Patient Journey, Disease Burden, Treatment Patterns, and the Impact of Delayed Diagnosis Among Patients With Sjögren's Disease: A Real‐World US Survey

Abstract

Objective

The aim of this study was to investigate disease burden, treatment patterns, and impact of delayed diagnosis on patient outcomes in Sjögren's disease (SjD) in a real‐world setting in the United States.

Methods

Data were drawn from the Adelphi Primary Sjögren's Syndrome Disease Specific Programme, a US‐based cross‐sectional collaborative survey between rheumatologists and their next six consecutive consulting patients with SjD. Rheumatologists provided patients’ demographic and clinical information using medical records and a patient‐completed questionnaire.

Results

Rheumatologists (N = 84) provided data on 511 patients with SjD. Calculated Clinical EULAR Sjögren's Syndrome Disease Activity Index (ClinESSDAI)–proxy scores (ps) indicated that patients had mild (44.2%), moderate (30.9%), or severe (24.9%) disease. Among 297 patients, 33.7% had a EULAR Sjögren's Syndrome Patient‐Reported Index (ESSPRI)‐ps of ≥5. Descriptive trends in patient‐reported EuroQoL 5‐Dimension 3‐Level (EQ‐5D‐3L) utility, EuroQoL Visual Analog Scale (EQ‐VAS), Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT‐F), and Work Productivity and Activity Impairment questionnaire scores suggested worsening impairment with increasing SjD severity. Mean (±SD) time to diagnosis was 2.3 (±5.6) years. EQ‐5D‐3L utility (P = 0.042), EQ‐VAS (P = 0.026), FACIT‐F (P = 0.007) scores, and ESSPRI‐ps (P = 0.001) differed significantly by duration of diagnostic delay. The odds of ClinESSDAI‐ps and ESSPRI‐ps ≥5 at data collection were 2.3 and 2.1 times greater, respectively, in patients diagnosed >12 versus <3 months after symptom onset.

Conclusion

Greater ClinESSDAI severity, longer diagnostic delays, and inefficient patient referrals were associated with worsened symptom management and poorer outcomes. Patients with moderate/severe disease who do not receive timely advanced treatment may experience greater disease burden, highlighting an unmet treatment need.

INTRODUCTION

Sjögren's disease (SjD) is a systemic, heterogeneous, autoimmune rheumatic disease that affects multiple organs and systems, potentially leading to serious complications and damage to joints, skin, lymph nodes, lungs, or virtually any organ system. ¹ The current prevalence of diagnosis in the United States is estimated to be between 0.0033% and 0.21% across all ages. ² SjD is characterized by dysfunction and destruction of the salivary and lacrimal glands. ³ , ⁴ SjD is also associated with a broad range of systemic manifestations ³ , ⁴ , ⁵ and the production of autoantibodies, ⁶ with complications such as lymphoma affecting about 2% to 10% of patients. ⁷ , ⁸ SjD symptoms—primarily dryness, pain, and fatigue—and ailments resulting from such symptoms negatively impact physical, psychological, and social components of health‐related quality of life (QoL). ⁹ , ¹⁰ , ¹¹ Patients with SjD have an increased risk of mortality compared with the general population, ¹² and the clinical burden of disease is considerable. ¹³

Diagnosis of SjD is challenging due to its heterogeneity, variable progression, and frequent subclinical characteristics of sicca manifestations. Additionally, because presenting symptoms are often subjective and nonspecific, many patients experience a substantial diagnostic delay, and more than half may remain undiagnosed. ¹⁴ SjD can occur alone (primary SjD) or in the presence of overlapping autoimmune diseases such as rheumatoid arthritis, systemic lupus erythematosus, or scleroderma (associated SjD). ⁴

The European Alliance of Associations for Rheumatology (EULAR) developed and validated disease‐specific activity indices to facilitate SjD evaluation in clinical trials. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ These include the EULAR Sjögren's Syndrome Disease Activity Index (ESSDAI) and the EULAR Sjögren's Syndrome Patient‐Reported Index (ESSPRI). ESSDAI evaluates systemic activity, whereas ESSPRI measures the primary symptoms of dryness, pain, and fatigue. Since its development, ESSDAI, which encompasses 12 domains, has been used frequently in randomized controlled trials as an outcome measure. ¹⁸ Clinical ESSDAI (ClinESSDAI) is closely related to ESSDAI and allows evaluation of disease activity independent of the biologic domain. Although ClinESSDAI does not fully capture patient‐reported burden, it provides complementary insight into systemic disease activity. The therapeutic management of SjD is based on relief of sicca symptoms with topical, oral, and ocular therapies and off‐label immunosuppression for systemic disease, ²⁰ but no cure or US Food and Drug Administration–approved systemic treatments exist. ¹ Current clinical guidelines also focus solely on symptomatic relief ²⁰ because therapies lack clinical evidence supporting their efficacy and are often poorly tolerated by patients. ¹⁴ Because newer treatments are being evaluated in SjD, it is important to establish the real‐world disease burden, including delay in diagnosis and its impact on patient outcomes. As part of this study, we describe the patient journey and disease burden in patients with primary SjD in a real‐world clinical setting, as well as diagnostic delay and its impact on disease severity, symptom burden, and patient‐reported outcomes (PROs).

