Arthritis Care & Research Vol. 67, No. 8, August 2015, pp 1086–1094 DOI 10.1002/acr.22574 C 2015, American College of Rheumatology V
ORIGINAL ARTICLE
Glucocorticoid Use in Patients With Systemic Lupus Erythematosus: Association Between Dose and Health Care Utilization and Costs SHIH-YIN CHEN,1 CHAN-BUM CHOI,2 QIAN LI,3 WEI-SHI YEH,1 YUAN-CHI LEE,3 AMY H. KAO,1 MATTHEW H. LIANG4
AND
Objective. To investigate the determinants of health care utilization and costs with use of glucocorticoid (GC) drugs among adult systemic lupus erythematosus (SLE) patients. Methods. This cross-sectional study analyzed established SLE patients identified by International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes from a large US insurance claims database in 2007–2011. Five patient groups were defined by their oral GC use during a 1-year period: non-GC users, 7.5 to £15 mg/day), or higher dosage (>15 mg/day). Annual health care utilization and costs were compared across these groups. Incremental costs of GC groups, calculated as the difference in total health care costs compared with those of the non-GC group, were estimated from multivariable regressions adjusting for demographic/clinical characteristics and stratified by concomitant immunosuppressant use. Results. A total of 50,230 SLE patients were identified (52% non-GC users, 20% 15 mg)
N (%) 50,230 (100) 26,319 (52) 10,073 (20) 5,017 (10) 4,808 (10) 4,013 (8) Demographics Age, mean 6 SD years 46.0 6 11.2 46.7 6 11.0 45.7 6 10.9† 47.0 6 11.0 45.1 6 11.5† 42.7 6 12.1† Women, % 90.7 90.7 93.1† 90.3 89.3† 86.6† Region of residence, % ‡ ‡ ‡ ‡ Northeast 16.2 16.8 15.5 15.0 14.8 17.2 Midwest 21.6 21.7 20.9 22.1 22.9 20.6 South 41.5 39.0 46.3 42.8 42.3 43.3 West 18.9 20.7 15.5 18.5 18.0 17.0 Unknown 1.8 1.8 1.8 1.6 2.0 1.9 Subset of lupus during 1-year followup, % Discoid/cutaneous lupus 15.2 13.1 16.1† 13.8 19.4† 23.4† Lupus nephritis 15.1 10.6 11.7† 21.3† 22.5† 36.7† Comorbidities during 1-year followup CCI score, mean 6 SD 2.5 6 1.9 2.3 6 1.7 2.6 6 1.9† 2.6 6 1.9† 2.8 6 2.0† 3.3 6 2.3† CCI score, % 1 38.4 43.4 34.5 36.5 33.1 23.6 2 24.9 25.4 26.8 23.9 22.8 20.7 3 15.5 14.4 17.0 15.5 17.0 17.3 $4 21.3 16.9 21.7 24.2 27.1 38.4 Oral GC utilization Days supplied, mean 6 SD N/A N/A 18.6 6 13.5 267.6 6 144.2 228.4 6 129.6 177.7 6 104.0 Daily dose, mean 6 SD mg N/A N/A 22.5 6 13.8 4.7 6 1.6 10.7 6 2.1 26.9 6 11.5 Cumulative dose, mean 6 SD mg N/A N/A 391.1 6 409.1 1,263.8 6 824.0 2,408.0 6 1427.5 4,529.9 6 2815 4.0 2.7 7.5† 3.1 4.6† 4.4† Use of intramuscular injectable steroid during 1-year followup, % Use of other lupus treatments during 1-year followup, % NSAIDs 35.5 32.2 46.0† 33.9‡ 36.0† 32.8 Antimalarial drugs 56.6 49.9 59.8† 68.9† 68.4† 63.4† Hydroxychloroquine or 56.6 49.8 59.8† 68.8† 68.3† 63.4† chloroquine Quinacrine 0.3 0.2 0.3 0.4† 0.3 0.4‡ Immunosuppressants 24.9 12.4 20.3† 44.0† 52.1† 61.4† Methotrexate 10.3 5.9 10.9† 17.4† 21.0† 15.3† Mycophenolate 8.5 3.4 4.7† 15.5† 18.4† 31.7† Azathioprine 7.2 3.0 5.2† 12.7† 16.8† 20.8† Cyclophosphamide 0.6 0.2 0.4‡ 0.5† 1.1† 3.5† Rituximab 0.8 0.3 0.7† 0.9† 1.4† 2.9† Other§ 1.6 0.7 1.1† 4.5† 3.6† 3.5† * GC 5 glucocorticoid; CCI 5 Charlson Comorbidity Index; N/A 5 not applicable; NSAIDs 5 nonsteroidal antiinflammatory drugs. † Statistically significant at P , 0.01 compared to non-GC users. ‡ Statistically significant at P , 0.05 compared to non-GC users. § Immunosuppressants in the other category include cyclosporine, tacrolimus, everolimus, sirolimus, antilymphocyte and antithymocyte globulin, basiliximab, daclizumab, muromonab, abatacept, and chlorambucil.
and others), and intramuscular GC injection, were identified using pharmacy claims and procedure codes on medical claims. Health care utilization of inpatient, emergency department (ED), and outpatient/physician office service, including the percentage of patients using the services and the frequency of events among the service users, was reported during the 1-year study period. For inpatient service, the
total number of hospital days among hospitalized patients was also summarized. The use of outpatient/physician office services was further specified to visits to primary care providers and specialists (i.e., rheumatologist, nephrologist, hematologist, neurologist, cardiologist, and dermatologist). Health care costs during the 1-year study period were calculated by combining patient- and plan-paid amounts from all medical and pharmacy claims. The total costs
Steroid Dose and Associated Costs in SLE
1089 characteristics, including demographics (age, sex, and region of residence), CCI, and subtypes of SLE (discoid/ cutaneous lupus and lupus nephritis), during the study period. Based on regression results, the incremental total costs were estimated for each of the GC cohorts, relative to no GC use, such that the expected instantaneous change in costs was a function of a change in GC variables while keeping all other covariates constant (21). Analyses were performed using SAS, version 9.2.
