Rheumatoid arthritis (RA) is a chronic, disabling autoimmune disorder that affects approximately 1% of American adults.1 There has been substantial progress in our understanding of RA disease progression and in the development of therapies for its treatment in recent years. Recommendations by the American College of Rheumatology and other organizations support early aggressive treatment with diseasemodifying antirheumatic drugs (DMARDs), including biologic therapies.2 Current guidelines recommend beginning DMARD therapy with nonbiologic agents, especially methotrexate, alone or in combination with other nonbiologic agents. Addition of a biologic DMARD may be indicated in patients whose disease does not respond to maximally tolerated doses of methotrexate or other nonbiologic DMARD therapy.2
RA is a costly disease, with the fourth highest direct medical expenditures per employee among 11 chronicconditions.3 In addition to its clinical burden and direct cost impact, RA has a significant adverse effect on the patient's ability to work, absenteeism, and productivity, because of the debilitating nature of the disease, and because most patients develop the disease between age 35 and 50 years, when most employed people are in their prime work years.4 Burden of illness studies have shown that overall, indirect costs may be greater than direct costs for RA.5
Factors associated with being unable to work among patients with RA include pain, impaired physical function, difficulty commuting, and failure to achieve significant improvement with antirheumatic drug therapy.6 Employment status is not simply a function of whether the patient is working but also of the patient's productivity while at work.
Biologic medications such as tumor necrosis factor (TNF)-alpha antagonists are frequently prescribed for treatment of moderate-to-severe RA.7 TNF-alpha antagonists inhibit the activity of the inflammatory mediator TNF-alpha. Etanercept (Enbrel) is a selfinjected TNF-alpha inhibitor approved for the treatment of moderate-to-severe RA. It has been shown to significantly improve clinical outcomes and functional status in patients with RA when added to multidrug nonbiologic DMARD regimens that have not succeeded in fully controlling the disease.8,9
From an economic perspective, the use of TNF-alpha inhibitors continues to increase.10 In 2008, etanercept was ranked the seventh top-selling pharmaceutical drug in dollar volume in the United States.11
The purpose of the Work Loss and Productivity (WLP) survey was to quantify the impact of etanercept therapy among employees with RA on both work loss and lost productivity while at work. Knowing this may enable employers to better understand the direct and indirect economic impact of TNF-alpha inhibitors such as etanercept in employees with moderate-to-severe RA.
Materials and Methods
Participants in the WLP survey were recruited from the participant pool of a previously conducted study—Rheumatoid Arthritis Payment (RAP)—that assessed willingness and ability to pay for etanercept among patients with moderate-to-severe RA at various income levels. The RAP study was conducted from April through November 2006 and included 217 patients aged ≥18 years with moderate-to-severe RA from 15 community-based rheumatology practices across the United States who had been taking etanercept for 3 to 24 months. Participants who were employed at the time of the RAP interviews (n = 118; 93 full-time, 25 part-time) were asked to participate in the WLP survey.
The WLP survey was conducted from December 2006 to January 2007; participants received $25 for participating in the study. This study was a retrospective analysis of the impact of etanercept treatment on work loss and productivity among employed patients with RA. Eligibility criteria for study participation included age ≥18 years, having a diagnosis of moderateto- severe RA, and being employed part-time (<40 weekly hours) or full-time (≥40 weekly hours) before and since starting etanercept therapy. Because the WLP study took place shortly after the RAP study, all WLP participants had been taking etanercept for at least 5 months, but not more than 28 months.
We limited the analyses to patients who were working before and after taking etanercept to be able to compare, during a telephone survey, each patient's work loss and productivity during that period. Of the 118 patients recruited for the study, 17 (14.4%) did not meet the eligibility criteria. These patients either stopped taking etanercept since the RAP study (n = 9), were not working before starting etanercept (n = 7), or were no longer working since starting etanercept (n = 1). Of the remaining 101 eligible patients, 3 were unwilling to participate, 3 did not return phone calls, and 9 were treated at sites that chose not to participate in the WLP study (total, n = 15; 14.9%); 86 eligible patients (85.1%) participated in the WLP study.
The WLP study protocol was approved by a centralized institutional review board, and all patients provided informed consent to participate in the study.
Survey Questionnaire and Administration
A standardized, telephone-administered survey was developed for this study to collect participant information. Patients were asked to rate their typical work loss and productivity experience before initiation of treatment with etanercept and at the time of the interview. The final questionnaire consisted of 13 items (see Appendix).