METHODS

Survey design

Data were drawn from the Adelphi Real World Primary SjD Disease Specific Programme (DSP), a real‐world, cross‐sectional survey of rheumatologists and their consulting patients with SjD in a clinical setting in the United States in 2018. The patient sample included only patients with primary SjD as classified by their physician. The DSP comprises physician surveys and a physician‐completed patient record form (PRF) paired with patient self‐completion (PSC) forms. The DSP methodology has been published and validated. ²¹ , ²² This study reflects real‐world clinical practice, in which diagnosis is often based on physician judgment rather than the strict application of classification criteria and therefore captures a broader and potentially more heterogeneous population. The survey gained ethical approval from the Western Institutional Review Board, study protocol number 1186583.

Physician and patient eligibility criteria

Rheumatologists with 3 to 30 years’ experience who were responsible for the management of and treatment decisions for patients with SjD and seeing at least seven patients with SjD in a typical month were eligible for inclusion in the study. Patients included in the study were ≥18 years old; had a confirmed diagnosis of SjD and medical history of current or previous noninfectious fever, lymphadenopathy/lymphoma, glandular swelling/parotid enlargement, arthralgia/synovitis, erythema/vasculitis/purpura, systemic involvement (pulmonary, renal, peripheral, or central nervous system), myositis, autoimmune cytopenia (with neutropenia), anemia, thrombocytopenia, lymphopenia, hypocomplementemia, hypergammaglobulinemia, cryoglobulinemia, or hypogammaglobulinemia; and/or were prescribed an immunosuppressant or biologic therapy for SjD. Patients were excluded if they were involved in a clinical trial.

Data collection

For this analysis, rheumatologists were asked to complete a PRF for their next six consecutive eligible patients with SjD. Data captured in the physician‐reported PRF included patient demographics and clinical characteristics such as journey to diagnosis, disease and symptom severity, symptomatology, physician involvement and referrals, treatment use and history, and treatment satisfaction. Physicians then invited the same eligible patients to complete an optional PSC form. Data captured in the PSC form included patient demographics, clinical characteristics, and PRO measures including the EuroQoL 5‐Dimension 3‐Level (EQ‐5D‐3L) utility score, EuroQoL Visual Analog Scale (EQ‐VAS), ²³ Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT‐F), ²⁴ , ²⁵ and Work Productivity and Activity Impairment (WPAI) questionnaire. ²⁶ , ²⁷ The EQ‐5D‐3L utility and EQ‐VAS scores have ranges of 0 to 1 and 0 to 100, respectively, with higher scores indicating better QoL. ²³ The FACIT‐F score ranges from 0 to 52, with lower scores indicating worse fatigue. ²⁴ , ²⁵ Patients reported whether they required caregivers, their relationship to caregivers, and the total number of hours per week they received care. Furthermore, on the PSC form, patients rated their agreement (on a scale of 1–5, with 1 being “completely disagree” and 5 being “completely agree”) with several statements regarding their psychological and emotional well‐being. Responses to the WPAI questionnaire were used to calculate the percentage of work time missed, impairment while working, overall work impairment, and activity impairment due to SjD. ²⁶ , ²⁷ ClinESSDAI proxy scores (ps) were calculated for each patient by assigning a score to the rheumatologists’ perception (“not present,” “mild,” “moderate,” or “severe”) for each clinical domain of the ESSDAI (excluding the biologic domain); scores were then grouped according to the definitions of mild (0–4), moderate (5–13), and severe (≥14) disease activity. ¹⁹ Physicians reported the time of onset of SjD symptoms and diagnosis in the PRF. Responses were collected as a continuous measure, and patients were subsequently categorized post hoc into specific time intervals (<3 months, 3–6 months, 6–12 months, and >12 months). ESSPRI‐ps were calculated using the patients’ rating of dryness, fatigue, and pain on a scale of 0 to 10. The patient‐acceptable symptom state is defined as an ESSPRI of <5. ¹⁹

All data relevant to the analysis are included in the article. All data that support the findings of this survey are the intellectual property of Adelphi Real World. Data are available upon reasonable request to Nicola Massey at nicola.massey@omc.com.

