Background

Rheumatoid arthritis (RA) is a chronic, inflammatory joint disease that can lead to joint destruction. RA particularly attacks peripheral joints and joint tissue, gradually resulting in bone erosion, destruction of cartilage, and, ultimately, loss of joint integrity. The prevalence of RA varies globally, ranging from 0.1- 2.0% of the population worldwide [1, 2]. RA significantly reduces functional capacity, quality of life, and results in an increase in sick leave, unemployment, and early retirement [3,4,5]. The loss of productivity due to RA is substantial [2, 5,6,7]. A 2015 American study estimated the cost of over $250 million annually from RA-related absenteeism in United States alone [8].

Research has highlighted the importance of maintaining a connection to the labour market [3, 9], Even a short cessation from work entails a pronounced risk of enduring work exclusion [10]. In Denmark merely 55% on sick leave for 13 weeks succeeded in re-joining the workforce within one year. Among those on sick leave for 26 weeks, only 40% returned to work within the same timeframe [11]. Sustained employment is associated with an improved health-related quality of life [12, 13]. Early and aggressive treatment of RA is crucial for importance in achieving remission and a favourable prognosis reducing the impact of the disease [2, 14,15,16]. Therefore, initiating treatment in a timely manner and supporting patients with RA in maintaining their jobs with inclusive and flexible workplaces if needed is critical [3, 17].

International studies have indicated, that many patients with RA are not employed [18]. In 2020, the average employment rate across Organization for Economic Co-operation and Development (OECD) countries was 69% in the general population (15 to 64 years of age), exhibiting variations among countries, ranging from 46–47% in South Africa and India to 85% in Iceland [19]. Employment rates were lower for individuals with educational levels below upper secondary level compared to those with upper secondary level or higher education [19]. For individuals suffering with chronic diseases, the employment rates tend to be lower. Prognostic determinants for employment in the context of other chronic diseases encompasses the disease’s severity, employment status prior to getting a chronic disease, and baseline educational level [20,21,22]. These somatic and social factors may similarly influence employment status of patients with RA. Several factors, including the type of job (especially physically demanding occupations), support from employers and co-workers, social safety net, and disease factors such as duration and severity, could have an impact on whether patients with RA are employed [17, 23, 24]. Over the years, politicians and social welfare systems have tried to improve the employment rates for patients with chronic diseases. In some countries, rehabilitation clinics have been instrumental in supporting patients to remain in paid work. Healthcare professionals who care for patients with RA occupy a pivotal role in preventing work-related disability and support the patients to remain in work. Consequently, knowledge of the factors that contribute to retention of patients with RA at work is imperative [17, 25].

The aim of this study is therefore to conduct a systematic review, with a primary focus on examining employment rates among patients with RA at the onset of the disease, at study entry, and throughout follow-up. Additionally, this study intends to identify predictors of employment. The predefined predictors, informed by the author’s comprehensive understanding of the field and specific to RA, encompass socioeconomic factors such as age, gender, level of education, employment status prior to the disease, disease stage and duration, treatment modalities, and comorbidities, including depression, which are relevant both to RA and other chronic conditions [26].

Methods

Protocol

This systematic review was carried out according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) for studies that included employment rate in patients with rheumatoid arthritis [27]. PROSPERO registration number: CRD42020189057.

Selection criteria and search strategies

A comprehensive literature search was conducted, covering the period from January 1966 to January 2023 across the PubMed, Embase, and Cochrane Library databases using the following search terms: (Rheumatoid arthritis OR RA) AND (employment OR return to work). Only studies featuring a minimum cohort size of thirty patients and articles in the English language were deemed eligible for inclusion.

The initial screening of articles was based on the titles and abstracts. Studies comprising a working-age population, with current or former employment status, and with no limitations to gender, demographics, or ethnicity were included in this review. Articles addressing topics of employment, work ability or disability, return to work or disability pension were encompassed within the scope of this review. Full-time and part-time employment, but not ‘working as housewives’ was included in this review’s definition of employment. Studies involving other inflammatory diseases than RA were excluded. Reference lists in the selected articles were reviewed, and more articles were included if relevant. A review of the reference lists in the initially selected articles was conducted, with additional articles incorporated if they proved relevant to the research objectives. The eligible study designs encompassed cohort studies, case–control studies, and cross-sectional studies. All other study designs, including reviews, case series/case reports, in vitro studies, qualitative studies, and studies based on health economics were systematically excluded from the review.

