Economic burden of interstitial lung disease in a commercially insured population with Sjögren syndrome in the United States

Abstract

BACKGROUND:

Patients with Sjögren syndrome (SjS) have substantial cost burden on the health care system; among these patients, those who develop interstitial lung disease (ILD) experience poorer quality of life and have a higher mortality risk. However, the economic burden of ILD has not been documented.

OBJECTIVE:

To estimate the direct health care costs associated with ILD among patients with SjS in a representative sample of the commercially insured population in the United States.

METHODS:

Individuals with a diagnosis of SjS between January 1, 2006, and September 30, 2015, with and without a diagnosis of ILD, were identified from the PharMetrics Plus for Academics database. The index date was defined as the later date of the first claim with a diagnosis of SjS or the first claim with a diagnosis of ILD for individuals with SjS and ILD (SjS-ILD), and the first claim with a diagnosis of SjS for SjS-only controls. All baseline variables were measured in the 180 days preindex period. A 5:1 propensity score matching was applied to controls for baseline demographic and geographic variables. The cost ratio and average marginal effect for total direct medical costs comparing SjS patients with and without ILD were estimated using a generalized linear model. Costs per health care resource utilization category were also reported. All costs were represented from a health plan payer perspective and inflated to 2020 US dollars.

RESULTS:

After applying the inclusion criteria, 815 SjS-ILD cases were identified and matched to 4,075 SjS-only controls based on the 5:1 propensity score matching procedure. The 180-day total cost of SjS-ILD cases was about 2 times higher compared with that of SjS-only controls (adjusted cost ratio = 1.95; 95% CI = 1.76-2.15). The average difference in total cost between patients with and without ILD was $8,814 (95% CI = $7,149-$10,479). Costs were mainly contributed from outpatient services other than physician office visit (such as radiological and pathological tests), inpatient services, and outpatient pharmacy cost components for both groups (39.4%, 38.8%, and 16.3% for SjS-ILD cases; 43.7%, 22.6%, and 22.9% for SjS-only controls, respectively).

CONCLUSIONS:

Total direct health care cost was substantially higher in patients with SjS and ILD compared with patients with SjS without ILD. Our findings provide the foundation for further economic evaluation for preventive strategies to reduce the clinical and economic burden imposed by ILD among patients with SjS.

Plain language summary

The cost of treating interstitial lung disease (ILD) in patients with Sjögren syndrome (SjS) is substantial. Health care costs were doubled among patients with SjS who had ILD compared with those who did not have ILD. The additional 6-month cost was $8,814 on average. Treatments are available to slow down or prevent ILD. Evidence regarding the economic value of these treatments can be developed based on the current study.

Implications for managed care pharmacy

ILD is irreversible and causes clinical and economic burdens. Treatments are available that can slow down the disease progression. This study documented the high cost of ILD in patients with SjS in a commercially insured population. To inform formulary decisions, this study is the necessary first step to quantify the cost of illness. Further economic evaluation can build on this study to assess the cost-effectiveness of new treatments for ILD.

Sjögren syndrome (SjS) is a chronic, multisystem autoimmune disease characterized by lacrimal and salivary gland inflammation due to lymphocytic infiltration, resulting in glandular symptoms of dry eyes and mouth. In addition, the spectrum of SjS-related extraglandular manifestations may involve a variety of systems, including musculoskeletal, neural, dermatological, hematological, gastrointestinal, hepatic, renal, cardiovascular, and pulmonary involvement.^1,2 Pulmonary involvement in SjS consists of various forms of small airways and interstitial lung disease (ILD).³ ILD is considered the most serious and frequent pulmonary complication in SjS.⁴ In the United States, SjS affects up to 3.1 million Americans and most commonly occurs in women in their 50s and 60s.^5-7 Among patients with primary SjS, approximately 10%-20% develop ILD.^2,8