Statistical analysis

Patient characteristics were summarized using descriptive analyses. Means and SDs were calculated for continuous variables, and frequencies and percentages were calculated for categorical variables. Because analyses were descriptive and not powered for statistical comparison, comparative findings cannot be generalized to the wider population.

Clinical characteristics were summarized descriptively (Supplementary Table S1), whereas Spearman's rho was used to assess the correlation between ClinESSDAI‐ps severity and PROs (Figure 1). To assess the impact of delayed diagnosis, patients were categorized post hoc according to the reported time between onset of SjD symptoms and diagnosis (<3 months, 3–6 months, 6–12 months, and >12 months). Analysis of variance and Kruskal–Wallis tests were used to examine the impact of delayed diagnosis on ClinESSDAI‐ps, ESSPRI‐ps, and PRO measures across groups. Multivariate linear and logistic regression analyses (both unadjusted and adjusted for time since diagnosis, age, sex, body mass index, and ethnicity) examined the impact of delayed diagnosis on SjD severity (ClinESSDAI‐ps), symptom burden (ESSPRI‐ps), and other PRO measures (FACIT‐F and WPAI).

Figure 1.

Patient‐reported outcome scores reported by proxy ClinESSDAI score in patients with SjD. Error bars denote SD. Higher EQ‐5D‐3L utility scores and EQ‐VAS scores indicate better quality of life; lower FACIT‐F scores indicate worse fatigue. ClinESSDAI, Clinical EULAR Sjӧgren’s Syndrome Disease Activity Index; EQ‐5D‐3L, EuroQoL 5‐Dimension 3‐Level; EQ‐VAS, EuroQoL Visual Analog Scale; FACIT‐F, Functional Assessment of Chronic Illness Therapy–Fatigue; SjD, Sjögren's disease.

RESULTS

Participant demographics and clinical characteristics

Rheumatologists (N = 84) provided data on 511 patients with SjD. At the time of the survey, patients had a mean (±SD) age of 53.3 (±14.2) years, 88.5% were female, and 79.8% were White; 67.0% had never smoked, 49.1% of patients were working full‐time, and the mean (±SD) time to diagnosis was 2.3 (±5.6) years (Supplementary Table S1). ClinESSDAI‐ps (mean [±SD] score 9.5 [±9.4]) were considered moderate or severe (ClinESSDAI‐ps ≥ 5) in over half (n = 285 [55.8%]) of patients at the time of data collection (Supplementary Table S1). Demographic characteristics, such as age and sex, were mostly consistent across groups by increasing ClinESSDAI severity, with a few exceptions. At the time of data collection, the mean time since diagnosis among patients with mild ClinESSDAI scores was 4.0 years, 4.8 years among those with moderate scores, and 5.1 years among patients with severe ClinESSDAI scores. Patients with severe disease activity were more likely to be smokers (14.2%) and African American (13.4%) than those with mild disease activity (3.5% and 6.2%, respectively).

Patient journey to diagnosis and physician referral patterns

The most common symptoms prompting first consultation were dry mouth (70.9%), dry eyes (69.2%), physical fatigue (44.5%), stiff/sore muscles (40.5%), and swollen glands (37.8%). Patients most frequently consulted a rheumatologist (38.0% of patients) or a primary care physician (PCP; 31.5%) following the initial onset of symptoms (Figure 2). A rheumatologist eventually made the SjD diagnosis for the large majority of patients (81.4%). Notably, there was a significant delay in SjD diagnosis, with a mean delay of 2.3 years. The mean (±SD) age of patients at diagnosis was 48.2 (±13.4) years. Top reasons cited by physicians for delayed diagnosis were awaiting test results, specialist referral, and inability to consult sooner. Ophthalmologists (43.4%), dentists (23.3%), and PCPs (21.1%) were most frequently involved in patients’ comanagement in the 12 months before data collection; patients had consulted with their rheumatologist on a mean (±SD) of 3.3 (±1.57) occasions (n = 503), an ophthalmologist on 1.1 (±0.96) occasions (n = 370), and a dentist on 1.0 (±1.16) occasions (n = 353). Physicians mostly used serum anti‐SSA/Ro testing for diagnosis; 408 patients had a known serum anti‐SSA/Ro result, of which 359 of 408 (88.0%) were positive. Additional commonly used diagnostic tests included serum anti‐SSB/La, platelet count, erythrocyte sedimentation rate, rheumatoid factor, C‐reactive protein, and antinuclear antibody tests. The most commonly used tests for monitoring patients were platelet count (71.2%), erythrocyte sedimentation rate (54.3%), C‐reactive protein (54.5%), and swollen/tender joint count (44.9%; Figure 2).

Figure 2.