RESULTS
Figure 1. Sample assembly. SLE 5 systemic lupus erythematosus; ICD-9-CM 5 International Classification of Diseases, Ninth Revision, Clinical Modification.
were then categorized into the following types: inpatient, ED, outpatient, and pharmacy costs. All costs were adjusted to 2012 US dollars using the Consumer Price Index Medical Component (20). Statistical analysis. The study measures of GC user cohorts were compared to the non-GC users using the Student’s t-test for continuous variables, the chi-square test for categorical variables, and the nonparametric Wilcoxon test for count and cost variables. Among the SLE patients with $60 days of GC use, the trends of health care utilization and costs from low-dose users to higher-dose users (i.e., the dose-response relationships) were assessed by the Cochran-Armitage trend test for binary variables, and by univariate regression for continuous variables. Differences were considered as statistically significant when P was less than 0.05 for a 2-sided test. Multivariable regression was conducted to assess the association between total health care costs and use of GCs in separate subgroup analyses stratified by immunosuppressant use during the 1-year study period. Immunosuppressant use was determined based on any associated claims identified (as listed in Table 1) during the 1-year study period. The use of GCs was defined by the cohorts of GC users, with non-GC users as the reference. The regressions followed the generalized linear model with log link and gamma distribution, and adjusted for patient
Patient characteristics and use of GCs. The final study sample consisted of 50,230 SLE patients (Figure 1). Overall, the mean age was 46 years, 90.7% were women, and most patients were from the south region (41.5%) (Table 1). The prevalence of discoid/cutaneous lupus was 15.2% and of lupus nephritis was 15.1%. Other than GCs, the most common SLE treatments were antimalarials (56.6%), followed by NSAIDs (35.5%) and immunosuppressants (24.9%). Among the SLE patients, 26,319 (52%) did not use oral GCs during the 1-year study period, 10,073 (20%) used GCs for ,60 days, and 13,838 (28%) used GCs for $60 days at low (10%), medium (10%), and higher (8%) daily GC doses. Although they had fewer total days on GCs and lower cumulative doses of GCs, users with ,60 days of GCs had higher daily doses (mean 22.5 mg) and were more likely to have used intramuscular GC injections (7.5%). Among users with $60 days of GCs, the average days of GC use decreased with a higher daily dose (from 267.6 days among low-dose users to 177.7 days among higherdose users), while the use of intramuscular GC injection slightly increased with higher daily doses (from 3.1% to 4.4%). Compared with the non-GC users, the use of intramuscular GC injections was higher among all oral GC users (all P , 0.01) except for the low-dose GC users. Patient characteristics varied across the study cohorts. Compared with non-GC users, those who used ,60 days of GCs were slightly more likely to be younger and female, while those with longer GC use and higher doses were more likely to be younger but male. In general, GC users were more likely to have discoid/cutaneous lupus (except for low-dose GC users) or lupus nephritis and higher CCI scores than non-GC users (all P , 0.01). The prevalence of lupus nephritis was the highest among higher-dose users (36.7%). GC users were also more likely to use other SLE treatments than non-GC users (all P , 0.01), except for the use of NSAIDs among higher-dose users. Use of concomitant immunosuppressants increased noticeably with the longer duration and higher dose of GC use, ranging from 20.3% among those with ,60 days of GC use to 61.4% among higher-dose GC users. Health care resource use and costs. During the 1-year study period, 21% of all SLE patients had inpatient stays, of whom had on average 1.6 hospitalizations and 8.4 hospital days (Table 2). Approximately one-third (31.9%) had ED visits, with a mean of 2.1 visits. Almost all SLE patients had some outpatient or physician office visits
1090
Chen et al
Table 2.
Health care resource utilization during 1-year followup by study cohorts* Users with ‡60 days of GCs
All subjects
Non-GC users
Users with 7.5–15 mg) (>15 mg) 4,808 (10) 24.7‡ 1.6 6 1.2‡ 9.3 6 14.4‡ 37.6‡ 2.2 6 2.2‡ 100 28.9 6 22.8‡ 79.0‡ 7.3 6 7.1‡ 60.9‡ 5.4 6 3.9‡ 12.8‡ 5.3 6 8.1‡ 4.9‡ 4.3 6 6.6 13.1‡ 2.6 6 2.7 27.2‡ 2.7 6 2.8‡ 19.9 2.6 6 2.9‡
4,013 (8) 42.7‡ 1.9 6 1.5‡ 12.3 6 16.4‡ 46.7‡ 2.5 6 2.9‡ 99.9 34.5 6 25.1‡ 77.3 8.3 6 8.4‡ 60.0‡ 6.1 6 4.5‡ 20.2‡ 6.5 6 7.6‡ 6.8‡ 5.9 6 7.2‡ 15.0‡ 3.3 6 3.4‡ 31.1‡ 3.0 6 4.2‡ 21.0 2.8 6 2.9‡
, 0.01 , 0.01 , 0.01 , 0.01 , 0.01 0.15 , 0.01 0.17 , 0.01 0.31 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 0.01 , 0.01
GC 5 glucocorticoid; ED 5 emergency department. Trend test was done only among users with $60 days of GCs. P , 0.01 compared to non-GC users. No. of hospital days was calculated only among hospitalized patients. Frequency was calculated only among users. P , 0.05 compared to non-GC users.
(99.9%), and the majority had visits to primary care providers (77.7%). Use of specialty care was the highest for rheumatologist visits (54.7%), followed by cardiologist (21.8%), dermatologist (20.6%), neurologist (11.8%), nephrologist (8.2%), and hematologist (3.6%) visits. As one might expect, the number of consultations to subspecialists, especially nephrologists and hematologists, increased from low-dose users to higher-dose users. In general, SLE patients who used GCs had higher health care utilization than those without use of GCs (P , 0.05). Users with ,60 days of GC use had similar health care utilization to the medium-dose GC users with $60 days of GC use. Among users with $60 days of GC use, the dose-response relationship was observed for most types of health care utilizations (P , 0.05). Specifically, the percentage of patients with inpatient stays increased by more than 2-fold from lowdose users (18.2%) to higher-dose users (42.7%), and among those who were hospitalized, the number of hospital days increased by 57% (mean hospital stays of 7.8 days and 12.3 days for low-dose users and higher-dose users, respectively). The percentage of patients with rheumatologist visits was stable over the GC dose cohorts, but frequency of visits increased with higher GC dose. Mean total health care costs were $21,535 among all SLE patients over the 1-year study period (Figure 2). NonGC users had mean total costs of $16,162. Compared to non-GC users, GC users had higher total health care costs,
as well as higher cost across all settings (inpatient, ED, outpatient, and pharmacy costs; all P , 0.05). Among users with $60 days of GC use, mean total health care costs were increased by 2-fold from low-dose users to higher-dose users ($21,869 versus $45,360; P , 0.01). A dose-response relationship was also observed in all cost components (all P , 0.01). Mean inpatient costs among higher-dose users ($20,887) were 3 times the costs of low-dose users ($6,249). Inpatient costs were the largest cost component among higher-dose users, but outpatient costs were the largest component among the other study cohorts. Users with ,60 days of GC use had similar costs to medium-dose users. The multivariable regression on total health care costs among SLE patients stratified by concomitant immunosuppressant use is shown in Table 3. Among SLE patients with concomitant immunosuppressant use, total costs were 1.3 times (P , 0.01) higher in patients who used GCs for ,60 days compared to those of non-GC users, with adjustment of patient characteristics. This corresponds to incremental costs of $7,297 (Figure 3). Incremental total costs for patients with $60 days of GC use were $1,285 for low dose, $6,913 for medium dose, and $15,019 for higher dose (all P , 0.05). Among patients who did not use concomitant immunosuppressants, incremental total costs were $4,656 for ,60 days of GC use, $2,514 for low dose, $5,319 for medium dose, and $12,517 for higher dose (all P , 0.01) when compared with those of non-GC users. Other variables associated with higher
Steroid Dose and Associated Costs in SLE
1091
Figure 2. Average health care costs during 1-year followup by study cohorts. USD 5 US dollars; GC 5 glucocorticoid; ‡ 5 trend test was done only among users with $60 days of GCs; * 5 P , 0.01 compared to non-GC users; † 5 P , 0.05 compared to non-GC users; ED 5 emergency department.
costs included female sex, higher CCI score, and discoid/ cutaneous lupus and/or lupus nephritis.