Patients with SjS have a diminished quality of life because of its high clinical burden.^9,10 Among patients with SjS, those manifested with ILD (SjS-ILD) have a further lower quality of life because of impaired physical function and higher mortality compared with those without pulmonary involvement.¹¹ A retrospective study that reviewed medical records in 2 academic centers demonstrated that patients with SjS with ILD have poor prognosis and that ILD is correlated with a decrease in functional status. Examining the clinical outcomes of ILD in 21 patients with SjS over a median follow-up of 24 months, the investigators reported that no patients experienced ILD resolution; rather, 7 patients (33.3%) experienced ILD deterioration despite therapy with steroids alone or in combination with azathioprine. Among the 7 deteriorated patients, 1 developed progressive respiratory failure and 1 died of ILD complication.¹² Another study of 62 patients with SjS-ILD reported that 50%, 10%, and 25% of patients, respectively, experienced ILD progression, had acute exacerbation, and died over 5 years of follow-up.¹³

The direct health care costs of SjS are significant.^14-16 One study in the United Kingdom reported that the annual total direct health care cost of SjS is similar to that of rheumatoid arthritis and more than 2 times greater than healthy controls.¹⁴ With respect to the United States, 1 commercial insurance claims study reported a 40% increased total direct health care cost of SjS 1 year after diagnosis in comparison with the year before diagnosis.¹⁶ However, to our knowledge, the economic burden of ILD among patients with SjS has not been documented. Information on the economic burden of ILD is particularly important given that recent trials showed the efficacy of antifibrotic drugs in slowing the progression of ILD.^17-19 This new finding provides opportunities for improving patient outcomes and reducing costs through early diagnosis and treatment. Therefore, a cost-identification analysis is warranted as the first step in performing an economic evaluation and generating evidence for decision-making, such as reimbursement decisions for new medicinal technologies.

The objective of this study is to quantify the direct health care costs associated with ILD in patients with SjS in comparison with those without ILD (SjS-only) in a national sample of the commercially insured population in the United States.

Methods

DATA SOURCE

This retrospective study used a 10% random sample of enrollees from the PharMetrics Plus for Academics database, which contains fully adjudicated health plan claims data and enrollment information for more than 70 health plans and more than 100 million commercially insured individuals throughout the United States from 2006 to 2015. The database includes medical and pharmacy claims data as well as demographic information. Each contributing health plan’s claims data undergo rigorous data quality review by IQVIA prior to its addition to the database. This study was deemed exempt by the Institutional Review Board at the University of Maryland, Baltimore (HP-00096910).

STUDY POPULATION

Individuals were included if they have at least 1 nonancillary claim with an SjS diagnosis (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 710.2) in any diagnosis position between January 1, 2006, and September 30, 2015. Patients were split into 2 mutually exclusive cohorts based on whether they had at least 1 claim with ILD diagnosis (ICD-9-CM codes 515, 516.30, 516.31, 516.32, 516.35, 516.36, 516.8, 517.8, and 518.89) during the same study period. Individuals with at least 1 claim with a diagnosis of ILD were categorized as SjS-ILD; individuals without a diagnosis of ILD were categorized as individuals with SjS only. All individuals were required to be 18 years and older and have 180-day pre-index continuous enrollment to capture baseline characteristic information and 180-day post-index continuous enrollment to ensure complete outcome measurement during follow-up.

