Abstract
Background:
Eosinophilic esophagitis (EoE) is characterized by eosinophil infiltration into the esophageal tissue and esophageal dysfunction. In the United States, EoE has an estimated prevalence of 26–163 cases per 100,000 people. Real-world data concerning the clinical burden of EoE and treatment patterns in the United States are limited.
Objectives:
To describe the demographics, clinical characteristics, symptoms, comorbidities, treatment pathways, and healthcare resource utilization (HCRU) and costs among patients with EoE in the United States.
Design:
Retrospective analysis.
Methods:
A study of pediatric and adult patients diagnosed with incident EoE (full incident EoE cohort) using Merative™ MarketScan® health insurance claims data between January 1, 2017, and June 30, 2020. A subset of patients (matched incident EoE cohort) was matched with patients without EoE for age, sex, and payor (matched control cohort). Follow-up was 12 months after the EoE diagnosis date. All statistics are descriptive.
Results:
The full incident EoE cohort included 20,290 patients (62.61% were male; median (range) age was 38 (1–93) years); 13,710 patients (matched incident EoE cohort) were matched to 54,727 patients without EoE (matched control cohort). During baseline in the full incident EoE cohort, the most common comorbidities within the Charlson Comorbidity Index (CCI) were chronic pulmonary disease (19.79%), hypertension (15.77%), and depression (9.71%); the most common non-CCI comorbidities were reflux esophagitis (38.84%), allergic rhinitis (19.14%), and depression/anxiety (19.07%). During follow-up, the most frequently reported symptoms were acid reflux/heartburn (56.02%), dysphagia (51.89%), and abdominal pain (30.50%). The most common medications first observed were proton pump inhibitors (42.51%) and oral corticosteroids (12.26%). Overall, a larger proportion of the matched incident EoE cohort had visits to any healthcare setting during baseline and follow-up than the matched control cohort. Correspondingly, the annualized, all-cause healthcare costs per patient were higher in the matched incident EoE cohort than in the matched control cohort at baseline (mean (standard deviation), $10,185 ($29,455); median, $3248 vs $4906 ($20,601); $632) and during follow-up ($15,103 ($35,484); $6708 vs $5200 ($21,314); $651).
Conclusion:
Considerable disease burden is experienced by patients with EoE (before and after diagnosis), which contributes to a high level of HCRU and increased costs.
Introduction
Eosinophilic esophagitis (EoE) is a chronic inflammatory disorder of the esophagus with an estimated prevalence of 13–49 cases per 100,000 people globally,1,2 and 26–163 cases per 100,000 people in the United States.3,4 EoE is associated with a number of clinical characteristics and comorbidities, with symptoms including dysphagia, feeding difficulties, esophageal food impaction, chest pain, abdominal pain, gastroesophageal reflux, and vomiting.5–7 These symptoms can significantly impact health-related quality of life. 8 Early diagnosis and treatment are key to minimizing disease progression and associated complications (including esophageal stricture, esophageal food impaction, esophageal perforation, and malnutrition), thereby reducing this impact.9,10
Treatment of EoE typically includes elimination or elemental diets, proton pump inhibitors (PPIs), corticosteroids (swallowed topical corticosteroids (STCs) or oral corticosteroids (OCSs)), and interventions such as endoscopic esophageal dilation.1,11,12 STCs have demonstrated greater effectiveness at inducing clinical remission compared with PPIs and are recommended in preference to OCSs. 12 OCSs, if prescribed, should be used sparingly due to associated systemic side effects.8,12 The monoclonal antibody dupilumab has recently received approval for the treatment of EoE in patients aged ⩾1 year weighing at least 15 kg. 13
EoE is associated with significant healthcare resource utilization (HCRU) and costs in the United States compared with matched control cohorts.14–17 Previous studies, for example, have shown that patients with EoE are more likely to have frequent inpatient, outpatient, specialist, emergency department (ED) visits, and pharmacy costs than matched non-EoE controls.16,17 However, US data concerning the clinical burden of EoE and treatment patterns are sparse. Here, we evaluate health insurance claims data to describe patient demographics and clinical characteristics, clinical symptoms and comorbidities, treatment patterns and pathways, and all-cause and EoE-related HCRU and associated costs in patients with EoE in the United States.
Methods
Study design
This was a retrospective cohort study of pediatric and adult patients with incident EoE in the United States (Supplemental Figure 1). The overall aim of the study was to describe real-world patient characteristics, treatment patterns, HCRU, and costs of patient with EoE in the United States. Routinely collected data were sourced from the Merative™ MarketScan® Commercial Claims and Encounters (CCAE), Medicare Supplemental and Coordination of Benefits (Medicare), and Multi-State Medicaid databases.