DISCUSSION To the best of our knowledge, this is the largest US claims database study of adult SLE patients ages ,65 years that investigated the relationship of health care costs and utilization with dose and duration of GC use. Our study population is approximately 3 or more times higher than the number of commercially-insured SLE patients previously described (5,22). Our findings add to the current knowledge of the impact of GC use on health care costs and their determinants at a population level different from the Medicaid population, which has further disadvantages of being impoverished and/or certain other major comorbid illnesses. We also found that this population with complex multisystem disease consulted their primary care physicians the majority of the time, while also relying on a wide variety of subspecialists led by rheumatologists, cardiologists, and dermatologists. In this 1-year snapshot, we observed the pervasive use of GCs in the disease. More than half of the SLE patients received oral GC therapies. Among GC users, more than
half were prescribed with at least 60 days of therapy. In particular, close to two-thirds of the patients using GCs for more than 60 days were prescribed with a daily prednisone or equivalent dose of .7.5 mg. Our study reported the average annual total cost of all SLE patients to be $21,535, which was similar to what has been reported previously for the commercially insured population ($15,000–21,000) (5,23,24). We further found that higher GC doses were associated with higher health care utilization and costs. This positive association cannot be explained by the costs of GCs (which were part of those health care costs), because GCs are inexpensive medications. In our study sample, GC costs were on average $50 per person per year among users with $60 days of use, regardless of dose. Two likely explanations are that higher disease severity confounds the observed relationship between high GC doses and costs, especially in a cross-sectional design, and/or that events specifically related to higher GC doses, such as GC-related adverse events, may increase the costs (15,25). Future studies using longitudinal designs with additional data on disease features would identify the driving factors of the association between dose of GCs and costs of SLE. We used immunosuppressant use as a surrogate of SLE disease severity. In multivariate analysis, we stratified by
1092
Chen et al
Table 3.
Multivariable regression on total health care costs during 1-year followup stratified by immunosuppressant use* Concomitant immunosuppressant use (n 5 12,495)
Use of steroid Non-GC user Users with ,60 days of GCs Users with $60 days of GCs Low dose Medium dose Higher dose Clinical/demographic characteristics Age group, years 18–44 45–54 55–64 Female sex Region of residence Northeast Midwest South West Unknown CCI score 1 2 3 $4 Discoid/cutaneous lupus Lupus nephritis
No immunosuppressant use (n 5 37,735)
Cost ratio
95% CI
P
Cost ratio
95% CI
P
Ref. 1.30
Ref. 1.24–1.37
Ref. , 0.01
Ref. 1.29
Ref. 1.26–1.32
Ref. , 0.01
1.05 1.29 1.63
1.00–1.11 1.23–1.35 1.55–1.71
0.04 , 0.01 , 0.01
1.16 1.33 1.78
1.11–1.20 1.28–1.39 1.69–1.87
, 0.01 , 0.01 , 0.01
Ref. 1.05 1.06 1.09
Ref. 1.02–1.09 1.00–1.12 1.04–1.15
Ref. , 0.01 0.06 , 0.01
Ref. 0.98 1.02 1.12
Ref. 0.96–1.00 0.98–1.05 1.08–1.16
Ref. 0.06 0.31 , 0.01
1.04 1.04 Ref. 1.19 0.98
0.99–1.09 0.99–1.08 Ref. 1.14–1.25 0.87–1.11
0.14 0.09 Ref. , 0.01 0.77
0.97 1.01 Ref. 1.14 0.88
0.94–1.00 0.98–1.03 Ref. 1.11–1.17 0.82–0.95
0.04 0.60 Ref. , 0.01 , 0.01
Ref. 1.56 2.04 3.82 1.16 1.71
Ref. 1.50–1.63 1.94–2.14 3.64–4.01 1.11–1.21 1.64–1.78
Ref. , 0.01 , 0.01 , 0.01 , 0.01 , 0.01
Ref. 1.55 2.33 4.73 1.07 2.00
Ref. 1.52–1.59 2.26–2.40 4.59–4.88 1.04–1.10 1.94–2.07
Ref. , 0.01 , 0.01 , 0.01 , 0.01 , 0.01
* 95% CI 5 95% confidence interval; GC 5 glucocorticoid; CCI 5 Charlson Comorbidity Index.
concomitant immunosuppressant use to assess the relationship between GC dose and total health care costs within the more homogeneous subgroups in disease severity. In general, a dose-response relationship between GC dose and total health care costs was observed in both subgroups. Incremental costs remained high for higher-dose (;$13,000) and medium-dose (;$6,000) users, regardless of the use of concomitant immunosuppressive drugs. As we hypothesized, SLE patients who used immunosuppressants and maintained GCs at a low dose had smaller incremental costs ($1,285 when compared with non-GC users; P 5 0.04) than the incremental costs among lowdose GC users who did not use immunosuppressants ($2,514 when compared with non-GC users; P , 0.01). Our study also identified SLE patients who used shortterm but high mean daily doses of GCs. These patients not only used high daily GC doses, but also had higher use of intramuscular steroid injections. This may indicate the preference of some clinicians to manage minor disease flares and/or to ensure absorption by using short parenteral courses of steroids, despite a lack of data for their superiority over simple oral regimens. These short-term GC users presented with similar health care resource utilization and costs to users with medium daily doses for $60 days. These patients also incurred significantly higher costs; future research should aim at understanding whether short-term exposure to high-dose GCs, combining
both oral and injectable GCs, may be associated with higher risk of adverse events associated with GCs. The concern of the observed short GC therapy duration might be due to insufficient followup to capture subsequent GC use. However, by examining the subgroup of patients who had a longer followup (58% of patients with at least 2 years), we found that the majority of patients (more than three-quarters) were still being categorized as having ,60 days of GC therapy (data not shown). Therefore, misclassification would be less likely. This study had several limitations. First, the use of administrative claims to identify SLE is subject to classification and coding errors. Although the large sample size and geographically diverse population would likely reduce the systematic biases, the coding algorithm to identify SLE patients using claims data has not been validated against medical records to determine positive/negative predictive value or sensitivity/ specificity. Second, findings from a cross-sectional study may be confounded by patient characteristics and SLE disease severity, neither of which could be captured comprehensively in an administrative database. The GC users may have been different from the non-GC users in ways other than GC use, such as SLE disease severity, with these differences between the groups accounting more for their differences in health care costs than did GC use itself. Therefore, we attempted to control for disease severity by stratifying patients by immunosuppressant utilization, although this is an imperfect surrogate.