For SjS-ILD cases, the index date was defined as the later date of the first claim with SjS diagnosis and the first claim with ILD diagnosis to capture the cost and resource use of having both conditions. For SjS-only controls, the index date was defined as the date of the first claim with SjS diagnosis. We allowed SjS-ILD cases to have their first ILD diagnosis claim occur before or after their first claim with SjS diagnosis because of the following reasons: (1) ILD can be the presenting manifestation of SjS²⁰ and the ILD diagnosis may occasionally precede the diagnosis of SjS⁴; (2) the claims database we used might not reflect the actual disease diagnosis date; and (3) the length of continuous enrollment in the commercial insurance database was not long enough to capture the full economic burden for every SjS-ILD case if we set the date of the first claim with SjS diagnosis as the index date. This approach of defining the index date as the later date of the first claim with an ILD diagnosis and the first claim with the diagnosis of the autoimmune disease of study interest has also been used in a previous study that examined the economic burden of systemic sclerosis-related ILD (SSc-ILD).²¹ We also conducted 2 sensitivity analyses: (1) we limited the duration of time between the first claim with an ILD diagnosis and the first claim with an SjS diagnosis to 15 months on the basis of, the median time reported in a previous study;¹² and (2) we excluded patients with other connective tissue diseases. All individuals were required to have an index date within the case identification period from July 1, 2006, to March 31, 2015, to allow for a minimum 180-day pre-index and a 180-day post-index period.

BASELINE MEASURES

Baseline characteristics, including age, sex, geographic region of residence, index year, and insurance plan product type, were measured at the index date. Baseline comorbidities were measured using the Charlson Comorbidity Index (CCI) modified by Quan et al²² in the 180-day preindex period. Other autoimmune diseases (ie, rheumatoid arthritis, systemic lupus erythematosus, and systemic sclerosis) were also included in the baseline covariates. Detailed information on variable definition and associated diagnosis codes are provided in Supplementary Table 1 ^{(215.5KB, pdf)} , available in online article.

OUTCOME MEASURES

All costs measured were direct health care costs obtained from the dollar amount paid by a health plan to a provider or facility for services supplied. We calculated the average all-cause health care cost per patient in the 180-day followup period, including the index date, by summing all costs in the claims file for that period and dividing them by the number of individuals in each group. We also stratified the cost by health care resource utilization categories (ie, inpatient, emergency department, pharmacy, physician office visit, and other outpatient services). Costs were compared between SjS-ILD cases and SjS-only controls as a ratio and as an average marginal cost. All costs were adjusted to 2020 US dollars using the annual average Consumer Price Index for the medical care component across the United States to account for inflation.²³

STATISTICAL ANALYSIS

To account for baseline covariates that are potential confounders, SjS-ILD cases were matched to SjS-only controls based on propensity scores with a 5:1 ratio of the number of controls matched to cases. The propensity scores were estimated based on logistic regression modeling with the following independent variables: index year, age, sex, geographic region of residence, and insurance product type.

Student’s t-test, chi-square test, and Cochran-Mantel-Haenszel test were conducted to compare baseline characteristics. Student’s t-test was used to test for differences in the continuous variable (ie, age), chi-square test was used to test for differences in the categorical variables (ie, sex), and Cochran-Mantel-Haenszel test was used to test for differences in nominal variables. Standardized differences were also calculated to examine and compare the distributions of the covariates between cases and controls before and after the propensity score matching.

A generalized linear model (GLM) with a γ distribution and a log link were fit to compare total costs for SjS-ILD cases vs SjS-only controls. The average marginal effect of ILD on the total costs of patients with SjS was estimated by computing the marginal effect for each individual and then averaging over the individual marginal effects.²⁴ The 95% CIs for the average marginal costs were estimated using the delta method. CCI and other related autoimmune diseases at baseline were not used in the propensity score model to achieve a higher rate of matched samples; instead, it was controlled for in the GLM. The remaining unbalanced variables after propensity score matching (ie, age and index year categories) were further controlled in the GLM to account for the potential residual confounding. Categorical variables (eg, CCI categories) were analyzed in the GLM by creating dummy variables indicating each of the categories. All analyses were performed using SAS studio (version 9.4, SAS Institute Inc). All differences between groups were tested at the 2-sided 5% significance level.

Results

BASELINE CHARACTERISTICS

A total of 9,836 patients met the study inclusion criteria. A flowchart presenting the attrition of the study sample is shown in Figure 1. After propensity score matching, a total of 4,890 study subjects were included in the analysis, including 815 SjS-ILD cases and 4,075 SjS-only controls. Patients’ baseline characteristics for cases and controls before and after propensity score matching are shown in Tables 1 and 2, respectively.