Patient journey to diagnosis, physician referral patterns, and tests used for diagnosis and monitoring. ANA, antinuclear antibody; CRP, C‐reactive protein; dsDNA, double‐stranded DNA; ESR, erythrocyte sedimentation rate; HCP, health care professional; Ig, Immunoglobulin; PCP, primary care physician; PFT, pulmonary function test; RF, rheumatoid factor; SPEP, serum protein electrophoresis; SSA, Sjögren's syndrome antigen A; SSB, Sjögren's syndrome antigen B; U1‐RNP, U1 ribonucleoprotein.

Disease burden (clinical, humanistic, and societal) overall and by disease severity

At the time of data collection, 91% of patients presented with sicca symptoms and systemic involvement. The most common organ systems involved were articular (73%), glandular (35%), and hematologic (25%), with <5% of patients experiencing no organ involvement. In terms of symptomatic burden, dry eyes (86.3%) and dry mouth (62.6%) were the most prevalent sicca symptoms; furthermore, 74.8% and 35.8% of patients experienced physical fatigue and muscle pain, respectively (Supplementary Table S1). Anxiety (20%), depression (19%), and diabetes (15%) were the most commonly reported comorbid conditions.

Of 297 patients with available patient‐reported ESSPRI‐ps data, 33.7% had an ESSPRI‐ps of ≥5 (Table 1). Overall, mean (±SD) ESSPRI‐ps was 4.0 (±2.1): 3.3 (±2.3) in patients with mild severity, 4.1 (±2.0) in those with moderate severity, and 4.8 (±1.7) in patients with severe disease (Table 1). Patient‐reported EQ‐5D‐3L utility, EQ‐VAS, and FACIT‐F scores were numerically lower (indicating greater impairment) across groups with higher ClinESSDAI activity levels, with the numerically highest difference occurring between mild and moderate disease activity levels. When assessing the domains of EQ‐5D‐3L (Figure 3), the scales for pain/discomfort and anxiety/depression were the most heavily affected in patients with mild disease severity. However, consistent with usual disease patterns, patients with severe disease activity reported usual activities and pain/discomfort as the domains most affected. Generally, the proportion of patients experiencing work and activity impairment was numerically greater across groups with higher disease severity (Table 1).

Table 1.

PRO scores by proxy ClinESSDAI scores

PRO	Overall (N = 511)	Mild ClinESSDAI‐ps (0–4; n = 226)	Moderate ClinESSDAI‐ps (5–13; n = 158)	Severe ClinESSDAI‐ps (≥14; n = 127)
EQ‐5D‐3L utility score, n a	303	119	97	87
Mean (±SD)	0.791 (±0.19)	0.843 (±0.19)	0.791 (±0.18)	0.722 (±0.19)
EQ‐VAS score, n a	295	117	91	87
Mean (±SD)	72.1 (±20.09)	78.6 (±19.78)	70.9 (±19.01)	64.5 (±18.83)
FACIT‐F, n b	302	117	98	87
Mean (±SD)	34.4 (±12.42)	37.9 (±12.75)	34.0 (±11.59)	30.2 (±11.57)
WPAI: percent absenteeism, n	147	60	36	51
Mean (±SD)	5.9 (±15.23)	4.0 (±13.44)	6.0 (±18.73)	8.0 (±14.49)
WPAI: percent presenteeism, n	177	79	43	55
Mean (±SD)	23.2 (±24.12)	14.9 (±22.01)	22.1 (±20.07)	35.8 (±24.92)
WPAI: percent overall work impairment, n	145	59	35	51
Mean (±SD)	24.8 (±24.74)	13.0 (±19.46)	24.0 (±20.50)	38.9 (±25.89)
WPAI: percent activity impairment, n	291	115	91	85
Mean (±SD)	30.6 (±24.71)	21.5 (±24.29)	32.5 (±23.03)	40.9 (±22.66)
ESSPRI‐ps, n	297	115	94	88
Mean (±SD)	4.0 (±2.12)	3.3 (±2.29)	4.1 (±2.00)	4.8 (±1.70)
<5, n (%)	197 (66.3)	90 (78.2)	59 (62.8)	48 (54.5)
≥5, n (%)	100 (33.7)	25 (21.8)	35 (37.2)	40 (45.5)
Patient‐reported SjD severity, n	296	119	91	86
Mild, n (%)	119 (40.2)	65 (54.6)	33 (36.3) c	21 (24.4) c
Moderate, n (%)	153 (51.7)	49 (41.2) d	50 (54.9)	54 (62.8) c
Severe, n (%)	24 (8.1)	5 (4.2) d	8 (8.8) d	11 (12.8)

ClinESSDAI, Clinical EULAR Sjögren's Syndrome Disease Activity Index; EQ‐5D‐3L, EuroQoL 5‐Dimension 3‐Level; EQ‐VAS, EuroQoL Visual Analog Scale; ESSPRI, EULAR Sjögren's Syndrome Patient‐Reported Index; FACIT‐F, Functional Assessment of Chronic Illness Therapy–Fatigue; PRO, patient‐reported outcome; ps, proxy score; SjD, Sjögren's disease; WPAI, Work Productivity and Activity Impairment.