The Merative MarketScan CCAE database includes medical and pharmacy claims for approximately 60 million individuals, primarily working-age adults, their spouses, and dependents, covered under certain employer-sponsored private health insurance. This population typically has higher reimbursement rates and broader access to healthcare services compared to other payor groups. 18 The Medicare Supplemental database contains medical and pharmacy claims for older adults or individuals with disabilities who are enrolled in Medicare and have additional Medicare supplemental coverage. This dataset includes both the Medicare-covered and supplemental insurance-covered portions of the claims, with reimbursement levels influenced by Medicare policies and supplemental plan structure. 19 The Medicaid database represents the healthcare utilization of approximately 70 million low-income individuals and those with end-stage renal disease enrolled in Medicaid across selected, geographically dispersed US states. This population is typically reimbursed at lower rates than those covered by commercial or Medicare plans. 20 The dataset captures claims from both fee-for-service and managed care Medicaid programs; however, individual states are not identifiable due to confidentiality restrictions. All three databases include longitudinal information on demographic characteristics, HCRU across care settings, pharmacy claims, and periods of continuous enrollment. Importantly, the payor groups represented differ not only in reimbursement structures but also in demographic profiles, comorbidity burden, and patterns of HCRU. All data were anonymized and compliant with the patient requirements of the Health Insurance Portability and Accountability Act. 21
Data from January 1, 2016, to June 30, 2021, were included in the analysis. The index date was the date of the first observed diagnosis of EoE during the selection window (January 1, 2017, through June 30, 2020). The baseline period was the 12-months before the EoE index date, and the follow-up period was the 12 months after the EoE index date (Supplemental Table 1). For the analyses of treatment patterns, a patient’s treatment index date were defined as the date of the first observed EoE-related treatment (index treatment) and could have occurred during either the baseline or follow-up period (Supplemental Table 2). The post-treatment period was from the treatment index date until the end of the follow-up period. This study has been reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (Supplemental Material). 22 To note, old terminology has been used throughout, as the database search and the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes preceded the 2022 nomenclature guidelines for eosinophilic gastrointestinal diseases. 23
EoE study population
Pediatric and adult patients who had a first diagnosis of EoE (ICD-10-CM code K20.0; Supplemental Table 3) recorded in the Merative MarketScan CCAE, Medicare, and Medicaid databases during the study selection window (January 1, 2017, to June 30, 2020) were eligible for inclusion in the full incident EoE cohort, subject to the following eligibility criteria: ⩾1 inpatient insurance claim or ⩾2 outpatient insurance claims (⩾30 days apart) containing a diagnosis code for EoE during the study selection window; ⩾12 months of continuous health plan enrollment before the EoE index date; ⩾12 months of continuous health plan enrollment after the EoE index date; no diagnosis of eosinophilic gastritis (ICD-10-CM code K52.81) or eosinophilic gastroenteritis (ICD-10-CM code K52.81) at any time; no diagnosis of EoE during the 12 months before the index date; and >$0 to <$1 million costs in the follow-up period (note: this criterion was added after a preliminary review of the study results to avoid skewing the results owing to patients with serious conditions (e.g., cancer) that did not impact EoE-related costs but whose costs were outliers; Supplemental Figure 1).
Analysis cohorts
This study evaluated three cohorts of patients as follows: all eligible patients with EoE (full incident EoE cohort); a subset of incident EoE patients for whom a matched control patient was assigned from the available pool of patients (matched incident EoE cohort); and matched patients without EoE (matched control cohort). Patient characteristics, symptoms, and treatment patterns were assessed in the full incident EoE cohort. The matched control cohort was used as a comparison group to assess HCRU and the costs compared with the matched incident EoE cohort.
The matched control cohort was randomly selected using a 5% random sample of the total Merative™ MarketScan® population and comprised individuals who had no documented diagnosis of EoE, gastroesophageal reflux disease, eosinophilic gastritis, or eosinophilic gastroenteritis at any time. Gastroesophageal reflux disease was excluded in the matched control cohort as gastroesophageal reflux disease can often be diagnosed in patients prior to EoE diagnosis. Matching was performed at a maximum ratio of one patient with EoE to four patients without EoE and was based on age (year of birth), sex, and health insurance payor type, using direct covariate matching (Supplemental Table 4). The matched controls had ⩾24 months of continuous health plan enrollment, of which the first 12 months comprised the baseline period and the second 12-month period was the follow-up period. The patient index date was assigned as the date occurring 12 months after the start of the 24-month period of continuous health plan enrollment.