FIGURE 1.

Study Sample Attrition Diagram for Assembling SjS-ILD Cases and SjS-Only Controls

TABLE 1.

Baseline Characteristics of SjS-ILD Cases vs SjS-Only Controls Before Propensity Score Matching Among a Total of 9,836 Individuals

Characteristics	SjS-ILD cases (n = 815)	SjS-only controls (n = 9,021)	P valuea	Standardized difference
Age, y, mean (SD)	58.47 (11.43)	52.54 (12.62)	< 0.01	0.49
Sex, n (%)
Female	696 (85.4)	7,862 (87.15)	0.15	0.05
Male	119 (14.6)	1,159 (12.85)
Residence region, n (%)
Northeast	197 (24.17)	2,191 (24.29)	0.84	0.05
Midwest	212 (26.01)	2,225 (24.66)
South	288 (35.34)	3,297 (36.55)
West	118 (14.48)	1,308 (14.5)
Index year, n (%)
2006-2008	160 (19.63)	2,265 (25.11)	< 0.01	0.12
2009-2011	330 (40.49)	3,373 (37.39)
2012-2015	325 (39.88)	3,383 (37.5)
Plan product type, n (%)
Preferred provider organization	608 (74.6)	6,732 (74.63)	0.65	0.03
Health maintenance organization	102 (12.52)	1,073 (11.89)
Other/unknownb	105 (12.88)	1,216 (13.48)
Quan Charlson Comorbidity Index
0	165 (20.25)	5,652 (62.65)	< 0.01	1.00
1	301 (36.93)	2,118 (23.48)
2 +	349 (42.82)	1,251 (13.87)
Other related autoimmune diseasesc
Rheumatoid arthritis	161 (19.75)	991 (10.99)	< 0.01	0.24
Systemic lupus erythematosus	125 (15.34)	587 (6.51)	< 0.01	0.29
Systemic sclerosis	58 (7.12)	118 (1.31)	< 0.01	0.29

^aStudent’s t-test was used for continuous variables, chi-square test was used for categorical variables, and Cochran-Mantel-Haenszel test was used for multilevel nominal variables.

^bOther health plan product types include consumer-directed health care, indemnity plan, and point of service.

^cDermatomyositis and polymyositis were not included because of rare or zero prevalence in our sample.

ILD = interstitial lung disease; SjS = Sjögren syndrome; y = years.

TABLE 2.

Baseline Characteristics of SjS-ILD Cases vs SjS-Only Controls After Propensity Score Matching Among a Total of 4,890 Individuals

Characteristics	SjS-ILD cases (n = 815)	SjS-only controls (n = 4,075)	P valuea	Standardized difference
Age, y, mean (SD)	58.47 (11.43)	53.47 (12.32)	< 0.01	0.41
Sex, n (%)
Female	696 (85.4)	3,548 (87.07)	0.20	0.05
Male	119 (14.6)	527 (12.93)
Residence region, n (%)
Northeast	197 (24.17)	991 (24.32)	0.90	0.03
Midwest	212 (26.01)	1,019 (25.01)
South	288 (35.34)	1,442 (35.39)
West	118 (14.48)	623 (15.29)
Index year, n (%)
2006-2008	160 (19.63)	1,012 (24.83)	< 0.01	0.13
2009-2011	330 (40.49)	1,478 (36.27)
2012-2015	325 (39.88)	1,585 (38.9)
Plan product type, n (%)
Preferred provider organization	608 (74.6)	3,050 (74.85)	0.44	0.06
Health maintenance organization	102 (12.52)	467 (11.46)
Others/unknownb	105 (12.88)	558 (13.69)
Quan Charlson Comorbidity Index
0	165 (20.25)	2,550 (62.58)	< 0.01	1.02
1	301 (36.93)	962 (23.61)
2 +	349 (42.82)	563 (13.82)
Other related autoimmune diseasesc
Rheumatoid arthritis	161 (19.75)	460 (11.29)	< 0.01	0.24
Systemic lupus erythematosus	125 (15.34)	264 (6.48)	< 0.01	0.29
Systemic sclerosis	58 (7.12)	69 (1.69)	< 0.01	0.27

Variables that are used as independent variables in the propensity score model include the following: age, sex, residence region, index year, and plan product type.