Self-reported productivity was measured on a 1-to-10 scale, in which lower numbers represented less impact and higher numbers represented more impact of RA on productivity. This scale was used to estimate the impact of productivity changes for full-time and part-time employees utilizing several different assumptions.
Univariate analysis was used to describe the characteristics of the study cohort, including age, sex, education, duration of RA, length of time taking etanercept, and type of employment. One-sample, paired t-tests were used to examine patient changes for all measures before and since taking etanercept, including weekly work hours, missed work days in a month, number of times patients missed ≥5 consecutive work days because of RA, number of days per month RA limited work, and the effect of RA on productivity. McNemar's tests were used to investigate the relationship between the percentage of patients who reported positive responses before and since taking etanercept for any days missed from work, ≥5 missed days from work, RA limiting the kind of work they could perform, or RA limiting the type of work for more than one half of the work days in a month.
To assess the validity of our findings, we compared results for the subset of 19 patients most recently switched to etanercept (5-10 months before the survey)—because these patients would be least prone to recall bias—with the results for the rest of the cohort. We selected the timeframe of 5 to 10 months to ensure a large enough sample size for statistical inference, while keeping the timeframe short enough to minimize potential recall bias. In addition, response to etanercept treatment can take 12 to 24 weeks, so the lower criterion was also based on providing sufficient time for the impact of treatment to be recognized by the patient.
We performed one-sample, paired t-tests and McNemar's tests on these subsets of patients. We also compared the results for patients with different employment status (full-time vs part-time), and for patients who worked in 2 different broad job categories (sedentary vs requiring manual labor or mobility).12 A P value of <.05 was considered significant. All analyses were conducted with SPSS software, 14.0 for Windows (SPSS Inc; Chicago, IL).
Calculation of Economic Impact
To derive the economic impact of etanercept treatment, we estimated the financial value of changes in employee absenteeism and productivity. For absenteeism, changes in work hours were multiplied by average salary wage estimates, based on patient-reported job categories, to determine the dollar value of the changes in hours worked. For productivity, several assumptions were used to estimate the economic impact of patient-reported productivity changes for full-time and part-time employees.
First, we assumed a 1:1 linear relationship between self-reported productivity and output, such that every 1-point reduction in productivity on the 1-to-10 scale was assumed to be equivalent to an 11.1% reduction in output. Under these assumptions, working 40 hours in a week at a productivity level of 2 (or a 1-point reduction in productivity) was equivalent to working 35.6 hours at full productivity.
To take a more conservative approach, we then assumed a 2:1 ratio between self-reported productivity and output. In this scenario, each 1-point reduction in productivity was the equivalent of a 5.5% reduction in output. Based on this 2:1 relationship, working 40 hours weekly at a self-reported productivity level of 2 was equivalent to working 37.8 hours at full productivity. These same percentage changes were used to estimate productivity changes for part-time employees. We used an annual cost of $16,389 for etanercept therapy based on the drug's 2008 average selling price (assuming a dose of 25 mg twice weekly).13
The majority of the study participants were female (69.8%), with a mean age of 50 years (Table 1). Mean disease duration was 7.9 years, and mean time taking etanercept was 16.5 months. Participants perceived themselves as generally healthy, with 75.6% rating their overall health status as good, very good, or excellent. Almost all (ie, 97.7%) of the patients reported some level of RA improvement since starting etanercept therapy. All had medical insurance, including some drug coverage, and were employed across a spectrum of job categories, characterized by the 2000 population survey of the Bureau of Labor Statistics (Table 2).12
Comparison Before and Since Using Etanercept
Patients experienced significant improvement (P <.05, one-sample, paired t-test) on all work loss and productivity measures after starting, compared with before starting, etanercept therapy, with the exception of weekly work hours (P = .059). This included missed days in a month declining from 1.8 to 0.1 days, days RA limited work in a month declining from 6.5 to 0.2 days, and the effect of RA on productivity improving 3.7 points on a 1-to-10 scale (P <.001 for all measures; Table 3). In addition, after starting etanercept, a significant reduction was seen in the percentage of patients who reported any days missed from work (from 58.1% to 11.6%, P <.001; McNemar's test), ≥5 days missed from work (from 19.8% to 2.3%, P = .01), RA limiting the kind of work patients could perform (from 62.8% to 8.1%, P <.001), or RA limiting the type of work patients could perform for more than half the work days per month (from 31.4% to 0%, P <.001).