Bias
To account for potential differences in underlying HCRU and costs typically accrued by patients within various demographic groups irrespective of an EoE diagnosis, patients from the incident EoE cohort were matched to patients without EoE based on the demographic characteristics described above.
Outcome measurements
The study described the baseline demographic characteristics and the clinical symptoms and comorbidities during the baseline and follow-up periods for all three cohorts (full incident EoE cohort, matched incident EoE cohort, and matched control cohort; see the Supplemental Materials for further details). Comorbidities within the Charlson Comorbidity Index (CCI) were reported, and the CCI score was determined for each patient during the baseline period only (Supplemental Table 5). 24 The predetermined non-CCI comorbidities to be reported included diseases associated with EoE, concomitant allergic diseases, and psychiatric comorbidities; see the Supplemental Material and Supplemental Table 6 for further details.
The study described treatment patterns in the full incident EoE cohort, including treatment classes and EoE-related agents used, first-line treatment pathways, and EoE-related interventions (including esophageal dilation, endoscopic bolus removal, and feeding tube insertion). Treatments that were considered EoE-related are shown in Supplemental Table 2; PPIs and OCSs were considered to be EoE-related if prescribed 60-days before or after an EoE diagnosis, while the others were considered EoE-related irrespective of the diagnosis for which they were prescribed.
HCRU and costs were assessed in the full incident EoE cohort and in the matched incident EoE and matched control cohorts for both all-cause and EoE-related costs (as defined in the Supplemental Materials). Healthcare settings are described in Supplemental Table 7.
Statistical analysis
Demographic characteristics, clinical symptoms, and comorbidities were summarized by the number and percentage of patients in each category (categorical data) or by mean with standard deviation (SD) or median with range, depending on the distribution of the data.
Descriptive analyses were reported for the first-line treatment pathways. The index treatment regimen was defined as that received on the date of the first EoE-related treatment in either the baseline or follow-up periods, and the order of the treatment classes was reported using dates subsequent to the index treatment regimen date. Patients were categorized according to whether they discontinued, switched, or augmented their treatment, as defined in the Supplemental Material. These analyses were conducted prior to the approval of dupilumab as a targeted therapy for EoE.
Descriptive analyses were used to compare the difference in total all-cause healthcare costs between the matched incident EoE and matched control cohorts. Healthcare costs were adjusted to 2021 US dollars using the medical care component of the US Consumer Price Index. Analyses were performed for total costs (all-cause and EoE-related) and for costs by the type of care setting.
Sub-analyses in the full incident EoE cohort characterized patients stratified by age group and by healthcare costs during the follow-up period. Patients were classified as those with “high costs” (defined as ⩾$30,000 per year during the follow-up period, that is, representing approximately the top 10% of the cost distribution) and those with “not-high costs” (<$30,000 per year during the follow-up period).
All analyses were conducted using SAS® statistical software version 9 (SAS Institute, Cary, NC, USA).
Results
Study population
Overall, 20,290 patients with newly diagnosed EoE were included in the full incident EoE cohort (Supplemental Figure 2); of these, 13,710 patients with EoE (matched incident EoE cohort) were matched to 54,727 patients without EoE (matched control cohort).
In the full incident EoE cohort, the majority of patients (62.61%) were male, and the median (range) age was 38 (1–93) years; 77.31% of patients were aged ⩾18 years (Table 1). Most patients (81.90%) were covered by commercial health insurance although this differed by age group: 89.26% of adults (⩾18 years) were covered by commercial health insurance, while children (<12 years old) were equally split between commercial health insurance and Medicaid (Supplemental Table 8). Demographic characteristics of age, sex, and payor type were similar between the matched incident EoE and matched control cohorts (Table 1).
Baseline demographics of the full incident EoE cohort, matched incident EoE cohort, and matched control cohort.
Age at index date, defined as the first observed EoE diagnosis between January 1, 2017, and June 30, 2020.
Geographic region available for patients in the Commercial and Medicare Supplemental and Coordination of Benefits databases only.
EoE, eosinophilic esophagitis; SD, standard deviation.
Clinical burden
EoE-related symptoms were common among the full incident EoE cohort during both baseline and follow-up periods, including acid reflux/heartburn, dysphagia, abdominal pain, and vomiting (Table 2). Acid reflux/heartburn tended to be more frequent among adults and children than adolescents during baseline; during follow-up, this trend was more evident in adults (Table 2). Dysphagia was more frequent among adults than adolescents and children during baseline; this trend was more evident during follow-up, when it was observed with higher frequencies. During baseline and follow-up, abdominal pain, nausea, and vomiting were more frequent among children and adolescents than adults. Chest pain and food impaction were less frequent among children than adolescents/adults. EoE-related symptoms for the matched incident EoE and matched control cohorts are shown in Supplemental Table 9.