^aStudent’s t-test was used for continuous variables, chi-square test was used for categorical variables, and Cochran-Mantel-Haenszel test was used for multilevel nominal variables.

^bOther health plan product types include consumer-directed health care, indemnity plan, and point of service.

^cDermatomyositis and polymyositis were not included because of rare or zero prevalence in our sample.

ILD = interstitial lung disease; SjS = Sjögren syndrome; y = years.

HEALTH CARE COSTS: DESCRIPTIVE ANALYSIS

The mean 180-day costs of individuals in the SjS-ILD cohort and the SjS-only cohort are presented in Table 3. The mean (SD) 180-day total costs were $23,192 ($48,189) for SjS-ILD cases and $8,648 ($17,478) for SjS-only controls. Outpatient services other than physician office visits (ie, laboratory, pathology, radiology, outpatient surgical, and ancillary services), inpatient services, and outpatient pharmacy were the major cost components in both groups (39.4%, 38.8%, and 16.3% for SjS-ILD cases, respectively; 43.7%, 22.6%, and 22.9% for SjS-only controls, respectively).

TABLE 3.

Mean 180-Day Costs in the SjS-ILD Cohort and the SjS-Only Cohort Among a Total of 4,890 Propensity Score–Matched Subjects

	SjS-ILD cases (n= 815)			SjS-only controls (n = 4,075)			Adjusted cost ratio (95% CI)a,b	Adjusted average marginal cost (95% CI)a,b,c
	Mean (SD), $	%	Median (IQR), $	Mean (SD), $	%	Median (IQR), $
Total cost	23,192 (48,189)	100.0	7,960 (18,668)	8,648 (17,478)	100.0	3,369 (7,107)	1.95 (1.76-2.15)d	8,814 (7,149-10,479)d
Cost per HCRU category
Inpatient	8,997 (36,824)	38.8	0 (1,532)	1,952 (11,769)	22.6	0 (0)	N/A	N/A
ED visit	232 (964)	1.0	0 (0)	140 (645)	1.6	0 (0)	N/A	N/A
Pharmacy	3,785 (8,212)	16.3	1,273 (3,575)	1,977 (4,472)	22.9	540 (1,879)	N/A	N/A
Physician office visit	1,036 (1,076)	4.5	720 (1,031)	799 (921)	9.2	545 (752)	N/A	N/A
Other outpatiente	9,142 (21,304)	39.4	2,929 (7,453)	3,781 (8,723)	43.7	1,062 (3,177)	N/A	N/A

Variables that are used as independent variables in the propensity score model include the following: age, sex, residence region, index year, and plan product type

^aGLM with a γ distribution and a log link was fit. Costs represented strictly to nonzero cost to fit γ distribution (N=4,836; 811 cases and 4,025 controls)

^bIndependent variables included in the GLM were comparison group (ie, SjS-ILD vs SjS-only), age, dummy variables for index year categories, dummy variables for CCI categories, baseline RA, baseline SLE, and baseline systemic sclerosis

^cCI was estimated using the delta method.

^dP < 0.01

^eOther outpatient services refer to outpatient services other than physician office visits, including laboratory, pathology, radiology, outpatient surgical, and ancillary services.

ED = emergency department; GLM = generalized linear model; HCRU = health care resource utilization; ILD = interstitial lung disease; N/A = not applicable; RA = rheumatoid arthritis; SjS = Sjögren syndrome; SLE = systemic lupus erythematosus.