Data extraction, quality assessment and risk-of-bias

The data extraction from the selected articles included author names, year of publication, study design, date for data collection, employment rate, study population, age, gender, educational level, ethnicity, disease duration, and pharmacological treatment. To ensure comprehensive evaluation of study quality and potential bias, quality assessment was independently assessed by two reviewers (LK and KB) using the Newcastle–Ottawa Scale (NOS) for cross-sectional and cohort studies [28]. Any disparities in the assessment were resolved by discussion until consensus was reached. For cross-sectional studies the quality assessment included: 1) Selection (maximum 5 points): representativeness of the sample, sample size, non-respondents, ascertainment of the risk factor; 2) Comparability (maximum 2 points); study controls for the most important, and any additional factor; 3) Outcome (maximum 3 points): assessment of outcome, and statistical testing. For cohort studies the assessment included: 1) Selection (maximum 4 points): representativeness of the exposed cohort, selection of the non-exposed cohort, ascertainment of exposure, demonstration that the outcome of interest was not present at start of study; 2) Comparability (maximum 2 points): comparability of cohorts on the basis of the design or analysis; 3) Outcome (maximum 3 points): assessment of outcome, was the follow-up long enough for outcomes to occur, and adequacy of follow up of cohorts. The rating scale was based on 9–10 items dividing the studies into high (7–9/10), moderate (4–6) or low (0–3) quality. A low NOS score (range 0–3) indicated a high risk of bias, and a high NOS score (range 7–9/10) indicated a lower risk of bias.

Analytical approach

For outcomes reported in numerical values or percentages, the odds ratio along with their 95% confidence intervals (CI) were calculated, whenever feasible. Weighted means were calculated, and comparisons between these were conducted using t-test for unpaired data. Furthermore, meta-analysis concerning the pre-determined and potentially pivotal predictors for employment status, both at disease onset, study entry, and follow-up was undertaken. The predictors included age, gender, ethnicity, level of education, duration of disease, treatment, and the presence of comorbities, contingent upon the availability of the adequate data. Additionally, attempts have been made to find information regarding on job categorizations, disease activity (quantified through DAS28; disease activity score for number of swollen joints), and quality of life (SF-36 scores ranging from 0 (worst) to 100 (best)). Age was defined as (< = 50/ > 50 years), gender (male/female), educational level college education or more/no college education), race (Caucasian/not Caucasian), job type (non-manual/manual), comorbidities (not present/present), MTX ever (no/yes), biological treatment ever (no/yes), prednisolone ever (no/yes), disease duration, HAQ score (from 0–3)), joint pain (VAS from 1–10), and DAS28 score. Age, disease duration, HAQ score, VAS score, SF36 and DAS28 were in the studies reported by mean values and standard deviations (SD). Challenges were encountered during attempts to find data which could be used for analysing predictors of employment status before disease onset, and at follow-up, as well as factors related to treatments beyond MTX, prednisolone, and biological as predictors for being employed after disease onset. Test for heterogeneity was done using Chi-squared statistics and I2, where I2 below 40% might not be important; 30–60% may represent moderate heterogeneity; 50–90% substantial heterogeneity; and 75–100% considerable heterogeneity. Meta-analysis for predictors for employment and odds ratio; confidence intervals; and test for heterogeneity were calculated using the software Review Manager (RevMan, version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

Results

General description of included studies

The search yielded a total of 2277 references addressing RA its association with employment. Following the initial title screen, 199 studies were considered relevant for further evaluation. Of those, 91 studies ultimately met the inclusion criteria. Figure 1 shows the results of the systematic search strategy.