Steroid Dose and Associated Costs in SLE
1093
Figure 3. Adjusted incremental total costs during 1-year followup by study cohorts comparedwith non-glucocorticoid (non-GC) users among A, patients with use of immunosuppressants and B, patients without use of immunosuppressants. Incremental costs were calculated based on regression results, adjusted for patient characteristics. The non-GC user cohort was the reference cohort to calculate the increment costs. Average healthcare costs during 1-year followup were $20,801 among non-GC users with use of immunosuppressants (n 5 3,273), and were $15,503 among non-GC users without use of immunosuppressants (n 5 23,046). * 5 P , 0.01 compared to non-GC users; † 5 P , 0.05 compared to non-GC users.
Third, this study population is most representative of a patient population with existing SLE (not newly diagnosed) with commercial insurance and age ,65 years. This study does not include SLE patients who were covered by either Medicare based on the minimum age of 65 years or Medicaid based on certain disabilities, end-stage renal disease, or income level. Health care services and costs paid completely out of pocket or by other supplemental insurances were not captured, but we expect it would be minimal. Finally, the higher-dose group consists of patients with a wide range of mean daily GC doses from at least 15 mg to 90 mg. There might be more variation within this higher-dose group that will warrant further investigation in future research. Regardless, this study also has several strengths. First, this study is approximately 3-fold the size of previous studies and has a very large sample of younger SLE popu-
lations from different regions in the US. Second, although we could not establish a causal relationship, the association observed between dose of GCs and health care costs warrants further investigation of the possible mechanisms of the relationship, such as reduced occurrence and management of GC-associated adverse events. In conclusion, use of GCs, especially at higher doses, is associated with higher health care utilization and costs in patients with SLE. Therapies with a GC-sparing effect may be associated with lower health care economic burden in the treatment of SLE. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors
1094
Chen et al
approved the final version to be submitted for publication. Dr. Chen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Chen, Choi, Li, Yeh, Kao, Liang. Acquisition of data. Chen, Li. Analysis and interpretation of data. Chen, Choi, Li, Lee, Kao, Liang.
ROLE OF THE STUDY SPONSOR Biogen reviewed and provided feedback on the article to the authors. The authors had full editorial control of the article and provided their final approval of all content. Publication of this article was not contingent on approval by Biogen.
REFERENCES 1. Rus V, Hajeer A, Hochberg MC. Systemic lupus erythematosus. In: Silman AJ, Hochberg MC, editors. Epidemiology of the rheumatic disease. 2nd ed. New York (NY): Oxford University Press; 2001. 2. Lim SS, Bayakly AR, Helmick CG, Gordon C, Easley KA, Drenkard C. The incidence and prevalence of systemic lupus erythematosus, 2002–2004: the Georgia Lupus Registry. Arthritis Rheum 2014;66:357–68. 3. Somers EC, Marder W, Cagnoli P, Lewis EE, DeGuire P, Gordon C, et al. Population-based incidence and prevalence of systemic lupus erythematosus: the Michigan Lupus Epidemiology and Surveillance program. Arthritis Rheum 2014; 66:369–78. 4. Slawsky KA, Fernandes AW, Fusfeld L, Manzi S, Goss TF. A structured literature review of the direct costs of adult systemic lupus erythematosus in the US. Arthritis Care Res (Hoboken) 2011;63:1224–32. 5. Furst DE, Clarke A, Fernandes AW, Bancroft T, Gajria K, Greth W, et al. Resource utilization and direct medical costs in adult systemic lupus erythematosus patients from a commercially insured population. Lupus 2013;22:268–78. 6. Kan HJ, Song X, Johnson BH, Bechtel B, O’Sullivan D, Molta CT. Healthcare utilization and costs of systemic lupus erythematosus in Medicaid. Biomed Res Int 2013;2013: 808391. 7. Thorn GW, Bayles TB, Massell BF, Forsham PH, Hill SR Jr, Smith S III, et al. Studies on the relation of pituitary-adrenal function to rheumatic disease. N Engl J Med 1949;241:529– 37. 8. American College of Rheumatology Ad Hoc Committee on Systemic Lupus Erythematosus Guidelines. Guidelines for referral and management of systemic lupus erythematosus in adults. Arthritis Rheum 1999;42:1785–96. 9. Zonana-Nacach A, Barr SG, Magder LS, Petri M. Damage in systemic lupus erythematosus and its association with corticosteroids. Arthritis Rheum 2000;43:1801–8.
10. Ad Hoc Working Group on Steroid-Sparing Criteria in Lupus. Criteria for steroid-sparing ability of interventions in systemic lupus erythematosus: report of a consensus meeting. Arthritis Rheum 2004;50:3427–31. 11. Ginzler EM, Aranow C. Prevention and treatment of adverse effects of corticosteroids in systemic lupus erythematosus. Baillieres Clin Rheumatol 1998;12:495–510. 12. McDonough AK, Curtis JR, Saag KG. The epidemiology of glucocorticoid-associated adverse events. Curr Opin Rheumatol 2008;20:131–7. 13. Poetker DM, Reh DD. A comprehensive review of the adverse effects of systemic corticosteroids. Otolaryngol Clin North Am 2010;43:753–68. 14. Sarnes E, Crofford L, Watson M, Dennis G, Kan H, Bass D. Incidence and US costs of corticosteroid-associated adverse events: a systematic literature review. Clin Ther 2011;33: 1413–32. 15. Petri M, Bechtel B, Dennis G, Shah M, McLaughlin T, Kan H, et al. Burden of corticosteroid use in patients with systemic lupus erythematosus: results from a Delphi panel. Lupus 2014;23:1006–13. 16. Li T, Carls GS, Panopalis P, Wang S, Gibson TB, Goetzel RZ. Long-term medical costs and resource utilization in systemic lupus erythematosus and lupus nephritis: a five-year analysis of a large Medicaid population. Arthritis Rheum 2009;61:755–63. 17. Chibnik LB, Massarotti EM, Costenbader KH. Identification and validation of lupus nephritis cases using administrative data. Lupus 2010;19:741–3. 18. Yeh WS, Clark DW, McCarty KM. Health resource utilization of lupus nephritis patients: a comparison of result across case identification algorithms. Value Health 2013;16:A21. 19. Charlson ME, Charlson RE, Peterson JC, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol 2008;61:1234–40. 20. Bureau of Labor Statistics. Consumer Price Index. URL: http://www.bls.gov/cpi/. 21. Graubard BI, Korn EL. Predictive margins with survey data. Biometrics 1999;55:652–9. 22. Pelletier EM, Ogale S, Yu E, Brunetta P, Garg J. Economic outcomes in patients diagnosed with systemic lupus erythematosus with versus without nephritis: results from an analysis of data from a US claims database. Clin Ther 2009; 31:2653–64. 23. Carls G, Li T, Panopalis P, Wang S, Mell AG, Gibson TB, et al. Direct and indirect costs to employers of patients with systemic lupus erythematosus with and without nephritis. J Occup Environ Med 2009;51:66–79. 24. Garris C, Jhingran P, Bass D, Engel-Nitz NM, Riedel A, Dennis G. Healthcare utilization and cost of systemic lupus erythematosus in a US managed care health plan. J Med Econ 2013;16:667–77. 25. Shah M, Chaudhari S, McLaughlin TP, Kan HJ, Bechtel B, Dennis GJ, et al. Cumulative burden of oral corticosteroid adverse effects and the economic implications of corticosteroid use in patients with systemic lupus erythematosus. Clin Ther 2013;35:486–97.