^a EQ‐5D utility score ranged from 0 to 1 and EQ‐VAS from 0 to 100, with higher scores indicating better quality of life.

^b FACIT‐F score ranged from 0 to 52, with lower scores indicating worse fatigue.

^c Patients reported less severe disease compared with ClinESSDAI‐ps.

^d Patients reported more severe disease compared with ClinESSDAI‐ps.

Figure 3.

Radar plot showing EQ‐5D‐3L index score for each of the five domains (mobility, self‐care, anxiety/depression, usual activities, and pain/discomfort) by proxy ClinESSDAI severity. EQ‐5D‐3L was scored according to responses for each of the five domains: 2 being “I have no problems in…,” 1 being “I have some problems in…,” and 0 being “I am unable to…”. A score of 0 represents the worst subjective health state. Decreasing scores correspond to increasing difficulty in domains of health‐related quality of life. ClinESSDAI, Clinical EULAR Sjӧgren’s Syndrome Disease Activity Index; EQ‐5D‐3L, EuroQoL 5‐Dimension 3‐Level.

Patient‐reported severity differed from calculated ClinESSDAI‐ps. When ClinESSDAI‐ps was <5, about half of patients overestimated their disease severity, reporting moderate to severe disease. When ClinESSDAI‐ps was 5 to 13, over half of patients reported moderate disease severity, whereas about one‐third reported milder disease; however, among patients with severe ClinESSDAI‐ps (≥14), the large majority (87%) reported lower disease severity. Sleep disturbances were reported by 27% of patients, with 51% of these patients reporting moderate sleep disturbances. Many patients (94%) reported no caregiver requirement for their daily needs.

Current treatment

Approximately half of patients (56%) used over‐the‐counter therapies, specifically eye drops/lubricants, moisturizers, pain killers, nasal spray, or saliva substitutes. The other most commonly used treatment classes were antimalarials (52%), conventional synthetic disease‐modifying antirheumatic drugs (csDMARDs; 38%), secretagogues (36%), nonprescription and prescription nonsteroidal anti‐inflammatory drugs (25%), glucocorticoids (22%), and biologic disease‐modifying antirheumatic drugs (bDMARDs; 12%). Methotrexate (28%) and azathioprine (5%) were the most commonly prescribed csDMARDs. An increase in the number of lines of therapy (Supplementary Figure S1) was associated with increased use of bDMARDs, glucocorticoids, and csDMARDs and decreased use of secretagogues.

Impact of delayed diagnosis on disease severity and PROs

Information on diagnostic delay and disease severity was available from 71 rheumatologists and 276 patients with SjD. Of the 276 patients, 29.3% were diagnosed within 3 months of symptom onset, 20.3% within 3 to 6 months, 17.8% within 6 to 12 months, and 32.6% >12 months after symptom onset. Demographic characteristics of the four patient groups were similar. The mean (±SD) time from symptom onset to diagnosis was 2.3 (±5.6) years (Supplementary Table S2). In unadjusted comparative analyses, changes in the delay from symptom onset to diagnosis were associated with significant differences in EQ‐5D‐3L utility (P = 0.042), EQ‐VAS (P = 0.026), FACIT‐F scores (P = 0.007), and ESSPRI‐ps (P = 0.001). EQ‐5D‐3L utility, EQ‐VAS, and FACIT‐F scores were numerically lower, whereas ESSPRI‐ps were numerically higher, suggesting reduced overall QoL with increased diagnostic delay (Figure 4).

Figure 4.

Comparisons of patient‐reported outcomes (A‐D), proxy ClinESSDAI scores (E), and proxy ESSPRI scores (F) by time from onset of symptoms to diagnosis in patients with SjD. Error bars denote 95% CIs. Higher EQ‐5D‐3L utility scores and EQ‐VAS scores indicate better quality of life; lower FACIT‐F scores indicate worse fatigue. CI, confidence interval; ClinESSDAI, Clinical EULAR Sjögren's Syndrome Disease Activity Index; EQ‐5D‐3L, EuroQoL 5‐Dimension 3‐Level; EQ‐VAS, EuroQoL Visual Analog Scale; ESSPRI, EULAR Sjögren's Syndrome Patient‐Reported Index; FACIT‐F, Functional Assessment of Chronic Illness Therapy–Fatigue; SjD, Sjögren's disease; WPAI, Work Productivity and Activity Impairment.