Symptoms and comorbidities during baseline and follow-up, and by age category (full incident EoE cohort).
CCI, Charlson comorbidity index; EoE, eosinophilic esophagitis; SD, standard deviation.
In the full incident EoE cohort, the baseline mean (SD) CCI score was 0.81 (1.31), and the three most common CCI comorbidities were chronic pulmonary disease (19.79%), hypertension (15.77%), and depression (9.71%; Table 2). Chronic pulmonary disease was more frequently reported among children and adolescents than in adults, whereas hypertension and depression occurred more frequently among adults (Table 2). In the matched control cohort, 6.02% had chronic pulmonary disease, 16.74% hypertension, and 5.90% depression, compared with 19.51%, 17.26%, and 10.45%, respectively, in the matched incident EoE cohort (Supplemental Table 9).
For the full incident EoE cohort, the most common comorbidities not included in the CCI during both the baseline and follow-up periods included reflux esophagitis (38.84% at baseline and 54.39% during follow-up), allergic rhinitis (19.14% at baseline and 32.53% during follow-up), depression/anxiety (19.07% at baseline and 22.97% during follow-up), and asthma (17.09% at baseline and 22.21% during follow-up; Table 2). Depression/anxiety affected adolescents and adults more frequently than children during both the baseline and follow-up periods, and reflux esophagitis was more frequent in adults during follow-up, but this difference was not evident during the baseline period. Allergic rhinitis and asthma tended to be reported more frequently among children than among adolescents or adults (Table 2). In the matched control cohort, baseline versus follow-up rates were 6.02% versus 6.16% for reflux esophagitis, 7.34% versus 7.19% for allergic rhinitis, 10.87% versus 12.24% for depression/anxiety, and 3.67% versus 3.81% for asthma (Supplemental Table 9).
Treatment patterns
In the full incident EoE cohort, 17,811 (87.78%) patients received EoE-related treatment during either the baseline or follow-up periods (Table 3). The most common EoE-related treatments were PPIs (received by 41.76% of patients overall), OCS (22.54%), antihistamines (17.85%), and STCs (11.23%) during the baseline period and PPIs (67.38%), STCs (28.04%), OCS (23.59%), and antihistamines (19.63%) during the follow-up period (Supplemental Table 10). Among patients who received an EoE-related treatment, the most common index treatments, which could have occurred in either the baseline or follow-up periods, were PPIs (received by 42.51% of patients), OCS (12.26%), antihistamines (8.81%), and STCs (6.69%; Table 3). Overall, 327 (1.61%) patients received enteral nutrition during the baseline period, compared with 518 (2.55%) patients during the follow-up period.
Treatment patterns in patients with incident EoE, overall, and by age category (full incident EoE cohort).
Percentages are calculated among patients who received treatment in the baseline or follow-up period unless otherwise noted.
Percentages among patients with augmentation.
Percentages among patients with treatment switch.
EoE, eosinophilic esophagitis; H2 blocker, histamine type 2 receptor blocker; OCS, oral corticosteroid; PPI, proton pump inhibitor; SD, standard deviation; STC, swallowed topical corticosteroid.
Among patients in the full incident EoE cohort who received EoE-related treatment in either the baseline or follow-up period, 95.42% of patients discontinued their index treatment, with a mean (SD) period from treatment initiation to discontinuation of 189.53 (313.20) days (Table 3). More than half (57.86%) of patients in the full incident EoE cohort receiving EoE-related treatment subsequently switched to another medication class, while 21.36% augmented their index treatment regimen with another medication class (Table 3). Among patients who received EoE-related treatment, the mean (SD) period from treatment initiation to switching was 354.47 (302.18) days, and from treatment initiation to augmentation was 194.79 (200.42) days. The most common classes of treatment to which patients switched were PPIs (32.02% of patients who switched), OCS (23.72%), and STC (16.37%), whereas the most common classes of treatment where augmentation was seen were OCS (31.75% of patients who received any augmented therapy), STCs (22.71%), and PPIs (16.69%).