Fig. 1
figure 1

Flow chart illustrating the systematic search for studies examining employment outcome in patients with rheumatoid arthritis

Full size image

Table 1 summarizes the general characteristics of the included studies. The publication year of the included studies ranged from 1971 to 2022. Among the studies, 60 (66%) adopted a cross-sectional research design [13, 18, 29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88, 129] with a total of 41,857 participants analysing data at a specific point in time. Concurrently, 31 studies (34%) adopted a cohort design [89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122, 130] with a total of 59,974 participants. Most of these studies exhibited a small to moderate sample size, with a median of 652 participants. Additionally, single centre studies and studies from high-income countries were predominant. Study details are shown in Table 1.

Table 1 Characteristics of the individual studies among patients with rheumatoid arthritis
Full size table

General description of study participants

On average, patients with RA were 51 years old, with an age range spanning from 42 to 64 years. Furthermore, the female population accounted for 75.9% of the patient cohort, with a range from 41 to 92%. The duration of the disease at study entry exhibited significant variability, ranging from less than one year up to more than 18 years on average.

Employment rate

At disease onset, the employment rate was 78.8% (weighted mean, range 45.4–100), at study entry 47.0% (range 18.5–100), and during the follow-up period 40.0% (range 4–88.2), as shown in Table 2. Notably, a comparative analysis of the employment rates between Europe and North America indicated no substantial difference (p = 0.93). However, the comparison between Europe, North America and ‘other continents’ did yield significant differences (or nearly differences) with p-values of 0.003 and 0.08, respectively.

Table 2 Employment rate in patients with RA, by continent (weighted mean, SD, range)
Full size table

The employment rate exhibited no change, when comparing studies from the 1980s through to 2022. Specifically, the weighted mean for the years 1981–2000 was 49.2%, aligning closely with the corresponding figures for the years 2001–2010 (49.2%) and 2011–2022 43.6%. These findings were statistically non-significant, with p-values of 0.80 for comparison between year 1981–2000 and 2001–2010; 0.66 for 2001–2010 and 2011–2022, and 0.94 for 1981–2000 and 2011–2022, shown in Figure S1, see Additional file.

Among the studies included in the analysis, nineteen studies included data of employment at follow-up, with durations ranging from 1 to 20 years, Table 2. For instance, Jäntti, 1999 [97] reported an employment rate 69% one year after disease onset, which gradually declined to 50% after 15 years and further to 20% after 20 years. Similarly, Mäkisara, 1982 [63] demonstrated that 60% of the patients were employed 5 years after disease onset, 50% after 10 years, and 33% after 15 years. Nikiphorou, 2012 [101] reported an employment rate of 67% at study entry, which decreased to 43% after 10 years.

In addition, seven studies included data of employment rate among patients comparing different medical treatments [18, 44, 56, 91, 105, 110, 119]. These studies indicated that, on average, 55.0% (weighted mean) of the patients were employed after receiving treatment with MTX, while 42.8% after undergoing treatment with a combination of MTX + Adalimumab (all patients were employed at disease onset in these specific studies).

Predictors for employment

Information of normative comparison data to use for meta-analysis of predictors for employment at study entry was available for age, gender, educational level, race, job type, comorbidities, MTX at any time, biological treatment at any time, prednisolone at any time, disease duration, HAQ score, joint pain (VAS-score), and disease activity (DAS28 score). Predictors for employment at study entry was being younger /age below 50 years, being a male, higher educational level (college or more), non-manual work, having no comorbidities, no medical treatment, short disease duration, and low HAQ score, VAS-score, or DAS28 score. Heterogeneity was small for age, gender, medical treatment, and moderate for educational level, and job type as indicted by the I2 values, Table 3, and shown in detail in Figures S2, S3, S4, S5, S6, S7, S8, S9, S10, S11, S12, S13, S14, S15 and S16, see Additional file.

Table 3 Predictors for employment among patients with RA
Full size table

Assessment of quality of included studies

All studies were subject to rigorous quality assessment. These assessments resulted in categorisation of either medium quality (n = 64; 70%) or high-quality studies (n = 27; 30%), with no studies falling into the low-quality category. The quality assessment is shown in Tables 4 and 5.