A subset analysis of patients (n = 19) who most recently started taking etanercept (5-10 months before the survey) showed significant decreases in missed work days monthly from 0.9 to 0.2 days (P = .003) and in number of days RA limited work monthly from 3.5 to 0.3 days (P = .022; Table 4). The effect of RA on productivity improved from a score of 4.2 to 2.1 days (P = .001). There was also a significant reduction in the percentage of patients who reported days missed from work (from 57.9% to 15.8%, P = .008) and RA limiting their work (from 47.4% to 10.5%, P = .016).
Other measures trended toward improvement but none of the results was significant. We performed the same analysis on the remainder of the 67 patients (ie, those who had been receiving etanercept for ≥11 months before the survey); significant improvement was seen on all work loss and productivity measures after starting etanercept compared with before starting etanercept therapy (Table 4).
We performed 2 additional subset analyses. The first looked at patients with different employment status (full-time vs part-time), and the second looked at patients in 2 different broad job categories (ie, sedentary vs requiring manual labor or mobility groups). In the first analysis, for the 70 patients who were employed fulltime, there were significant improvements in all work loss and productivity measures. For the 16 patients who were employed part-time, there was a significant improvement in some measures and a trend toward improvement in the others (Table 5).
For the subset analysis comparing patients in sedentary jobs with those in jobs involving manual labor or mobility, the magnitude of the difference between before and after starting etanercept was similar, except formissed days in a month and missing ≥5 days because of RA (both greater improvement in the manual labor group) and RA limiting the kind of work (greater improvement in the sedentary group) (Table 6).
Average Annual Economic Impact
Using standard wage rates to estimate the economic value resulting from reductions in absenteeism, after starting etanercept therapy, the average annual economic impact on employers was $3889 per employed patient with RA. Stratifying patients by income levels, average annual economic impacts resulting from decreased absenteeism alone were $2493, $4351, and $4850 per employed patient with RA with annual household income levels of <$40,000; $40,000 to <$80,000; and ≥$80,000, respectively.
Using standard wage rate conversions and different productivity-to-output assumptions, after starting etanercept therapy, the average annual economic impacts on employers per employed patient with RA as a result of improved productivity were $19,065 and $9533 for productivity- to-output ratios of 1:1 and 2:1, respectively. At a 1:1 productivity-to-output ratio, the average annual economic impacts as a result of improved productivity were $10,092, $27,292, and $33,442 per employed patient with RA with annual household income levels of <$40,000; $40,000 to <$80,000; and ≥$80,000, respectively. At a 2:1 productivity-to-output ratio, the average annual economic impacts resulting from improved productivity for these income groups were $5046; $13,646; and $16,721, respectively.
Taking into account decreased absenteeism and improved productivity as increased value or "savings" to the employer, and using a per-patient etanercept cost of $16,389, net annual savings to the employer for each employed patient with RA after starting etanercept was $6566 at a 1:1 productivity-to-output ratio. Net annual cost to the employer was $2967 at a 2:1 productivity- to-output ratio. Stratifying patients by income levels, assuming a 1:1 productivity-to-output ratio and combining the impact of drug costs with absenteeism and productivity, resulted in a net annual cost of $3803 for each patient with an annual household income level of <$40,000, but net annual savings of $15,254 and $21,904 for patients with annual household income levels of $40,000 to <$80,000 and ≥$80,000, respectively (Table 7).
The WLP study documents significant improvements in self-reported absenteeism and productivity among employed patients with RA who were taking etanercept and continued using it for 5 to 28 months. Amongpatients in our study, average hours worked weekly increased by 1.8 hours after starting etanercept therapy and average days missed monthly decreased by 1.7 days (13.4 hours) or 3.35 hours weekly, for an overall increase in hours worked weekly of 5.15 hours.
In a methodologically similar study regarding absenteeism among employed patients with RA receiving etanercept in which patients reported their work status via telephone interviews, Yelin and colleagues found unadjusted and adjusted values of 5.4 and 7.4 more hours worked per week, respectively, among patients with RA receiving etanercept versus matched patients who were not using etanercept.14 The lower numbers in our study may relate to the fact that Yelin and colleagues included patients with RA who were formerly not working but returned to the workforce after starting etanercept, whereas our study looked only at patients who remained employed before and after starting etanercept.