ORIGINAL ARTICLE
Glucocorticoid Use in Patients With Systemic Lupus Erythematosus: Association Between Dose and Health Care Utilization and Costs SHIH-YIN CHEN,1 CHAN-BUM CHOI,2 QIAN LI,3 WEI-SHI YEH,1 YUAN-CHI LEE,3 AMY H. KAO,1 MATTHEW H. LIANG4
AND
Objective. To investigate the determinants of health care utilization and costs with use of glucocorticoid (GC) drugs among adult systemic lupus erythematosus (SLE) patients. Methods. This cross-sectional study analyzed established SLE patients identified by International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes from a large US insurance claims database in 2007–2011. Five patient groups were defined by their oral GC use during a 1-year period: non-GC users, 7.5 to £15 mg/day), or higher dosage (>15 mg/day). Annual health care utilization and costs were compared across these groups. Incremental costs of GC groups, calculated as the difference in total health care costs compared with those of the non-GC group, were estimated from multivariable regressions adjusting for demographic/clinical characteristics and stratified by concomitant immunosuppressant use. Results. A total of 50,230 SLE patients were identified (52% non-GC users, 20% 15 mg)
N (%) 50,230 (100) 26,319 (52) 10,073 (20) 5,017 (10) 4,808 (10) 4,013 (8) Demographics Age, mean 6 SD years 46.0 6 11.2 46.7 6 11.0 45.7 6 10.9† 47.0 6 11.0 45.1 6 11.5† 42.7 6 12.1† Women, % 90.7 90.7 93.1† 90.3 89.3† 86.6† Region of residence, % ‡ ‡ ‡ ‡ Northeast 16.2 16.8 15.5 15.0 14.8 17.2 Midwest 21.6 21.7 20.9 22.1 22.9 20.6 South 41.5 39.0 46.3 42.8 42.3 43.3 West 18.9 20.7 15.5 18.5 18.0 17.0 Unknown 1.8 1.8 1.8 1.6 2.0 1.9 Subset of lupus during 1-year followup, % Discoid/cutaneous lupus 15.2 13.1 16.1† 13.8 19.4† 23.4† Lupus nephritis 15.1 10.6 11.7† 21.3† 22.5† 36.7† Comorbidities during 1-year followup CCI score, mean 6 SD 2.5 6 1.9 2.3 6 1.7 2.6 6 1.9† 2.6 6 1.9† 2.8 6 2.0† 3.3 6 2.3† CCI score, % 1 38.4 43.4 34.5 36.5 33.1 23.6 2 24.9 25.4 26.8 23.9 22.8 20.7 3 15.5 14.4 17.0 15.5 17.0 17.3 $4 21.3 16.9 21.7 24.2 27.1 38.4 Oral GC utilization Days supplied, mean 6 SD N/A N/A 18.6 6 13.5 267.6 6 144.2 228.4 6 129.6 177.7 6 104.0 Daily dose, mean 6 SD mg N/A N/A 22.5 6 13.8 4.7 6 1.6 10.7 6 2.1 26.9 6 11.5 Cumulative dose, mean 6 SD mg N/A N/A 391.1 6 409.1 1,263.8 6 824.0 2,408.0 6 1427.5 4,529.9 6 2815 4.0 2.7 7.5† 3.1 4.6† 4.4† Use of intramuscular injectable steroid during 1-year followup, % Use of other lupus treatments during 1-year followup, % NSAIDs 35.5 32.2 46.0† 33.9‡ 36.0† 32.8 Antimalarial drugs 56.6 49.9 59.8† 68.9† 68.4† 63.4† Hydroxychloroquine or 56.6 49.8 59.8† 68.8† 68.3† 63.4† chloroquine Quinacrine 0.3 0.2 0.3 0.4† 0.3 0.4‡ Immunosuppressants 24.9 12.4 20.3† 44.0† 52.1† 61.4† Methotrexate 10.3 5.9 10.9† 17.4† 21.0† 15.3† Mycophenolate 8.5 3.4 4.7† 15.5† 18.4† 31.7† Azathioprine 7.2 3.0 5.2† 12.7† 16.8† 20.8† Cyclophosphamide 0.6 0.2 0.4‡ 0.5† 1.1† 3.5† Rituximab 0.8 0.3 0.7† 0.9† 1.4† 2.9† Other§ 1.6 0.7 1.1† 4.5† 3.6† 3.5† * GC 5 glucocorticoid; CCI 5 Charlson Comorbidity Index; N/A 5 not applicable; NSAIDs 5 nonsteroidal antiinflammatory drugs. † Statistically significant at P , 0.01 compared to non-GC users. ‡ Statistically significant at P , 0.05 compared to non-GC users. § Immunosuppressants in the other category include cyclosporine, tacrolimus, everolimus, sirolimus, antilymphocyte and antithymocyte globulin, basiliximab, daclizumab, muromonab, abatacept, and chlorambucil.
and others), and intramuscular GC injection, were identified using pharmacy claims and procedure codes on medical claims. Health care utilization of inpatient, emergency department (ED), and outpatient/physician office service, including the percentage of patients using the services and the frequency of events among the service users, was reported during the 1-year study period. For inpatient service, the
total number of hospital days among hospitalized patients was also summarized. The use of outpatient/physician office services was further specified to visits to primary care providers and specialists (i.e., rheumatologist, nephrologist, hematologist, neurologist, cardiologist, and dermatologist). Health care costs during the 1-year study period were calculated by combining patient- and plan-paid amounts from all medical and pharmacy claims. The total costs
Steroid Dose and Associated Costs in SLE
1089 characteristics, including demographics (age, sex, and region of residence), CCI, and subtypes of SLE (discoid/ cutaneous lupus and lupus nephritis), during the study period. Based on regression results, the incremental total costs were estimated for each of the GC cohorts, relative to no GC use, such that the expected instantaneous change in costs was a function of a change in GC variables while keeping all other covariates constant (21). Analyses were performed using SAS, version 9.2.
RESULTS
Figure 1. Sample assembly. SLE 5 systemic lupus erythematosus; ICD-9-CM 5 International Classification of Diseases, Ninth Revision, Clinical Modification.