In adjusted regression models (Table 2), significant differences were observed in EQ‐5D‐3L utility scores and EQ‐VAS scores between patients diagnosed within <3 months of symptom onset and >12 months after onset (P = 0.013 and P = 0.004, respectively), with lower scores (indicating reduced QoL) in patients diagnosed later. Furthermore, significant differences were observed in ClinESSDAI‐ps (<3 vs >12 months, P = 0.018) and ESSPRI‐ps (<3 months vs 3–6 months, P = 0.048; <3 months vs 6–12 months, P = 0.035), with higher scores indicating greater disease activity and poorer QoL in patients diagnosed later. No significant difference was found in ESSPRI‐ps in patients diagnosed <3 and >12 months following symptom onset. FACIT‐F and WPAI scores did not significantly differ with length of delay in diagnosis (Table 2). Adjusted logistic regression models indicate that the odds of having ClinESSDAI‐ps and ESSPRI‐ps ≥5 at the time of data collection are 2.3 and 2.1 times greater, respectively, in patients who received an SjD diagnosis >12 months after symptom onset versus those who were diagnosed in <3 months.

Table 2.

Adjusted logistic and linear regression analyses of ClinESSDAI and ESSPRI proxy scores by time from onset of symptoms to diagnosis in patients with SjD

Time from onset of symptoms to diagnosis	<3 months (n = 81)	3–6 months (n = 56)	6–12 months (n = 49)	>12 months (n = 90)	P value (<3 months: reference)
Linear regression analyses, a coefficient (95% confidence interval)
EQ‐5D‐3L utility (n = 163)	0 (reference)	0.00509 (−0.082 to 0.092)	−0.036 (−0.113 to 0.042)	−0.102 (−0.181 to −0.023)	3–6 months: 0.907; 6–12 months: 0.363; >12 months: 0.013
EQ‐VAS (n = 160)	0 (reference)	−0.048 (−9.207 to 9.111)	−0.905 (−10.402 to 8.591)	−9.819 (−16.325 to −3.313)	3–6 months: 0.992; 6–12 months: 0.849; >12 months: 0.004
FACIT‐F (n = 162)	0 (reference)	2.487 (−2.748 to 7.721)	2.284 (−1.919 to 6.487)	−4.409 (−9.827 to 1.009)	3–6 months: 0.345; 6–12 months: 0.281; >12 months: 0.109
WPAI percent overall work impairment (n = 82)	0 (reference)	2.137 (−14.443 to 18.718)	4.996 (−12.293 to 22.286)	12.343 (−2.674 to 27.359)	3–6 months: 0.796; 6–12 months: 0.563; >12 months: 0.105
Logistic regression analyses, a odds ratio (95% confidence interval)
ClinESSDAI proxy score (n = 276)	1 (reference)	1.115 (0.544 to 2.286)	1.830 (0.849 to 3.947)	2.339 (1.160 to 4.720)	3–6 months: 0.766; 6–12 months: 0.123; >12 months: 0.018
ESSPRI proxy score (n = 161)	1 (reference)	0.236 (0.056 to 0.988)	0.344 (0.128 to 0.925)	2.106 (0.877 to 5.061)	3–6 months: 0.048; 6–12 months: 0.035; >12 months: 0.096

ClinESSDAI, Clinical EULAR Sjögren's Syndrome Disease Activity Index; EQ‐5D‐3L, EuroQoL 5‐Dimension 3‐Level; EQ‐VAS, EuroQoL Visual Analog Scale; ESSPRI, EULAR Sjögren's Syndrome Patient‐Reported Index; FACIT‐F, Functional Assessment of Chronic Illness Therapy–Fatigue; SjD, Sjögren's disease; WPAI, Work Productivity and Activity Impairment.

^a Multivariate linear and logistic regression analyses were adjusted for time since diagnosis, age, sex, body mass index, and ethnicity.

DISCUSSION

This real‐world analysis in patients with SjD describes the patient journey, clinical and patient‐reported burden of SjD, effects of disease severity on PRO measures, and impact of delayed diagnosis on disease severity. The findings of this analysis reaffirm that SjD is a systemic disease with substantial impact beyond sicca manifestations, with dry eyes, physical fatigue, and articular involvement affecting the majority of the cohort. Symptom severity, disease activity, and PROs indicated that the burden of SjD impacts patients’ daily lives and that delays in diagnosis worsen both systemic and symptom severity. Although a variety of treatments were used, in light of current strategies frequently having little evidence of efficacy or being poorly tolerated, ²⁸ our study suggests an unmet treatment need for patients with SjD.