The proportions of patients undergoing endoscopic bolus removal or endoscopic dilation were low during the baseline period (3.20% and 5.86%, respectively), but increased to 4.82% and 24.41%, respectively, during the follow-up period. Among patients who underwent an EoE-related procedure, the most common treatment classes received in the 3-month period prior to the procedure were PPIs (5.80%) and OCS (5.68%), whereas the most common treatment classes received in the 3-month period after the procedure were PPIs (40.56%) and STC (17.43%; Supplemental Table 11).
The use of psychiatric medications was higher in the matched incident EoE cohort than in the matched control cohort during baseline and follow-up periods, except for antipsychotics during the follow-up period for which usage was similar (Supplemental Table 12).
HCRU and costs
In the full incident EoE cohort, there was a large proportion of patients with all-cause and EoE-related outpatient hospital visits, pharmacy encounters, and physician office visits during baseline and follow-up (Figure 1). There was a trend for visits to these care settings to increase during the follow-up period (all-cause outpatient hospital visits, 67.00% at baseline vs 78.77% at follow-up; all-cause physician office visits, 95.27% at baseline vs 98.06% at follow-up; and all-cause pharmacy encounters, 89.32% at baseline vs 93.77% at follow-up; Figure 1). The greatest increase between baseline and follow-up periods for all-cause HCRU was seen in ambulatory surgical center visits (12.04% at baseline vs 33.73% at follow-up), driven in part by a rise in endoscopic interventions. During the baseline period, 5.86% of patients had all-cause inpatient visits compared with 7.69% during the follow-up period. All-cause ED visits were reported in 32.31% of patients during the baseline period compared with 28.39% during follow-up (Figure 1). All-cause HCRU was higher in the matched incident EoE patient cohort compared with the matched control cohort across all settings at baseline (Figure 2) and follow-up (Figure 3).
Proportion of patients with (a) all-cause and (b) EoE-related health care resource utilization during baseline and follow-up (full incident EoE cohort (N = 20,290)).
Proportion of patients with all-cause healthcare resource utilization during (a) baseline and (b) follow-up (matched incident EoE (N = 13,710) and matched control cohorts (N = 54,727)).
Summary of mean (SD) healthcare resource utilization visits per patient during baseline and follow-up in the matched incident EoE (N = 13,710) and matched control (N = 54,727) cohorts.
In the full incident EoE cohort, mean (SD) total all-cause and total EoE-related healthcare costs per patient were higher in the follow-up period compared with the baseline period (all-cause: mean (SD), $14,996 ($33,958); median (minimum, maximum), $6741 ($2, $926,526) at follow-up vs mean (SD), $10,201 ($31,832); median (minimum, maximum), $3225 ($0, $1,444,720) at baseline; EoE-related: mean (SD), $3401 ($5616); median (minimum, maximum), $1733 ($0, $200,937) at follow-up vs mean (SD) $584 ($2234); median (minimum, maximum), $0 ($0, $117,879) at baseline; Figure 4(a)). The largest proportions of both all-cause and EoE-related healthcare costs per patient per year were accounted for by outpatient hospital visits (Figure 4(b)). The second and third largest contributors to all-cause healthcare costs in the full incident EoE cohort were pharmacy and inpatient hospital visits. Although a smaller proportion of patients in the full incident EoE cohort had inpatient hospital visits than other types of all-cause healthcare encounters (with the exception of telemedicine visits; Figure 1(a)), inpatient hospital visits disproportionately contributed to all-cause healthcare costs per patient per year (Figure 4(b)) at both baseline (mean (SD), $1816 ($19,260); median (minimum, maximum), $0 ($0, $1,287,140), representing 17.80% of the all-cause costs) and follow-up (mean (SD), $2231 ($16,893); median (minimum, maximum), $0 ($0, $765,051), representing 14.88% of the all-cause costs), likely because a relatively small proportion of patients had encounters that were high cost.
(a) Total all-cause and EoE-related healthcare costs, and (b) all-cause and EoE-related healthcare costs by healthcare setting, per patient per year, during baseline, and follow-up (full incident EoE cohort (N = 20,290)).
The mean (SD) annualized, all-cause healthcare cost per patient was $10,185 ($29,455; median (minimum, maximum), $3248 ($0, $1,097,036)) at baseline and $15,103 ($35,484; median (minimum, maximum), $6708 ($2, $896,866)) during follow-up in the matched incident EoE cohort compared with $4906 ($20,601; median (minimum, maximum), $632 ($0, $1,767,015)) and $5200 ($21,314; median (minimum, maximum), $651 ($0, $1,208,066)), respectively, for the matched control cohort (Figure 5(a)). During the follow-up period, outpatient hospital visits, pharmacy costs, and inpatient visits comprised the largest cost drivers for the matched incident EoE cohort (Figure 5(b)).