were then categorized into the following types: inpatient, ED, outpatient, and pharmacy costs. All costs were adjusted to 2012 US dollars using the Consumer Price Index Medical Component (20). Statistical analysis. The study measures of GC user cohorts were compared to the non-GC users using the Student’s t-test for continuous variables, the chi-square test for categorical variables, and the nonparametric Wilcoxon test for count and cost variables. Among the SLE patients with $60 days of GC use, the trends of health care utilization and costs from low-dose users to higher-dose users (i.e., the dose-response relationships) were assessed by the Cochran-Armitage trend test for binary variables, and by univariate regression for continuous variables. Differences were considered as statistically significant when P was less than 0.05 for a 2-sided test. Multivariable regression was conducted to assess the association between total health care costs and use of GCs in separate subgroup analyses stratified by immunosuppressant use during the 1-year study period. Immunosuppressant use was determined based on any associated claims identified (as listed in Table 1) during the 1-year study period. The use of GCs was defined by the cohorts of GC users, with non-GC users as the reference. The regressions followed the generalized linear model with log link and gamma distribution, and adjusted for patient
Patient characteristics and use of GCs. The final study sample consisted of 50,230 SLE patients (Figure 1). Overall, the mean age was 46 years, 90.7% were women, and most patients were from the south region (41.5%) (Table 1). The prevalence of discoid/cutaneous lupus was 15.2% and of lupus nephritis was 15.1%. Other than GCs, the most common SLE treatments were antimalarials (56.6%), followed by NSAIDs (35.5%) and immunosuppressants (24.9%). Among the SLE patients, 26,319 (52%) did not use oral GCs during the 1-year study period, 10,073 (20%) used GCs for ,60 days, and 13,838 (28%) used GCs for $60 days at low (10%), medium (10%), and higher (8%) daily GC doses. Although they had fewer total days on GCs and lower cumulative doses of GCs, users with ,60 days of GCs had higher daily doses (mean 22.5 mg) and were more likely to have used intramuscular GC injections (7.5%). Among users with $60 days of GCs, the average days of GC use decreased with a higher daily dose (from 267.6 days among low-dose users to 177.7 days among higherdose users), while the use of intramuscular GC injection slightly increased with higher daily doses (from 3.1% to 4.4%). Compared with the non-GC users, the use of intramuscular GC injections was higher among all oral GC users (all P , 0.01) except for the low-dose GC users. Patient characteristics varied across the study cohorts. Compared with non-GC users, those who used ,60 days of GCs were slightly more likely to be younger and female, while those with longer GC use and higher doses were more likely to be younger but male. In general, GC users were more likely to have discoid/cutaneous lupus (except for low-dose GC users) or lupus nephritis and higher CCI scores than non-GC users (all P , 0.01). The prevalence of lupus nephritis was the highest among higher-dose users (36.7%). GC users were also more likely to use other SLE treatments than non-GC users (all P , 0.01), except for the use of NSAIDs among higher-dose users. Use of concomitant immunosuppressants increased noticeably with the longer duration and higher dose of GC use, ranging from 20.3% among those with ,60 days of GC use to 61.4% among higher-dose GC users. Health care resource use and costs. During the 1-year study period, 21% of all SLE patients had inpatient stays, of whom had on average 1.6 hospitalizations and 8.4 hospital days (Table 2). Approximately one-third (31.9%) had ED visits, with a mean of 2.1 visits. Almost all SLE patients had some outpatient or physician office visits
1090
Chen et al
Table 2.
Health care resource utilization during 1-year followup by study cohorts* Users with ‡60 days of GCs
All subjects
Non-GC users
Users with 7.5–15 mg) (>15 mg) 4,808 (10) 24.7‡ 1.6 6 1.2‡ 9.3 6 14.4‡ 37.6‡ 2.2 6 2.2‡ 100 28.9 6 22.8‡ 79.0‡ 7.3 6 7.1‡ 60.9‡ 5.4 6 3.9‡ 12.8‡ 5.3 6 8.1‡ 4.9‡ 4.3 6 6.6 13.1‡ 2.6 6 2.7 27.2‡ 2.7 6 2.8‡ 19.9 2.6 6 2.9‡
4,013 (8) 42.7‡ 1.9 6 1.5‡ 12.3 6 16.4‡ 46.7‡ 2.5 6 2.9‡ 99.9 34.5 6 25.1‡ 77.3 8.3 6 8.4‡ 60.0‡ 6.1 6 4.5‡ 20.2‡ 6.5 6 7.6‡ 6.8‡ 5.9 6 7.2‡ 15.0‡ 3.3 6 3.4‡ 31.1‡ 3.0 6 4.2‡ 21.0 2.8 6 2.9‡
, 0.01 , 0.01 , 0.01 , 0.01 , 0.01 0.15 , 0.01 0.17 , 0.01 0.31 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01 0.01 , 0.01
GC 5 glucocorticoid; ED 5 emergency department. Trend test was done only among users with $60 days of GCs. P , 0.01 compared to non-GC users. No. of hospital days was calculated only among hospitalized patients. Frequency was calculated only among users. P , 0.05 compared to non-GC users.
(99.9%), and the majority had visits to primary care providers (77.7%). Use of specialty care was the highest for rheumatologist visits (54.7%), followed by cardiologist (21.8%), dermatologist (20.6%), neurologist (11.8%), nephrologist (8.2%), and hematologist (3.6%) visits. As one might expect, the number of consultations to subspecialists, especially nephrologists and hematologists, increased from low-dose users to higher-dose users. In general, SLE patients who used GCs had higher health care utilization than those without use of GCs (P , 0.05). Users with ,60 days of GC use had similar health care utilization to the medium-dose GC users with $60 days of GC use. Among users with $60 days of GC use, the dose-response relationship was observed for most types of health care utilizations (P , 0.05). Specifically, the percentage of patients with inpatient stays increased by more than 2-fold from lowdose users (18.2%) to higher-dose users (42.7%), and among those who were hospitalized, the number of hospital days increased by 57% (mean hospital stays of 7.8 days and 12.3 days for low-dose users and higher-dose users, respectively). The percentage of patients with rheumatologist visits was stable over the GC dose cohorts, but frequency of visits increased with higher GC dose. Mean total health care costs were $21,535 among all SLE patients over the 1-year study period (Figure 2). NonGC users had mean total costs of $16,162. Compared to non-GC users, GC users had higher total health care costs,
as well as higher cost across all settings (inpatient, ED, outpatient, and pharmacy costs; all P , 0.05). Among users with $60 days of GC use, mean total health care costs were increased by 2-fold from low-dose users to higher-dose users ($21,869 versus $45,360; P , 0.01). A dose-response relationship was also observed in all cost components (all P , 0.01). Mean inpatient costs among higher-dose users ($20,887) were 3 times the costs of low-dose users ($6,249). Inpatient costs were the largest cost component among higher-dose users, but outpatient costs were the largest component among the other study cohorts. Users with ,60 days of GC use had similar costs to medium-dose users. The multivariable regression on total health care costs among SLE patients stratified by concomitant immunosuppressant use is shown in Table 3. Among SLE patients with concomitant immunosuppressant use, total costs were 1.3 times (P , 0.01) higher in patients who used GCs for ,60 days compared to those of non-GC users, with adjustment of patient characteristics. This corresponds to incremental costs of $7,297 (Figure 3). Incremental total costs for patients with $60 days of GC use were $1,285 for low dose, $6,913 for medium dose, and $15,019 for higher dose (all P , 0.05). Among patients who did not use concomitant immunosuppressants, incremental total costs were $4,656 for ,60 days of GC use, $2,514 for low dose, $5,319 for medium dose, and $12,517 for higher dose (all P , 0.01) when compared with those of non-GC users. Other variables associated with higher
Steroid Dose and Associated Costs in SLE
1091
Figure 2. Average health care costs during 1-year followup by study cohorts. USD 5 US dollars; GC 5 glucocorticoid; ‡ 5 trend test was done only among users with $60 days of GCs; * 5 P , 0.01 compared to non-GC users; † 5 P , 0.05 compared to non-GC users; ED 5 emergency department.
costs included female sex, higher CCI score, and discoid/ cutaneous lupus and/or lupus nephritis.