The patient population comprised mainly of White female patients. A wide range of symptoms and systemic manifestations were reported by physicians. Overall, 94% of patients presented with classic sicca symptoms, including 91% with symptoms of dryness and/or organ involvement and 5% with organ involvement only. ²⁹ In addition, patients reported a broad range of symptoms or conditions. In a study using patient online forums and one‐to‐one questions, >75% of patients reported ≥10 symptoms or conditions, with 98% to 100% mentioning dry eyes, dry mouth, and fatigue. ³⁰ Patients reported that the severity of many SjD symptoms varies from day to day, ³⁰ indicating that patients must constantly attend to their symptoms. More recently, similar findings were reported in a collaborative survey generated by patients with SjD and their providers. Among the 2,961 respondents included in the analysis, patients reported >40 SjD‐related symptoms, with dry eyes (94%), dry mouth (93%), and fatigue (82%) occurring in the majority of patients. ³¹ According to the Sjögren's Foundation survey, 72% of patients reported that they were paying out of pocket for their SjD medication, with average annual expenses approaching $6,000 per patient. ³²

Most patients consult their PCP when SjD symptoms first occur, are diagnosed and managed by rheumatologists, and typically involve ophthalmologists, dentists, and PCPs in their management. Due to its variability and insidious presentation, SjD diagnosis by an expert clinician, ideally a rheumatologist, remains the gold standard, ³³ and management should involve a multidisciplinary approach. ²⁰ , ³⁴ In this study, physicians generally followed treatment guidelines, ²⁰ , ³⁴ treating patients with over‐the‐counter therapies, particularly artificial tears, glucocorticoids, and csDMARDs (especially hydroxychloroquine), as well as some biologic therapies. These findings highlight that nonspecific symptoms, including fatigue, muscle pain, and swollen glands, may be overlooked at the first consultation, particularly among nonrheumatologists, whereas delayed diagnostic testing and follow‐up on test results, along with inefficiencies in specialist referral pathways, may substantially contribute to delayed diagnosis and suboptimal early disease management.

Patient‐perceived severity was not synonymous with ClinESSDAI‐ps. Although some patients may overestimate their perceived disease severity, particularly those with ClinESSDAI‐ps <5, most patients with moderate or severe ClinESSDAI‐ps (≥5) tend to underestimate the systemic effects of the disease when reporting their perceived severity. Since implementing ClinESSDAI or assessing the component domains of the disease in clinical practice might not always be feasible, physician assessment of disease severity should be used to complement patient‐reported severity.

Patients experienced deterioration in QoL, worsening fatigue, and greater impairment in work and activity with increasing ClinESSDAI‐ps. In patients with mild ClinESSDAI‐ps, patient‐reported EQ‐VAS and EQ‐5D‐3L index scores were worse than US population norms (79.3 and 0.867, respectively). ³⁵ Furthermore, patients with mild disease activity exhibited mean (±SD) FACIT‐F scores lower than the US general population norm of 43.6 (±9.4). ³⁶ Previous studies have demonstrated that patients’ daily lives, including physical, social, and emotional aspects, are impacted by SjD symptoms ⁹ , ¹⁰ , ³⁰ , ³⁷ and that the SjD disease burden is influenced by levels of fatigue, pain, and organ involvement. ⁹ , ¹³ Baseline sicca symptoms have been correlated with greater fatigue at baseline and increasing fatigue over time. ³⁸ Similarly, another analysis of the Adelphi Primary Sjögren's Syndrome DSP focusing on patients with current or historical systemic organ involvement found that higher proportions of patients with severe disease reported reduced QoL and impairment in work and activity than those with mild and moderate disease. ¹³ These studies, together with the findings of the current analysis, highlight the substantial impact of systemic pain and fatigue on patient functioning and the unmet need of appropriate treatments to address these serious burdens.

Establishing a diagnosis of SjD is challenging due to its symptomatic heterogeneity and organ involvement, which may be subclinical, nonspecific, or subjective, and its management often requires a multidisciplinary team of experts. ¹⁴ , ²⁰ SjD is therefore frequently misdiagnosed, underdiagnosed, or diagnosed in the late stages. ¹⁴ , ²⁰ In assessing the effects of a delay in SjD diagnosis on patient outcomes, our analyses found that approximately one‐third of patients were diagnosed >12 months after symptom onset; the mean time to diagnosis was 2.3 years. This may underestimate the true diagnostic delay experienced by patients with SjD because earlier symptom onset may not have been captured or initial symptoms may not have been attributed to SjD given its heterogeneous nature. Diagnosis is typically delayed for many years after symptom onset, possibly because the presenting symptoms are nonspecific and physician awareness of SjD is generally poor. A US study reported a mean time of seven years from first symptom to diagnosis in patients with a confirmed diagnosis of SjD. ³⁷