Table 4 Assessment of quality of the included cohort studies. NOS heat map [89,90,91,92,93,94,95,96,97,98,99,100,101,102,103, 106,107,108,109,110,111,112,113,114,115,116,117, 119, 121, 122, 130]
Full size table
Table 5 Assessment of quality of the included cross-sectional studies. NOS heat map [13, 18, 29,30,31,32,33, 35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60, 62,63,64,65,66,67,68, 70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87, 129]
Full size table

Notably, many studies were characterised by several common attributes, including cross-sectional study design, single-centre-settings, relatively small sample sizes, and the reliance on self-reported patient data. When including only the high-quality studies in the analyses, the employment rates at study entry changed from 47% (weighted mean, all studies) to 50% (weighted mean, high quality studies).

Discussion

Key findings

This systematic review has identified a decline in the employment rate among patients with RA, with a notable decrease from disease onset during the study entry to follow-up, where only half of the patients were employed. These findings corroborate earlier research that indicated a substantial decline in employment rates among patients with RA over time. Notably, previous studies have reported that approximately one third of patients with RA stopped working within 2 to 3 years after disease onset, and more than half was unable to work after 10 to 15 years [23, 63, 93, 97, 101]. Only few studies have included data from the general population, comparing the employment rates with the rates for patients with RA [89, 90]. Comparisons with the general population further underscored the challenges faced by RA patients, as their employment rates were consistently lower.

Despite changes in medical treatment, social security systems, and societal norms over the past decades, there was no significant improvement in the employment for patients with RA. This pattern aligns with data from the Global Burden of Disease studies, highlighting the persistent need for novel approaches and dedicated efforts to support patients with RA in sustaining employment [2, 123]. Recent recommendations from EULAR (European Alliance of Associations for Rheumatology) and ACR (American College of Rheumatology) have emphasized the importance of enabling individuals with rheumatic and musculoskeletal diseases to engage in healthy and sustainable work [17, 124, 125].

While different countries possess different social laws and health care systems for supporting patients with chronic diseases, the variations in the weighted mean of employment rates across countries were relatively minor.

In the meta-analysis, one of the strongest predictors for maintaining employment was younger age at disease onset [43, 51, 101, 116]. Verstappen, 2004 found that older patients with RA had an increased risk of becoming work disabled, potentially caused by the cumulative effects of long-standing RA, joint damage, and diminished coping mechanisms, compared to younger patients [23].

More women than men develop RA, however this study showed that a higher proportion of men managed to remain employed compared to women [18, 36, 42, 43, 46, 62, 71, 89, 101, 116]. Previous studies have shown inconsistent results in this regard. Eberhart, 2007 found that a significantly higher number of men with RA worked even though there was no difference in any disease state between the sexes [93]. De Roos,1999 showed that work-disabled women were less likely to be well-educated and more likely to be in a nonprofessional occupation than working women. Interestingly, there was no association of these variables among men. Type of work and disease activity may influence work capacity more in women than in men [46]. Sokka, 2010 demonstrated a lower DAS28 and HAQ-score in men compared to women among the still working patients with RA, which indicated that women continued working at higher disability and disease activity levels compared with men [18].

Disease duration also played a significant role as a predictor of employment outcomes [33, 36, 45, 71, 77, 86, 102, 111]. Longer disease duration correlate with decreased employment likelihood, which could be attributed to older age and increased joint damage and disability in patients with longer-standing RA.

Higher educational levels were associated with a greater possibility of employment [30, 43, 45, 46, 51, 62, 86]. This is probably due to enhanced job opportunities, flexibility, lower physical workload, better insurance coverage, and improved health care for well-educated individuals. This is further supported by the fact that having a manual work was a predictor for not being employed [30, 39, 43,44,45].

Furthermore, health-related quality of life, as measured by SF 36, lower disease activity (DAS28 scores), reduced joint pain (VAS-score), and lower disability (HAQ score) were additionally predictors for being employed [33, 35, 36, 45, 71, 86]. This support the statement that the fewer symptoms from RA, the greater the possibility of being able to work.