A study by Kavanaugh and colleagues showed an average increase of 3.54 hours worked weekly among patients with RA treated with methotrexate and either infliximab or placebo who had a clinically meaningful improvement in their Health Assessment Questionnaire (HAQ) score of ≥0.25.15 In terms of work loss among employed patients with RA before starting etanercept, the mean value in our study was 21.6 days per year as compared with a median estimate of mean values from a systematic review of work loss among patients with RA of 39 days per year, with a range of 7 to 84 days.4 The increased number of days of work lost in this systematic review versus our study may reflect the fact that this review included studies from the United States and Europe, as well as studies that were published as early as 1987.4 In addition to improvements in absenteeism and work loss after starting etanercept, our study showed significant improvements in productivity while at work. After monetizing the value of the improvements seen in our study, results suggest that the costs of etanercept may be partially or completely offset by increased time at work and improved productivity. These findings are supportive of the national Healthy People 2010 arthritis objective to reduce unemployment among adults with physician-diagnosed RA and to decrease the proportion who are limited in their ability to work for pay.16
Several limitations must be kept in mind when interpreting the results of this study. First, we could not find a suitable, validated questionnaire for this type of retro-spective study, so we created a questionnaire closely modeled after the validated Work Productivity and Activity Impairment Questionnaire, but modified it to include patients' pre- and post-etanercept experiences.17
Second, information about days lost from work and productivity at work were based on patients' self-reporting and may have been subject to recall bias, particularly for the time period before they started treatment. To address this issue, we performed a subset analysis of patients most recently started on etanercept (5-10 months before the survey)—because these patients would likely be least prone to recall bias—and still found statistically significant improvement in missed work days in a month, number of days RA limited work in a month, and the effect of RA on productivity. The absolute magnitude of these changes was lower than for the rest of the group, suggesting either some degree of recall bias or that patients receiving etanercept for a longer period of time had greater improvements in their RA.
Rather than asking each patient about his or her productivity and absenteeism before and after starting etanercept, an alternative would have been to compare patients who started treatment with untreated patients, as was done in a recent nonrandomized study by Wolfe and colleagues.18 Although when the 2 groups were compared at baseline, they differed in many key attributes that were known to contribute to the risk of disability (and, by extension, productivity and absenteeism), such as corticosteroid use; analgesic use; HAQ, pain, and severity index scores; joint scores; use of DMARDs in the past; and joint replacement.18 The authors of the study attempted to control for all of these effects,18 but as the authors of an accompanying editorial to the study stated, "These effects and differences cannot simply be disentangled by statistical techniques."19
Third, we collected data only on patients who were continuously employed before and since starting etanercept, rather than including those who were unemployed or on temporary disability leave before starting etanercept and may have returned to work because of clinical improvement after starting etanercept, or those who may have stopped work because of clinical deterioration after starting etanercept. We do not know of any published data on the impact of etanercept on unemployed patients with RA who desire employment. We do know that 7 patients in our study returned to work but only 1 patient stopped working after starting etanercept, although we do not know the reasons for these changes. The overall impact of excluding these cohorts depends on the relative numbers of patients in these 2 groups.
Fourth, by restricting the study to patients who had been receiving etanercept for at least 5 months, we were essentially looking only at responders, because nonresponders were unlikely to have remained on treatment long enough to have been part of the study. Non-responders would likely have been limited to no productivity improvements, but this study did not capture information on these patients. Nonresponders would still incur some drug costs until etanercept was discontinued, although these costs were not captured in this study. However, the majority of the costs of drugs such as etanercept are likely incurred by patients who have an adequate clinical response, and thus take the medications long-term.
Fifth, we examined only the cost impact to employers of providing etanercept to their employees; employers offering dependent care coverage would incur additional costs without the corresponding work gain and productivity benefits.
Finally, because this study focused on the impact of only 1 TNF-alpha inhibitor, etanercept, it is not possible to know how these results would compare with findings for the other available TNF-alpha inhibitors.
This study documents significant improvements in absenteeism and work productivity among employed patients with RA who started taking etanercept. Although the annual direct costs of biologic DMARDs such as etanercept are higher than the costs of regimens used in patients with less-severe disease, a broader view of the economic impact of these treatments for employed individuals with RA and for their employers should also incorporate the impact of these therapies on indirect costs, including work loss and productivity.
Studies such as this, that estimate the total economic impact of treatment from an employer's perspective, can be used to develop rational benefit plan designs that incorporate the full economic impact of novel biologic therapies for chronic diseases, including RA.