Although studies have reported on the burden of SjD symptoms ¹⁰ , ¹³ , ³⁰ , ³⁷ and diagnostic delays, ¹⁴ , ³⁰ to the best of our knowledge, this study is the first to correlate patient outcomes and duration of diagnostic delay using ClinESSDAI‐ps and ESSPRI‐ps. A key finding from the regression analyses was that longer diagnostic delays were associated with moderate/severe disease. Patients diagnosed >12 months after symptom onset were over twice as likely to have ClinESSDAI‐ps and ESSPRI‐ps of ≥5 compared with those diagnosed in <3 months. Comparative analyses demonstrated an inverse correlation between patient outcomes and duration of diagnostic delay, with significant worsening of QoL, fatigue, and ESSPRI‐ps as the delay increases. Moreover, adjusted regression models showed that QoL significantly deteriorated from a <3‐month delay to a >12‐month delay in diagnosis. These findings suggest that patients with longer diagnostic delays experience prolonged uncontrolled disease activity, contributing to cumulative disease burden and poorer patient outcomes.

Work and activity impairment did not significantly differ with length of delay in diagnosis. An interview analysis found that many SjD symptoms were secondary to dryness, such as inflammation of the eyes and speaking difficulties, which subsequently led to difficulties at work; for example, eye dryness caused problems with computer work and driving, and oral dryness made long conversations with customers challenging. ¹⁰ Patients also reported reducing their hours, changing from full‐time to part‐time work, and taking early retirement. ³⁰

Our findings should be considered with regard to the limitations of the DSP survey. Diagnostic heterogeneity is inherent to real‐world data and reflects routine clinical practice, in which diagnoses are typically established based on physician discretion. Therefore, participating rheumatologists were not required to adhere to one or more standardized SjD diagnostic criteria. Given the difficulties of accurately diagnosing SjD, we cannot exclude the possibility that some patients in the population were misdiagnosed with SjD. Similarly, although the literature has shown that antimalarials and DMARDs are ineffective for treating SjD, ³⁹ , ⁴⁰ relatively high rates of antimalarial and DMARD use were observed in this study. This could reflect real‐world attempts to manage systemic symptoms in the absence of approved therapies or the possibility of patients with diagnosed SjD receiving treatment for concomitant autoimmune conditions. Although rheumatologists were asked to provide information for their next six consecutive eligible patients to limit selection bias, no formal verification of patient selection occurred. This analysis may represent the burden and outcomes in patients who visit their rheumatologist frequently and may be more severely affected by SjD than patients who do not consult their rheumatologist as frequently. The low sample size of patients with severe disease and PRO scores limited interpretation of patient burden and outcomes at this severity level. The survey findings relied on the accuracy of retrospective data reported by rheumatologists and patients, particularly the estimated time from symptom onset to diagnosis. Recall bias is a common limitation of surveys; however, survey data were collected at clinic visits and using patients’ medical records to reduce the possibility of recall bias. Furthermore, ClinESSDAI‐ps and ESSPRI‐ps were derived from proxy measures in the survey rather than direct use of the measures, and regression analyses could not control for systemic activity at SjD symptom onset or during the patient's journey to diagnosis. The cross‐sectional design of this analysis does not allow for conclusions about causal relationships, although identification of significant associations is possible. The 3‐month time intervals up to >12 months were selected to provide greater granularity within the first year of diagnostic delay while providing easier interpretation of worsening outcomes for each subsequent 3‐month interval. Despite the limited available data on shorter diagnostic delays in SjD, further research evaluating longer diagnostic delays of two to three years would provide additional valuable insights.

In conclusion, this real‐world analysis indicates that for patients with SjD, the broad range of symptoms and organ involvement presents a significant burden of disease impacting daily life and work. Greater ClinESSDAI severity and delays in diagnosis are associated with worsening symptom severity and poorer patient outcomes. This analysis suggests an unmet treatment need for patients with SjD. Although clinical guidelines exist to help manage treatment decisions, ²⁰ treatment of SjD remains based on the symptomatic treatment of sicca manifestations and immunosuppression for systemic disease. Moreover, the available therapies lack evidence of differential efficacy and are frequently ineffective or poorly tolerated. ²⁰ , ²⁸ Delays in diagnosis may be attenuated by SjD management based in centers of expertise and a multidisciplinary approach, with a principal role for specialists in autoimmune diseases and involvement of PCPs and other health care professionals. ²⁰

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Mr Milligan confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

ROLE OF THE STUDY SPONSOR

Novartis Pharmaceuticals Corporation was involved in the study design, the analysis and interpretation of the data, and the decision to submit the manuscript for publication.

Supporting information

Disclosure form.

Supplementary Table S1: Demographic and clinical characteristics of patients with SjD and characteristics of participating rheumatologists.

Supplementary Table S2: Patient demographic and clinical characteristics by time from onset of symptoms to diagnosis.

Supplementary Figure S1: Treatment use by line of therapy and treatment class.