HEALTH CARE COSTS: ADJUSTED REGRESSION ANALYSIS

The adjusted effect of ILD on total direct health care cost among patients with SjS in both the multiplicative and additive scale are presented in Table 3. The 180-day total costs of individuals with SjS-ILD were approximately 2 times higher than individuals with SjS-only (adjusted cost ratio = 1.95; 95% CI = 1.76-2.15; P < 0.01). The adjusted average marginal effect, that is, the adjusted average difference in the total costs comparing SjS-ILD cases and SjS-only controls, was $8,814 (95% CI = $7,149-$10,479; P < 0.01). The results of the sensitivity analyses were consistent with the result in the main analysis with respect to adjusted cost ratio and adjusted average marginal effect (Supplementary Tables 2 and 3 ^{(215.5KB, pdf)} ).

Discussion

ILD is one of the most common pulmonary complications in SjS. It has been shown to increase the morbidity and mortality of patients with SjS. Although the significant economic impact of SjS has been documented, it is not yet examined in this clinically unique subset of patients. In this study, we used a representative sample of the commercially insured population aged 65 years and younger to estimate the economic burden among patients with SjS-ILD in the United States. We found that patients with SjS-ILD incurred $23,192 in the 180 days after being identified as having both conditions. In comparison with matched SjS-only controls, patients with SjS-ILD had roughly 2 times higher total direct health care cost and an average marginal cost of $8,814 after controlling for baseline comorbidities. The costs of inpatient admissions, outpatient pharmacy, and other outpatient services were the main cost drivers.

To our knowledge, this is the first study to estimate costs among patients with SjS-ILD. Previous commercial claims studies have reported substantial costs for SSc-ILD and rheumatoid arthritis–related ILD (RA-ILD) in the United States.^21,25 The annual total direct health care costs for patients with SSc-ILD and RA-ILD were $33,195 and about $30,000-$50,000, respectively. The costs we identified for SjS-ILD ($22,562) are difficult to compare with these numbers since we captured 180-day costs in our study. Additionally, 1 commercial claims study documented the annual total direct health care costs for SjS to be $20,417.¹⁶ This annual estimate does not readily compare with the 180-day costs we observed for the SjS-only cohort in our study ($8,648).

The prognosis of SjS-ILD is poor. The condition is characterized by a chronically deteriorating process, can cause irreversible lung damage, and is associated with reduced physical functioning and increased risks of acute exacerbation and mortality.^11-13 Against this backdrop, the current study quantifies the high economic impact of SjS-ILD on the health care system. Our results show that patients with SjS-ILD have higher health care expenditures overall and the cost is consistently higher across each of the service settings. This finding of cost differences is not surprising. The higher average costs in emergency department and hospitalization among patients with SjS-ILD are likely because of the progression and acute exacerbation of ILD.¹³ The higher average pharmacy cost and physician office visit cost are likely because of the need of more aggressive therapy (eg, azathioprine, cyclophosphamide, or rituximab) as well as more frequent follow-up with a pulmonologist or rheumatologist.¹² Lastly, the higher cost of other unclassified outpatient services are likely because of increased laboratory (eg, serologic studies), pathology (eg, lung biopsy), radiology (eg, chest X-ray and high-resolution computed tomography), and other ancillary services (eg, pulmonary functional tests, pulmonary rehabilitation, and oxygen therapy) for the diagnosis, evaluation, and management of ILD.²⁶

Recent findings from clinical trials have shown benefits of 2 existing antifibrotic therapies (nintedanib^18,19 and pirfenidone¹⁷) for the treatment of progressive fibrosing ILD. This benefit was consistent across ILD diagnostic subgroups, including connective tissue disease-associated ILDs in both trials of the 2 antifibrotic therapies.^17,19 More importantly, the finding of the benefits in reducing ILD progression rate in patients treated with nintedanib led to the US Food and Drug Administration’s approval in March 2020, with a new indication for the treatment of chronic fibrosing ILD with a progressive phenotype.²⁷ Our study findings can be used in future value assessment studies to support reimbursement decisions regarding patient access to medications that are intended to slow down irreversible lung function degeneration in this vulnerable population.