(a) Total all-cause healthcare costs and (b) all-cause healthcare costs by healthcare setting, per patient per year, during baseline and follow-up (matched incident EoE (N = 13,710) and matched control (N = 54,727) cohorts).
Annualized EoE-related healthcare costs per patient at follow-up were higher in pediatric patients (aged <12 years; mean (SD), $5301 ($7161); median (minimum, maximum), $3062 ($0, $88,444)) and adolescents (aged 12–17 years; mean (SD), $5420 ($8690); median (minimum, maximum), $2904 ($0, $200,937)) than adults (aged ⩾18 years; mean (SD), $2825 ($4561); median (minimum, maximum), $1513 ($0, $136,185); Supplemental Figure 3(a)). The main cost drivers for this increase were outpatient visits, physician office visits, and pharmacy visits for both pediatric and adolescent patients. In adults, the main drivers for EoE-related healthcare costs were inpatient and ambulatory surgical center visits (Supplemental Figure 3(b)).
Further analysis split the full incident EoE cohort into a group with healthcare “high costs” (costs ⩾$30,000 during follow-up) and a group with “not high costs” (costs <$30,000 during follow-up). In total, 2121 patients (10.45%) in the full incident EoE cohort were categorized as having “high costs,” and 18,169 patients (89.55%) as having “not high costs.” At baseline, a larger proportion of patients having “high costs” versus “not high costs” had chronic pulmonary disease (28.15% vs 18.81%), hypertension (21.97% vs 15.05%), and depression (17.63% vs 8.79%; Supplemental Table 13). The symptom of abdominal pain had the largest absolute difference at baseline in the percentage of patients with a symptom between the “high costs” and “not high costs” groups (38.28% vs 28.85%), with similar findings observed during the follow-up period (Supplemental Table 13). The mean (SD) total annualized, all-cause healthcare costs per patient for those having “high costs” was $45,227 ($83,885) at baseline and $76,406 ($80,212) during follow-up (median (minimum, maximum), $22,128 ($0, $1,444,720) at baseline and $51,772 ($30,033, $926,526) at follow-up), with inpatient hospital visits, outpatient hospital visits, and pharmacy claims being the largest contributors (Supplemental Figure 4(a) and (b)). In those having “not high costs,” the mean (SD) total annualized, all-cause healthcare cost per patient was $6112 ($12,269) during baseline and $7827 ($6732) during follow-up (median (minimum, maximum), $2775 ($0, $359,940) at baseline and $5851 ($2, $29,990) at follow-up; Supplemental Figure 4(a)), with outpatient hospital visits, physician office visits, and pharmacy costs being the largest contributors (Supplemental Figure 4(b)). The difference in all-cause costs between those having “high costs” and those having “not high costs” was driven by inpatient visits costs (Supplemental Figure 4(b)), which reflected a larger proportion of patients having “high costs” experiencing an inpatient visit compared with those having “not high costs” (Supplemental Figure 5(a) and (b)).
Mean (SD) annualized EoE-related healthcare cost per patient having “high costs” was $1030 ($4147) during baseline and $7851 ($12,911) during the follow-up period (median (minimum, maximum), $6 ($0, $69,282) at baseline and $3206 ($0, $200,937) at follow-up; Supplemental Figure 4(c)), reflecting an increase of more than seven-fold, driven mainly by an increase in outpatient hospital visits and second by inpatient visits (Supplemental Figure 4(d)). By contrast, the mean (SD) annualized EoE-related healthcare cost per patient having “not high costs” was $532 ($1882) at baseline and $2881 ($3631) at follow-up (median (minimum, maximum), $0 ($0, $117,879) at baseline and $1632 ($0, $28,421) at follow-up; Supplemental Figure 4(c)), with the costs at follow-up driven mainly by outpatient hospital and physician office visits (Supplemental Figure 4(d)). EoE-related ambulatory surgical center visits were slightly more frequent in those having “not high costs” than in those having “high costs” (Supplemental Figure 5(c) and (d)). The difference in EoE-related costs between the two groups was driven by outpatient hospital costs and inpatient hospital visits.
Discussion
This retrospective cohort study from commercial and noncommercial insurance data provides valuable information on real-world patient characteristics and the clinical burden of EoE, including treatment pathways and patterns, and related HCRU and costs for pediatric and adult patients with EoE in the United States.