DISCUSSION To the best of our knowledge, this is the largest US claims database study of adult SLE patients ages ,65 years that investigated the relationship of health care costs and utilization with dose and duration of GC use. Our study population is approximately 3 or more times higher than the number of commercially-insured SLE patients previously described (5,22). Our findings add to the current knowledge of the impact of GC use on health care costs and their determinants at a population level different from the Medicaid population, which has further disadvantages of being impoverished and/or certain other major comorbid illnesses. We also found that this population with complex multisystem disease consulted their primary care physicians the majority of the time, while also relying on a wide variety of subspecialists led by rheumatologists, cardiologists, and dermatologists. In this 1-year snapshot, we observed the pervasive use of GCs in the disease. More than half of the SLE patients received oral GC therapies. Among GC users, more than
half were prescribed with at least 60 days of therapy. In particular, close to two-thirds of the patients using GCs for more than 60 days were prescribed with a daily prednisone or equivalent dose of .7.5 mg. Our study reported the average annual total cost of all SLE patients to be $21,535, which was similar to what has been reported previously for the commercially insured population ($15,000–21,000) (5,23,24). We further found that higher GC doses were associated with higher health care utilization and costs. This positive association cannot be explained by the costs of GCs (which were part of those health care costs), because GCs are inexpensive medications. In our study sample, GC costs were on average $50 per person per year among users with $60 days of use, regardless of dose. Two likely explanations are that higher disease severity confounds the observed relationship between high GC doses and costs, especially in a cross-sectional design, and/or that events specifically related to higher GC doses, such as GC-related adverse events, may increase the costs (15,25). Future studies using longitudinal designs with additional data on disease features would identify the driving factors of the association between dose of GCs and costs of SLE. We used immunosuppressant use as a surrogate of SLE disease severity. In multivariate analysis, we stratified by
1092
Chen et al
Table 3.
Multivariable regression on total health care costs during 1-year followup stratified by immunosuppressant use* Concomitant immunosuppressant use (n 5 12,495)
Use of steroid Non-GC user Users with ,60 days of GCs Users with $60 days of GCs Low dose Medium dose Higher dose Clinical/demographic characteristics Age group, years 18–44 45–54 55–64 Female sex Region of residence Northeast Midwest South West Unknown CCI score 1 2 3 $4 Discoid/cutaneous lupus Lupus nephritis
No immunosuppressant use (n 5 37,735)
Cost ratio
95% CI
P
Cost ratio
95% CI
P
Ref. 1.30
Ref. 1.24–1.37
Ref. , 0.01
Ref. 1.29
Ref. 1.26–1.32
Ref. , 0.01
1.05 1.29 1.63
1.00–1.11 1.23–1.35 1.55–1.71
0.04 , 0.01 , 0.01
1.16 1.33 1.78
1.11–1.20 1.28–1.39 1.69–1.87
, 0.01 , 0.01 , 0.01
Ref. 1.05 1.06 1.09
Ref. 1.02–1.09 1.00–1.12 1.04–1.15
Ref. , 0.01 0.06 , 0.01
Ref. 0.98 1.02 1.12
Ref. 0.96–1.00 0.98–1.05 1.08–1.16
Ref. 0.06 0.31 , 0.01
1.04 1.04 Ref. 1.19 0.98
0.99–1.09 0.99–1.08 Ref. 1.14–1.25 0.87–1.11
0.14 0.09 Ref. , 0.01 0.77
0.97 1.01 Ref. 1.14 0.88
0.94–1.00 0.98–1.03 Ref. 1.11–1.17 0.82–0.95
0.04 0.60 Ref. , 0.01 , 0.01
Ref. 1.56 2.04 3.82 1.16 1.71
Ref. 1.50–1.63 1.94–2.14 3.64–4.01 1.11–1.21 1.64–1.78
Ref. , 0.01 , 0.01 , 0.01 , 0.01 , 0.01
Ref. 1.55 2.33 4.73 1.07 2.00
Ref. 1.52–1.59 2.26–2.40 4.59–4.88 1.04–1.10 1.94–2.07
Ref. , 0.01 , 0.01 , 0.01 , 0.01 , 0.01
* 95% CI 5 95% confidence interval; GC 5 glucocorticoid; CCI 5 Charlson Comorbidity Index.
concomitant immunosuppressant use to assess the relationship between GC dose and total health care costs within the more homogeneous subgroups in disease severity. In general, a dose-response relationship between GC dose and total health care costs was observed in both subgroups. Incremental costs remained high for higher-dose (;$13,000) and medium-dose (;$6,000) users, regardless of the use of concomitant immunosuppressive drugs. As we hypothesized, SLE patients who used immunosuppressants and maintained GCs at a low dose had smaller incremental costs ($1,285 when compared with non-GC users; P 5 0.04) than the incremental costs among lowdose GC users who did not use immunosuppressants ($2,514 when compared with non-GC users; P , 0.01). Our study also identified SLE patients who used shortterm but high mean daily doses of GCs. These patients not only used high daily GC doses, but also had higher use of intramuscular steroid injections. This may indicate the preference of some clinicians to manage minor disease flares and/or to ensure absorption by using short parenteral courses of steroids, despite a lack of data for their superiority over simple oral regimens. These short-term GC users presented with similar health care resource utilization and costs to users with medium daily doses for $60 days. These patients also incurred significantly higher costs; future research should aim at understanding whether short-term exposure to high-dose GCs, combining
both oral and injectable GCs, may be associated with higher risk of adverse events associated with GCs. The concern of the observed short GC therapy duration might be due to insufficient followup to capture subsequent GC use. However, by examining the subgroup of patients who had a longer followup (58% of patients with at least 2 years), we found that the majority of patients (more than three-quarters) were still being categorized as having ,60 days of GC therapy (data not shown). Therefore, misclassification would be less likely. This study had several limitations. First, the use of administrative claims to identify SLE is subject to classification and coding errors. Although the large sample size and geographically diverse population would likely reduce the systematic biases, the coding algorithm to identify SLE patients using claims data has not been validated against medical records to determine positive/negative predictive value or sensitivity/ specificity. Second, findings from a cross-sectional study may be confounded by patient characteristics and SLE disease severity, neither of which could be captured comprehensively in an administrative database. The GC users may have been different from the non-GC users in ways other than GC use, such as SLE disease severity, with these differences between the groups accounting more for their differences in health care costs than did GC use itself. Therefore, we attempted to control for disease severity by stratifying patients by immunosuppressant utilization, although this is an imperfect surrogate.
Steroid Dose and Associated Costs in SLE
1093
Figure 3. Adjusted incremental total costs during 1-year followup by study cohorts comparedwith non-glucocorticoid (non-GC) users among A, patients with use of immunosuppressants and B, patients without use of immunosuppressants. Incremental costs were calculated based on regression results, adjusted for patient characteristics. The non-GC user cohort was the reference cohort to calculate the increment costs. Average healthcare costs during 1-year followup were $20,801 among non-GC users with use of immunosuppressants (n 5 3,273), and were $15,503 among non-GC users without use of immunosuppressants (n 5 23,046). * 5 P , 0.01 compared to non-GC users; † 5 P , 0.05 compared to non-GC users.