LIMITATIONS

This study has the following limitations, many of which are characteristic of claims database studies. First, the diagnosis date is not available in the administrative claims database. Guided by our exploratory analysis and previous literature, we chose the later date of the first claim with a diagnosis of SjS and the first claim with a diagnosis of ILD and then assessed the economic burden 180 days after the later date to ensure we captured the costs of having both conditions. Therefore, this study may not capture the full disease burden from the disease onset of SjS for the SjS-ILD cohort. Although this may have led to an underestimation of the economic impact comparing SjS-ILD cases to SjS-only controls, it may provide a conservative estimate for the cost outcome. Second, the case definition relied on administrative diagnosis codes, as opposed to being based on documented clinical examination and test information. As such, misclassification of SjS and ILD may have occurred. The ICD-9 code to identify patients with SjS (710.2) is classified as sicca syndrome and may not be solely used for SjS. However, there is precedence and justification for using this code based on prior literature. A previous study applied the same code for estimating the health care burden of SjS in another commercial database.¹⁶ In addition, the sensitivity and specificity of the ICD-9 code 710.2 for SjS has been estimated to be 95.5% and 95.8%, respectively, in a study that compared administrative claims against data abstracted from medical charts.²⁸ Uncertainty in the clinical diagnosis of ILD and SjS guided our definition of ILD and SjS, which allowed the ILD diagnosis to precede the one for SjS in some instances. In our study, the time between the first ILD and SjS claims (ie, 406 days) among those with ILD preceding SjS was similar to findings from prior studies.^12,29 More details on the distribution of the time (in days) between the first claim for ILD and SjS are available in Supplementary Figures 1 and 2 ^{(215.5KB, pdf)} . Furthermore, the mean age of 53 years and 58 years in the SjS-only and the SjS-ILD cohorts, respectively; the high proportion of female patients (85%-87%), and the approximate 8% prevalence of ILD among patients with SjS reported in the current study are consistent with the descriptions of patients with SjS and SjS-ILD in the published literature (ie, mean age of 50 and 60 years in patients with SjS and SjS-ILD, a female-to-male ratio of 9:1 in patients with SjS, and a 10% 1-year cumulative incidence of ILD in patients with SjS).^7,8 Third, most enrollees in the PharMetrics Plus for Academics database are covered by commercial insurance, with a limited subset of Medicare Advantage and Medicaid patients. Therefore, the results may not be generalizable to other populations, such as those publicly insured or uninsured. Additionally, patients who are older than 65 years may be partially covered by Medicare, potentially underestimating the costs. Finally, our study provides benchmarking evidence and may not reflect the contemporary economic burden of SjS-ILD. Future studies using more recent data (eg, International Classification of Diseases, Tenth Revision, Clinical Modification codes) will be needed to confirm our findings.

Despite the limitations considered above, this study provides generalizable information for an understudied and clinically relevant subgroup of individuals diagnosed with SjS. Our results regarding the economic burden of SjS-ILD are timely given recent therapeutic innovations and can be used for planning and implementing policies regarding resource allocation for innovative technologies. Future studies in the publicly insured population, with longer follow-up, and more recent data are warranted to increase the generalizability of the finding of this study. Studies that examine the utilization outcomes of each service setting will be a constructive next step to further understand the pattern of resource use. Studies with richer clinical information are also encouraged to disentangle the impact of disease severity on the burden of this clinically and economically relevant subgroup.

Conclusions

Compared with individuals with SjS alone, individuals with SjS-ILD incurred substantial economic burden from the commercial payer perspective, in large part because of outpatient services other than physician visits, outpatient pharmacy, and inpatient costs. Economic evaluations of preventive strategies to reduce the clinical and economic burden imposed by ILD among patients with SjS are warranted.