The results showed that the presence of comorbidity was a predictor for not being employed, aligning with findings from previous studies that chronic diseases such as cardiovascular disease, lung disease, diabetes, cancer, and depression reduced the chances of being employed [126]. Moreover, the risk of exiting paid work increased with multimorbidity [127].

While limited data were available for assessing the impact of treatment on employment, indications suggested that patients with RA were receiving medical treatments, such as MTX or biological medicine, were more likely to be unemployed. One possible explanation for this phenomenon could be that patients with RA, who were receiving medical treatment, had a more severe and a longer duration of RA compared to those, who had never been on medical treatment. However, the scarcity of relevant studies necessitates caution when drawing definitive conclusions in this regard.

Therefore, the predictors for employment found in this review were being younger, being a male, having higher education, low disease activity, low disease duration, and being without comorbidities. This is supported by previous studies [93, 116]

In summary, this review underscores the importance of managing disease activity, offering early support to patients upon diagnosis, and reducing physically demanding work to maintain employment among patients with RA. Achieving success in this endeavour requires close cooperation among healthcare professionals, rehabilitation institutions, companies, and employers. Furthermore, it is important that these efforts are underpinned by robust social policies that ensure favourable working conditions and provide financial support for individuals with physical disabilities, enabling them to remain active in the labour market.

Strengths and limitations

The strength of this review and meta-analysis lies in the inclusion of a large number of articles originating from various countries. Furthermore, the data showed a consistent employment rate in high quality studies compared to all studies. However, there are some limitations to this review. No librarian was used to define search terms and only three databases were searched. Furthermore, the initial search, selection of articles, data extraction, and analysis was undertaken only by one author, potentially leading to the omission of relevant literature and data. The review also extended back to 1966, with some articles from the 1970s and 1980s included. Given the significant changes in medical treatment, social security systems, and society over the past decades, the generalizability of the findings may be limited.

Moreover, the majority of studies did not include a control group from the general population, which limited the ability to compare employment rates with the general population in the respective countries. Many studies were cross-sectional in design, which limits the evidence of causality between employment rate and having RA. However, the employment rate was approximately the same in high quality studies compared to all studies, which supports an association. A substantial number of studies relied on self-reported employment rates, introducing the potential for recall bias. Additionally, many studies did not account for all relevant risk factors for unemployment failing to control for all relevant confounders.

EULAR have made recommendation for point to consider when designing, analysing, and reporting of studies with work participation as an outcome domain in patients with inflammatory arthritis. These recommendations include study design, study duration, and the choice of work participation outcome domains (e.g., job type, social security system) and measurement instruments, the power to detect meaningful effects, interdependence among different work participation outcome domains (e.g., between absenteeism and presentism), the populations included in the analysis of each work participation outcome domain and relevant characteristics should be described. In longitudinal studies work-status should be regularly assessed and changes reported, and both aggregated results and proportions of predefined meaningful categories should be considered [128]. Only some of the studies in this review met the requirements for high quality studies. In both older and newer studies methodological deficiencies persisted in study design, analysis, and reporting of results, as recommended by EULAR.

Perspectives for future studies

Future research in this area should focus on developing and evaluating new strategies to address the ongoing challenges faced by patients with RA in maintaining employment. Despite many initiatives over the years, there has been no success in increasing employment rates for patients with RA in many countries. Therefore, there is a pressing need for controlled studies that investigated the effectiveness of interventions such as education, social support, and workplace adaptations in improving employment outcomes for these individuals.

Conclusion

This systematic review underscores the low employment rate among patients with RA. Key predictors of sustained employment include being younger, having higher educational level, short disease duration, and lower disease activity, along with fewer comorbidities. Importantly, the review reveals that the employment rate has not changed significantly across different time periods. To support patients with RA in maintaining their employment, a comprehensive approach that combines early clinical treatment with social support is crucial. This approach can play a pivotal role in helping patients with RA stay connected to the labour market.