Patients with EoE had a high disease burden in terms of comorbidities and EoE-related symptoms, with large proportions of patients experiencing acid reflux/heartburn, dysphagia, and abdominal pain. We observed higher reporting of symptoms during the follow-up period, which may be because of greater awareness following diagnosis. The most common comorbidity was reflux esophagitis reported in approximately 39% of patients during the baseline period, which is consistent with previous results (between 35% and 51% of patients).17,25 The proportions of patients undergoing endoscopic bolus removal or endoscopic dilation were low and also lower than previously reported, 26 which may reflect inconsistencies in coding. Notably, patients with EoE were already experiencing a high disease burden at diagnosis, consistent with previous reports.17,27 In particular, children and adolescents had a higher frequency of concomitant allergic diseases compared with adults, as has been previously reported. 25 Depression/anxiety was one of the most common comorbidities observed in patients with EoE during the baseline and follow-up periods in the current study and was reported by almost one in four adult patients during the follow-up period. This is consistent with the impact of EoE on patient psychological health that has been demonstrated in a number of earlier studies.27–30 Indeed, the use of psychiatric medications was much higher in EoE patients than in matched controls in the present analysis, highlighting the need for a multidisciplinary approach to supporting patients with EoE.
The most common EoE-related index treatment for patients with EoE was a PPI, in accordance with current management guidelines, 12 and retrospective and real-world studies.31–33 More than half (~58%) of patients with EoE who received treatment switched to a second-line therapy, which is consistent with previous observations of 41%–52%.32,33 Approximately 21% of treated patients augmented their index treatment with another medication class. These results underscore the limitations of standard-of-care therapies, which often necessitate the use of rescue therapies, such as OCS, and highlight the unmet need for targeted therapies.
Notably, nearly all patients (~95%) receiving treatment for EoE discontinued their index treatment regimen. This high discontinuation rate may reflect a combination of factors, including suboptimal symptom control, side effects, treatment burden, transition to dietary or nonpharmacologic medications, limited access to certain therapies or fluctuating disease activity. Additionally, variations in provider practice patterns and the historical lack of standardized long-term treatment strategies may have contributed to early treatment cessation. In addition, it is important to recognize that our definition of discontinuation, based on a 60-day prescription gap, relies on assumptions inherent to claims data. Some patients classified as discontinued may have continued therapy but experienced delays in refilling their prescriptions or may have been taking medications less frequently than prescribed, patterns that cannot be fully captured in this dataset. Interestingly, only ~58% of those receiving treatment switched to a second-line treatment, implying that a notable proportion of patients may have opted not to pursue any subsequent therapy at all. The reasons for this trend remain unclear: potential factors may include no need to manage their symptoms, transition to dietary or nonpharmacological medications, challenges adhering to treatment regimes, a lack of efficacy or perceived efficacy, a lack of motivation to explore alternative options, a lack of viable alternatives being available, the intermittent nature of EoE symptoms, and a lack of awareness that EoE is a chronic disease with a need for ongoing therapy.
Although dupilumab is an approved, targeted biologic therapy for EoE based on phase III results,34,35 the analysis presented in this article was conducted prior to its approval for EoE. Biologic treatments may prove advantageous in treating the underlying pathology of EoE in routine clinical practice, but further studies are needed to establish the potential merits of real-world biologic use occurring post approval.
Consistent with the high disease burden observed in this analysis, our findings also showed higher use of healthcare resources and increased costs in patients with incident EoE than in the directly comparable matched control cohort without EoE. Notably, healthcare costs remained high after 1 year of treatment in patients with EoE, with the cost of outpatient visits being a major cost driver. These findings are consistent with earlier reports.14,17 Among the subset of patients with “high cost” healthcare, inpatient hospital visits accounted for much of these higher costs. It is important to note that cost estimates in the study are based on insurer-paid amounts as captured in the claims data, reflecting healthcare expenditures from the payor perspective. The three insurance databases analyzed—commercial, Medicare Supplemental and Medicaid—cover distinct US populations with differing reimbursement levels and access to care. As such, the reported costs represent negotiated payments rather than the full economic or opportunity costs and may vary by insurance type, geographic region, and provider contracts. Therefore, the exact numerical cost estimates are not likely to be directly comparable across healthcare systems or generalizable outside the United States.