Third, this study population is most representative of a patient population with existing SLE (not newly diagnosed) with commercial insurance and age ,65 years. This study does not include SLE patients who were covered by either Medicare based on the minimum age of 65 years or Medicaid based on certain disabilities, end-stage renal disease, or income level. Health care services and costs paid completely out of pocket or by other supplemental insurances were not captured, but we expect it would be minimal. Finally, the higher-dose group consists of patients with a wide range of mean daily GC doses from at least 15 mg to 90 mg. There might be more variation within this higher-dose group that will warrant further investigation in future research. Regardless, this study also has several strengths. First, this study is approximately 3-fold the size of previous studies and has a very large sample of younger SLE popu-
lations from different regions in the US. Second, although we could not establish a causal relationship, the association observed between dose of GCs and health care costs warrants further investigation of the possible mechanisms of the relationship, such as reduced occurrence and management of GC-associated adverse events. In conclusion, use of GCs, especially at higher doses, is associated with higher health care utilization and costs in patients with SLE. Therapies with a GC-sparing effect may be associated with lower health care economic burden in the treatment of SLE. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors
1094
Chen et al
approved the final version to be submitted for publication. Dr. Chen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Chen, Choi, Li, Yeh, Kao, Liang. Acquisition of data. Chen, Li. Analysis and interpretation of data. Chen, Choi, Li, Lee, Kao, Liang.
ROLE OF THE STUDY SPONSOR Biogen reviewed and provided feedback on the article to the authors. The authors had full editorial control of the article and provided their final approval of all content. Publication of this article was not contingent on approval by Biogen.
REFERENCES 1. Rus V, Hajeer A, Hochberg MC. Systemic lupus erythematosus. In: Silman AJ, Hochberg MC, editors. Epidemiology of the rheumatic disease. 2nd ed. New York (NY): Oxford University Press; 2001. 2. Lim SS, Bayakly AR, Helmick CG, Gordon C, Easley KA, Drenkard C. The incidence and prevalence of systemic lupus erythematosus, 2002–2004: the Georgia Lupus Registry. Arthritis Rheum 2014;66:357–68. 3. Somers EC, Marder W, Cagnoli P, Lewis EE, DeGuire P, Gordon C, et al. Population-based incidence and prevalence of systemic lupus erythematosus: the Michigan Lupus Epidemiology and Surveillance program. Arthritis Rheum 2014; 66:369–78. 4. Slawsky KA, Fernandes AW, Fusfeld L, Manzi S, Goss TF. A structured literature review of the direct costs of adult systemic lupus erythematosus in the US. Arthritis Care Res (Hoboken) 2011;63:1224–32. 5. Furst DE, Clarke A, Fernandes AW, Bancroft T, Gajria K, Greth W, et al. Resource utilization and direct medical costs in adult systemic lupus erythematosus patients from a commercially insured population. Lupus 2013;22:268–78. 6. Kan HJ, Song X, Johnson BH, Bechtel B, O’Sullivan D, Molta CT. Healthcare utilization and costs of systemic lupus erythematosus in Medicaid. Biomed Res Int 2013;2013: 808391. 7. Thorn GW, Bayles TB, Massell BF, Forsham PH, Hill SR Jr, Smith S III, et al. Studies on the relation of pituitary-adrenal function to rheumatic disease. N Engl J Med 1949;241:529– 37. 8. American College of Rheumatology Ad Hoc Committee on Systemic Lupus Erythematosus Guidelines. Guidelines for referral and management of systemic lupus erythematosus in adults. Arthritis Rheum 1999;42:1785–96. 9. Zonana-Nacach A, Barr SG, Magder LS, Petri M. Damage in systemic lupus erythematosus and its association with corticosteroids. Arthritis Rheum 2000;43:1801–8.
10. Ad Hoc Working Group on Steroid-Sparing Criteria in Lupus. Criteria for steroid-sparing ability of interventions in systemic lupus erythematosus: report of a consensus meeting. Arthritis Rheum 2004;50:3427–31. 11. Ginzler EM, Aranow C. Prevention and treatment of adverse effects of corticosteroids in systemic lupus erythematosus. Baillieres Clin Rheumatol 1998;12:495–510. 12. McDonough AK, Curtis JR, Saag KG. The epidemiology of glucocorticoid-associated adverse events. Curr Opin Rheumatol 2008;20:131–7. 13. Poetker DM, Reh DD. A comprehensive review of the adverse effects of systemic corticosteroids. Otolaryngol Clin North Am 2010;43:753–68. 14. Sarnes E, Crofford L, Watson M, Dennis G, Kan H, Bass D. Incidence and US costs of corticosteroid-associated adverse events: a systematic literature review. Clin Ther 2011;33: 1413–32. 15. Petri M, Bechtel B, Dennis G, Shah M, McLaughlin T, Kan H, et al. Burden of corticosteroid use in patients with systemic lupus erythematosus: results from a Delphi panel. Lupus 2014;23:1006–13. 16. Li T, Carls GS, Panopalis P, Wang S, Gibson TB, Goetzel RZ. Long-term medical costs and resource utilization in systemic lupus erythematosus and lupus nephritis: a five-year analysis of a large Medicaid population. Arthritis Rheum 2009;61:755–63. 17. Chibnik LB, Massarotti EM, Costenbader KH. Identification and validation of lupus nephritis cases using administrative data. Lupus 2010;19:741–3. 18. Yeh WS, Clark DW, McCarty KM. Health resource utilization of lupus nephritis patients: a comparison of result across case identification algorithms. Value Health 2013;16:A21. 19. Charlson ME, Charlson RE, Peterson JC, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol 2008;61:1234–40. 20. Bureau of Labor Statistics. Consumer Price Index. URL: http://www.bls.gov/cpi/. 21. Graubard BI, Korn EL. Predictive margins with survey data. Biometrics 1999;55:652–9. 22. Pelletier EM, Ogale S, Yu E, Brunetta P, Garg J. Economic outcomes in patients diagnosed with systemic lupus erythematosus with versus without nephritis: results from an analysis of data from a US claims database. Clin Ther 2009; 31:2653–64. 23. Carls G, Li T, Panopalis P, Wang S, Mell AG, Gibson TB, et al. Direct and indirect costs to employers of patients with systemic lupus erythematosus with and without nephritis. J Occup Environ Med 2009;51:66–79. 24. Garris C, Jhingran P, Bass D, Engel-Nitz NM, Riedel A, Dennis G. Healthcare utilization and cost of systemic lupus erythematosus in a US managed care health plan. J Med Econ 2013;16:667–77. 25. Shah M, Chaudhari S, McLaughlin TP, Kan HJ, Bechtel B, Dennis GJ, et al. Cumulative burden of oral corticosteroid adverse effects and the economic implications of corticosteroid use in patients with systemic lupus erythematosus. Clin Ther 2013;35:486–97.