This study was funded by a grant from Amgen, Inc., Thousand Oaks, CA. Dr Globe and Dr Huang are employees of Amgen; Dr Mazonson, Mr Santas, Ms Murphy, and Ms Cheng are consultants to McKesson Corporation and have received honoraria or consultancy fees from Amgen; Dr Kavanaugh has nothing to disclose.
- American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines. Guidelines for the management of rheumatoid arthritis: 2002 update. Arthritis Rheum. 2002;46:328-346.
- Saag KG, Teng GG, Patkar NM, et al. American College of Rheumatology 2008 recommendations for the use of nonbiologic and biologic disease-modifying antirheumatic drugs in rheumatoid arthritis. Arthritis Rheum. 2008;59:762-784.
- Ozminkowski RJ, Burton WN, Goetzel RZ, et al. The impact of rheumatoid arthritis on medical expenditures, absenteeism, and short-term disability benefits. J Occup Environ Med. 2006;48:135-148.
- Burton W, Morrison A, Maclean R, Ruderman E. Systematic review of studies of productivity loss due to rheumatoid arthritis. Occup Med (Lond). 2006;56:18-27.
- Pugner KM, Scott DI, Holmes JW, Hieke K. The costs of rheumatoid arthritis: an international long-term view. Semin Arthritis Rheum. 2000;29:305-320.
- Lacaille D, Sheps S, Spinelli JJ, et al. Identification of modifiable work-related factors that influence the risk of work disability in rheumatoid arthritis. Arthritis Rheum. 2004;51:843-852.
- Furst DE, Keystone EC, Fleischmann R, et al. Updated consensus statement on biological agents for the treatment of rheumatic diseases, 2009. Ann Rheum Dis. 2010;69(suppl 1):i2-i29.
- Genovese MC, Bathon JM, Martin RW, et al. Etanercept versus methotrexate in patients with early rheumatoid arthritis: two-year radiographic and clinical outcomes. Arthritis Rheum. 2002;46:1443-1450.
- Weinblatt ME, Kremer JM, Bankhurst AD, et al. A trial of etanercept, a recombinant tumor necrosis factor receptor:Fc fusion protein, in patients with rheumatoid arthritis receiving methotrexate. N Engl J Med. 1999;340:253-259.
- Goldman DP, Joyce GF, Lawless G, et al. Benefit design and specialty drug use. Health Aff (Millwood). 2006;25:1319-1331.
- IMS Health. Top 15 US pharmaceutical products by sales. www.imshealth.com/ deployedfiles/imshealth/Global/Content/StaticFile/Top_Line_Data/2008_Top_15_Products_by_U.S._Sales.pdf. Accessed April 14, 2010.
- Bureau of the Census for the Bureau of Labor Statistics. Current population survey, March 2000. www.census.gov/apsd/techdoc/cps/cpsmar00.pdf. Accessed April 12, 2010.
- Centers for Medicare & Medicaid Services. Payment allowance limits for Medicare Part B drugs, effective April 1, 2008 through June 30, 2008. www.cms.gov/apps/ama/license.asp?file=/McrPartBDrugAvgSalesPrice/downloads/April08ASPbyHCPCS.zip. Accessed April 14, 2010.
- Yelin E, Trupin L, Katz P, et al. Association between etanercept use and employment outcomes among patients with rheumatoid arthritis. Arthritis Rheum. 2003;48:3046-3054.
- Kavanaugh A, Han C, Bala M. Functional status and radiographic joint damage are associated with health economic outcomes in patients with rheumatoid arthritis. J Rheumatol. 2004;3:849-855.
- Yelin E, Murphy L, Cisternas MG, et al. Medical care expenditures and earnings losses among persons with arthritis and other rheumatic conditions in 2003, and comparisons with 1997. Arthritis Rheum. 2007;56:1397-1407.
- Reilly MC, Zbrozek AS, Dukes EM. The validity and reproducibility of a work productivity and activity impairment instrument. Pharmacoeconomics. 1993;4:353-365.
- Wolfe F, Allaire S, Michaud K. The prevalence and incidence or work disability in rheumatoid arthritis, and the effect of anti-tumor necrosis factor on work disability. J Rheumatol. 2007;34:2211-2217.
- Verstappen SM, Jacobs JW, Hyrich KL. Effect of anti-tumor necrosis factor on work disability. J Rheumatol. 2007;34:2126-2128.