The limitations of this study are common to retrospective observational studies using routinely collected insurance claims data. This study included individuals with commercial health coverage, private Medicare supplemental coverage, or Medicaid coverage, which provides a broader patient pool than is observed in studies that only include patients with commercial insurance; however, the results may not be generalizable to those without health insurance coverage. Only patients with 12 months of continuous health plan enrollment after their first observed EoE diagnosis were evaluated, so the study may have been biased toward healthier patients with greater longevity; however, it was not anticipated that a large percentage of patients with EoE would have died during follow-up. Diagnoses of EoE were based on ICD-10-CM codes from insurance claims data, which may be subject to under-reporting or misclassification; while a requirement for two outpatient claims 30 days apart was used to improve specificity, the absence of clinical data or endoscopic confirmation limits diagnostic certainty. Additionally, the analysis was limited by an absence of detailed clinical information, such as index endoscopy, dilation procedures, and initial treatment regimes, within the claims data; further research integrating clinical datasets may help elucidate how these factors influence cost variations in patients with EoE. Second, the datasets represent information collected for clinical and routine use rather than specifically for research purposes. Although extensive quality control and validity checks are conducted at the practice level, the validity and completeness of individual patient records cannot be assessed. Geographic region was not included as a matching variable which may have introduced variability in outcomes due to regional differences in insurance providers, healthcare access, treatment practices, and resource availability. As this was a descriptive study, no formal statistical tests were conducted to compare healthcare costs across children, adolescents, and adults; therefore, the observed cost differences should be interpreted as descriptive trends rather than statistically validated effects. Third, there is potential for misclassification of disease status or covariates as patients were identified through administrative claims data, and these data are subject to data coding limitations and data entry error. For instance, medication use was based on prescription collections, which do not necessarily fully reflect a patient’s real-world medication use in that medications that were paid for out of pocket or over the counter, as well as medication adherence, were not captured in the databases. Comprehensive information on dietary interventions, such as elimination diets, could not be reliably captured in the data source unless associated with an insurance billing claim and therefore could not be fully evaluated. As a result, the impact of dietary management on treatment patterns and outcomes in this study could not be thoroughly assessed. Moreover, as claims data do not capture clinical intent or confirm concurrent medication use, there is an inherent limitation in definitively identifying true treatment augmentation. Fourth, there is a risk of selection bias as patients receiving corticosteroids may have had more contact with their doctors and/or be more closely monitored and, thus, may have been more likely to receive a diagnosis than patients not exposed to corticosteroids. Fifth, the study was conducted and analyzed before the approval of dupilumab in the United States; thus, it does not capture any practice changes resulting from the approval of the first targeted therapy for EoE. Lastly, the potential impacts of the coronavirus 2019 pandemic were not assessed.
Conclusion
This study includes wide-ranging commercial and noncommercial insurance data, and it provides new insights into treatment patterns, switching, and augmentation therapies for patients with EoE. Our findings indicate a high level of disease burden experienced by patients with EoE in the United States that was already present before diagnosis and persisted during follow-up. This high disease burden is reflected in higher HCRU and related costs for patients with EoE, which suggests the need for new treatment approaches.
Supplemental Material
sj-docx-1-tag-10.1177_17562848251391090 – Supplemental material for Characteristics, treatment patterns, healthcare resource utilization, and costs in patients with eosinophilic esophagitis in the United States: a retrospective analysis of insurance claims data
Supplemental material, sj-docx-1-tag-10.1177_17562848251391090 for Characteristics, treatment patterns, healthcare resource utilization, and costs in patients with eosinophilic esophagitis in the United States: a retrospective analysis of insurance claims data by Xiao Xu, Ekaterina Maslova, Danuta Kielar, Heide Stirnadel-Farrant, Juliana Meyers, Mirna Chehade, Rohit Katial and Stephanie Y. Chen in Therapeutic Advances in Gastroenterology
Supplemental Material
sj-docx-2-tag-10.1177_17562848251391090 – Supplemental material for Characteristics, treatment patterns, healthcare resource utilization, and costs in patients with eosinophilic esophagitis in the United States: a retrospective analysis of insurance claims data
Supplemental material, sj-docx-2-tag-10.1177_17562848251391090 for Characteristics, treatment patterns, healthcare resource utilization, and costs in patients with eosinophilic esophagitis in the United States: a retrospective analysis of insurance claims data by Xiao Xu, Ekaterina Maslova, Danuta Kielar, Heide Stirnadel-Farrant, Juliana Meyers, Mirna Chehade, Rohit Katial and Stephanie Y. Chen in Therapeutic Advances in Gastroenterology
Footnotes
Acknowledgements
The authors would like to thank Seema Aceves (University of California, San Diego) for contributions to the study design. Medical writing support was provided by Caroline Ridley, PhD, and Juliette Gray, PhD, of Springer Health+, Springer Healthcare Ltd, UK, which was funded by AstraZeneca in accordance with Good Publication Practice (GPP 2022 guidelines (
).
Author’s note
Danuta Kielar: Affiliation at time of study conduct.
Declarations
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
