectory specifically in patients treated with biologics and non-biologic DMARDs is needed, as a complement to studies using cross-sectional data,15 studies lacking comparator groups16 and studies not analysing sick leave and disability pension jointly.13 17
A strength of this study was the population-based design. Access to general population comparators matched by important confounders further enabled estimation of the burden attributable to RA. Using registered rather than self-reported data on productivity losses eliminated recall bias.18 19 The authors are unaware of any previous study reporting nationwide productivity data for patients with RA, and, although some studies have used general population comparators to estimate costs attributable to RA,1 a review noted that few use this methodology.20 Instead, full productivity is assumed as benchmark or questionnaires ask for work loss that the patients themselves attribute to RA.
This study also had limitations. First, Sweden has a generous welfare safety net. Although this affects RA patients and comparators, it may be differential and limit generalisability to other countries. A systematic review of productivity losses in RA recently concluded that ‘rates of work disability were similar in the USA and in Northern European countries, despite differences in social systems and study methodologies’.5 It is unclear whether this would hold had the studies been fully comparable, as considerable subgroup heterogeneity was found in the current study with, for example, larger differences across educational levels than between RA and the general population.
Second, the data source used does not include sick leave episodes ≤14 days, resulting in underestimation. In Finland, Puolakka et al4 reported 3% of sick leave episodes being <10 days in patients with RA, constituting 0.2% of total days, indicating that underestimation is likely to be small. Furthermore, some additional underestimation is likely to have taken place as some workers enter into old-age retirement prior to age 65 years. Third, although SRQ is a nationwide register and ARTIS, its subregister for biologics-treated patients, covers an estimated 87% of all target patients,21 there are currently no estimates of the coverage and generalisability of SRQ. Fourth, this study did not include premature death or presenteeism as sources of productivity losses associated with RA.5 22 23 Adding these elements would increase the indirect cost estimate further. Finally, no data were available on smoking or obesity status, two important and independent predictors of productivity losses due to disability pension and premature death.24 25
In conclusion, work ability was found to deteriorate dramatically around RA diagnosis but stabilised shortly thereafter in the 1999–2007 period, a period including the introduction of new antirheumatic drugs and therapeutic strategies. The current findings highlight a large unmet medical need remaining in the RA population today, and further need to improve treatment.
The ARTIS Study Group E Baecklund (Uppsala University), L Cöster (Linköping University), C Dackhammar (Sahlgrenska Academy), N Feltelius (Medical Products Agency), P Geborek (Lund University), L Jacobsson (Lund University), L Klareskog (Karolinska Institutet), S Lindblad (Karolinska Institutet), S Rantapaa-Dahlqvist (Umeå University), T Saxne (Lund University) and R van Vollenhoven (Karolinska Institutet).
Competing interests The ARTIS Study Group conducts scientific analyses using data from the Swedish Biologics Register ARTIS run by the Swedish Society for Rheumatology. For the maintenance of this register, the Swedish Society for Rheumatology has received funding, independent of the conduct of these scientific analyses, from Schering-Plough, BMS, Wyeth, Abbott Laboratories and Roche.
Ethics approval This study was conducted with the approval of the Regional Ethics Committee, Karolinska Institutet, Stockholm, Sweden.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.